{
  "name": "Neocloud rankings",
  "url": "https://neocloudrankings.com/",
  "license": "CC BY 4.0",
  "cite": "Neocloud Rankings, https://neocloudrankings.com/",
  "updated": "2026-11-11",
  "fields": [
    {
      "key": "rank",
      "label": "Rank",
      "type": "number",
      "description": "Position by composite score, computed at build time from the signals below with the weights on /methodology/. Ties share the higher position. Never set by hand and never for sale.",
      "filterable": false,
      "sortable": true
    },
    {
      "key": "score",
      "label": "Score",
      "type": "number",
      "unit": "/100",
      "description": "Composite score out of 100: ClusterMAX tier 30%, contracted power 20%, contracted backlog 15%, capital raised 15%, GPUs 10%, named customers 10%. A signal the company does not disclose scores 0.",
      "filterable": true,
      "sortable": true
    },
    {
      "key": "clustermax_tier",
      "label": "ClusterMAX",
      "type": "enum",
      "description": "Tier in SemiAnalysis ClusterMAX 3.0 (23 Sep 2026), the public hands-on rating of GPU clouds. Medal tiers run Platinum, Gold, Silver, Bronze; Participation Ribbon sits between Bronze and Underperforming. Underperforming and Unavailable are the two groups SemiAnalysis lists under \"Not Recommended\": Unavailable covers providers it could not test or judged too small to rate. Empty when the provider is not listed.",
      "filterable": true,
      "sortable": true,
      "enumValues": [
        "Platinum",
        "Gold",
        "Silver",
        "Bronze",
        "Participation Ribbon",
        "Underperforming",
        "Unavailable"
      ]
    },
    {
      "key": "contracted_mw",
      "label": "Contracted power",
      "type": "number",
      "unit": "MW",
      "description": "Power the company states as contracted, secured or committed in its portfolio, in MW. Speculative pipeline and planned campuses are excluded; the power note quotes the company's own word for it.",
      "filterable": true,
      "sortable": true
    },
    {
      "key": "active_mw",
      "label": "Active power",
      "type": "number",
      "unit": "MW",
      "description": "Power the company states as active, live, operational or managed today, in MW.",
      "filterable": false,
      "sortable": true
    },
    {
      "key": "power_note",
      "label": "Power note",
      "type": "string",
      "description": "What the power figures cover, the as-of date, and targets or planned campuses that are not counted.",
      "filterable": false,
      "sortable": false
    },
    {
      "key": "backlog_usd_m",
      "label": "Backlog",
      "type": "number",
      "unit": "$M",
      "description": "Contracted revenue backlog, remaining performance obligations or total contracted value as the company states it, in USD millions.",
      "filterable": true,
      "sortable": true
    },
    {
      "key": "backlog_note",
      "label": "Backlog note",
      "type": "string",
      "description": "Which backlog measure the company uses and its as-of date.",
      "filterable": false,
      "sortable": false
    },
    {
      "key": "revenue_q_usd_m",
      "label": "Quarterly revenue",
      "type": "number",
      "unit": "$M",
      "description": "Latest reported quarterly revenue, in USD millions, from the company's results release or filing. Company-level unless the note says otherwise.",
      "filterable": false,
      "sortable": true
    },
    {
      "key": "arr_usd_m",
      "label": "Stated ARR",
      "type": "number",
      "unit": "$M",
      "description": "Annualized run-rate revenue exactly as the company states it, in USD millions. We never annualize a quarter ourselves.",
      "filterable": false,
      "sortable": true
    },
    {
      "key": "revenue_note",
      "label": "Revenue note",
      "type": "string",
      "description": "Period, basis and any revenue targets the company has stated.",
      "filterable": false,
      "sortable": false
    },
    {
      "key": "gpus",
      "label": "GPUs",
      "type": "number",
      "description": "GPUs deployed, owned or under contract as the company states it.",
      "filterable": true,
      "sortable": true
    },
    {
      "key": "gpus_note",
      "label": "GPU note",
      "type": "string",
      "description": "What the GPU figure counts and when it was stated.",
      "filterable": false,
      "sortable": false
    },
    {
      "key": "capital_usd_m",
      "label": "Capital raised",
      "type": "number",
      "unit": "$M",
      "description": "Equity and debt raised, in USD millions: the total when the company states one, otherwise the rounds and facilities named in the capital note.",
      "filterable": true,
      "sortable": true
    },
    {
      "key": "capital_note",
      "label": "Capital note",
      "type": "string",
      "description": "Exactly which rounds, notes or facilities the capital figure covers, with dates.",
      "filterable": false,
      "sortable": false
    },
    {
      "key": "customer_count",
      "label": "Named customers",
      "type": "number",
      "description": "Count of the customers listed in the customers field, computed at build time.",
      "filterable": false,
      "sortable": true
    },
    {
      "key": "customers",
      "label": "Customers",
      "type": "string",
      "description": "Customers named publicly by the company, in its filings or in reputable press.",
      "filterable": false,
      "sortable": false
    },
    {
      "key": "listing",
      "label": "Ownership",
      "type": "enum",
      "description": "Publicly listed, privately held, or a subsidiary of another company.",
      "filterable": true,
      "sortable": false,
      "enumValues": [
        "public",
        "private",
        "subsidiary"
      ]
    },
    {
      "key": "ticker",
      "label": "Ticker",
      "type": "string",
      "description": "Exchange and ticker symbol for listed companies.",
      "filterable": false,
      "sortable": true
    },
    {
      "key": "parent_company",
      "label": "Parent",
      "type": "string",
      "description": "Owner or majority shareholder when the neocloud is part of another group.",
      "filterable": false,
      "sortable": false
    },
    {
      "key": "capex_q_usd_m",
      "label": "Quarterly capex",
      "type": "number",
      "unit": "$M",
      "description": "Latest disclosed quarterly capital expenditures, in USD millions. Context only; excluded from the composite score.",
      "filterable": false,
      "sortable": false
    },
    {
      "key": "capex_note",
      "label": "Capex note",
      "type": "string",
      "description": "Period and scope of capital expenditures, including whether figures are actuals or guidance.",
      "filterable": false,
      "sortable": false
    },
    {
      "key": "gpu_models",
      "label": "GPU models",
      "type": "string",
      "description": "Hardware advertised or tested, with reservations and future availability explicitly noted. Context only; not a fleet count or scoring input.",
      "filterable": false,
      "sortable": false
    },
    {
      "key": "regions",
      "label": "Operating footprint",
      "type": "string",
      "description": "Company or service geography as sourced; a general cloud footprint does not guarantee GPU stock in each region.",
      "filterable": false,
      "sortable": false
    },
    {
      "key": "orchestration",
      "label": "Deployment and orchestration",
      "type": "string",
      "description": "Available provisioning and cluster management services; future features are labelled as such.",
      "filterable": false,
      "sortable": false
    },
    {
      "key": "interconnect",
      "label": "Cluster interconnect",
      "type": "string",
      "description": "Published cluster network technology or bandwidth; not a measured service-quality claim.",
      "filterable": false,
      "sortable": false
    },
    {
      "key": "capacity_note",
      "label": "Capacity scope",
      "type": "string",
      "description": "Partner supply, network coverage and other non-fleet capacity signals excluded from GPU scoring.",
      "filterable": false,
      "sortable": false
    }
  ],
  "records": [
    {
      "id": "coreweave",
      "name": "CoreWeave",
      "url": "https://www.coreweave.com",
      "summary": "The largest listed neocloud: NVIDIA GPU clusters rented to AI labs, enterprises and trading firms from US and European data centers. It holds a $132.5B revenue backlog against 4.6 GW of contracted power.",
      "hq": "US",
      "founded": 2017,
      "tags": [
        "Public",
        "US",
        "Europe"
      ],
      "fields": {
        "listing": "public",
        "ticker": "NASDAQ: CRWV",
        "parent_company": null,
        "clustermax_tier": "Platinum",
        "contracted_mw": 4600,
        "active_mw": 1700,
        "power_note": "Over 4.6 GW contracted (up more than 900 MW in the quarter) and over 1.7 GW active, reported with Q3 2026 results (10 Nov 2026).",
        "backlog_usd_m": 132500,
        "backlog_note": "Revenue backlog of about $132.5B at 30 Sep 2026.",
        "revenue_q_usd_m": 3661,
        "arr_usd_m": null,
        "revenue_note": "Q3 2026 revenue, quarter to 30 Sep 2026 (up from $1,365M a year earlier).",
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": 10301,
        "capital_note": "Proceeds from issuance of debt, net, for the three months to 30 Sep 2026 (cash flow statement), including a ~$4.2B upsized convertible notes offering and the $2.6B DDTL 5.5 facility.",
        "customers": "OpenAI, Jane Street, Perplexity, Cohere, Hippocratic AI, Chan Zuckerberg Initiative, Canva, Wayve, Substack, Serve Robotics, Parallel Web Systems, Mirendil",
        "rank": 1,
        "score": 89,
        "customer_count": 12
      },
      "sources": [
        {
          "url": "https://investors.coreweave.com/news/news-details/2026/CoreWeave-Reports-Strong-Third-Quarter-2026-Results/default.aspx",
          "title": "CoreWeave Reports Strong Third Quarter 2026 Results",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "ticker",
            "contracted_mw",
            "active_mw",
            "power_note",
            "backlog_usd_m",
            "backlog_note",
            "revenue_q_usd_m",
            "revenue_note",
            "capital_usd_m",
            "capital_note",
            "customers"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "crusoe",
      "name": "Crusoe",
      "url": "https://www.crusoe.ai",
      "summary": "Denver-based builder of AI data center campuses that also runs Crusoe Cloud, renting NVIDIA and AMD GPUs. It built the Abilene, Texas campus used by OpenAI and states 6 GW+ of contracted capacity.",
      "hq": "US",
      "tags": [
        "Private",
        "US",
        "Europe"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Bronze",
        "contracted_mw": 6000,
        "active_mw": 1000,
        "power_note": "6 GW+ of gross contracted capacity across data centers and cloud; 1 GW of gross capacity delivered and operational (Series F release, 17 Sep 2026).",
        "backlog_usd_m": 140000,
        "backlog_note": "$140B+ total contracted value (TCV) stated in the Series F release, Sep 2026.",
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": 3900,
        "capital_note": "$3.9B Series F at a $30.9B post-money valuation, co-led by Atreides Management, Mubadala Capital and Valor Equity Partners; initial close announced 17 Sep 2026.",
        "customers": "Cognition, Figure, Perplexity",
        "rank": 2,
        "score": 67.5,
        "customer_count": 3
      },
      "sources": [
        {
          "url": "https://www.crusoe.ai/resources/newsroom/crusoe-announces-series-f-funding",
          "title": "Crusoe Raises $3.9 Billion Series F",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "contracted_mw",
            "active_mw",
            "power_note",
            "backlog_usd_m",
            "backlog_note",
            "capital_usd_m",
            "capital_note",
            "customers"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "nebius",
      "name": "Nebius",
      "url": "https://nebius.com",
      "summary": "Amsterdam-based AI cloud listed on Nasdaq, running NVIDIA clusters in Europe and the US for AI labs and enterprises. It reached $3.0B ARR at the end of Q2 2026 with more than $40B in customer commitments.",
      "hq": "NL",
      "tags": [
        "Public",
        "Europe",
        "US"
      ],
      "fields": {
        "listing": "public",
        "ticker": "NASDAQ: NBIS",
        "parent_company": null,
        "clustermax_tier": "Platinum",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": "Q2 shareholder letter targets 5 GW contracted power at year-end 2026; this is guidance, not a current contracted-MW disclosure. Q3 results scheduled for 12 Nov 2026 before market open.",
        "backlog_usd_m": 40000,
        "backlog_note": "More than $40B of investment-grade customer commitments in long-term contracts (Citi facility release, 22 Oct 2026).",
        "revenue_q_usd_m": 582.3,
        "arr_usd_m": 3000,
        "revenue_note": "Quarter ended 30 Jun 2026, released 12 Aug 2026: group revenue $582.3M; ARR $3.0B at quarter-end (last month of AI cloud revenue x 12, company definition). Latest reported quarter verified at the 11 Nov 2026 snapshot.",
        "gpus": null,
        "gpus_note": "Q2 shareholder letter describes initial Vera Rubin NVL72 systems for validation, but gives no fleet total; GPU count remains undisclosed.",
        "capital_usd_m": 3500,
        "capital_note": "$3.5B 364-day senior secured financing facility from Citi, 22 Oct 2026.",
        "customers": "Microsoft, Meta, Reflection, Cohere",
        "capex_q_usd_m": 5700,
        "capex_note": "Quarter ended 30 Jun 2026, released 12 Aug 2026: capital expenditures approximately $5.7B, primarily GPUs, GPU-related hardware and data center expansion. Latest reported quarter verified at the 11 Nov 2026 snapshot.",
        "rank": 3,
        "score": 58.9,
        "customer_count": 4
      },
      "sources": [
        {
          "url": "https://nebius.com/about",
          "title": "About Nebius",
          "accessed": "2026-11-09",
          "fields": [
            "listing",
            "ticker"
          ]
        },
        {
          "url": "https://nebius.com/newsroom/nebius-secures-3-5-billion-financing-facility-with-citi",
          "title": "Nebius secures $3.5 billion financing facility with Citi",
          "accessed": "2026-11-11",
          "fields": [
            "capital_usd_m",
            "capital_note",
            "backlog_usd_m",
            "backlog_note"
          ]
        },
        {
          "url": "https://nebius.com/newsroom/nebius-reports-second-quarter-2026-financial-results",
          "title": "Nebius reports second quarter 2026 financial results",
          "accessed": "2026-11-11"
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "https://assets.nebius.com/assets/a6ecfd85-a6cb-4967-8ef7-9a25bd261f9c/SHLQ226.pdf",
          "title": "Nebius Q2 2026 shareholder letter, pp. 3–5 and 9–11",
          "accessed": "2026-11-11",
          "fields": [
            "arr_usd_m",
            "revenue_q_usd_m",
            "revenue_note",
            "power_note",
            "gpus_note",
            "capex_q_usd_m",
            "capex_note",
            "customers"
          ]
        },
        {
          "url": "https://nebius.com/newsroom/nebius-group-announces-date-of-third-quarter-2026-results-and-conference-call",
          "title": "Nebius Q3 2026 reporting date (30 Oct 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "power_note",
            "capex_note"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "nscale",
      "name": "Nscale",
      "url": "https://www.nscale.com",
      "summary": "London-based AI cloud building GPU data centers in Europe and the US for Microsoft and Anthropic, with an S-1 filed for a NYSE listing. It reports $103.4B of contract value and 50,000+ GPUs deployed in 2026.",
      "hq": "GB",
      "tags": [
        "Private",
        "Europe",
        "US"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Unavailable",
        "contracted_mw": 1370,
        "active_mw": null,
        "power_note": "Approximately 1.37 GW of active and contracted IT capacity at 31 Aug 2026 (S-1, 18 Sep 2026). Potential 10 GW development runway is excluded. No separate active-MW total verified in the filing.",
        "backlog_usd_m": 103400,
        "backlog_note": "$103.4B active and contracted total contract value at 31 Aug 2026, including $2.6B active TCV. TCV is not recognized revenue or a directly comparable ASC 606 RPO measure.",
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": "H1 2026 revenue $140.6M, up 1,252% (S-1). No quarterly figure published.",
        "gpus": 50000,
        "gpus_note": "Over 50,000 GPUs deployed in 2026, 24,000 of them in the prior month; six active data centers in five countries (company, 1 Oct 2026).",
        "capital_usd_m": 3360,
        "capital_note": "$3.36B pre-IPO convertible loan notes led by Third Point, 25 Sep 2026: a $2.36B initial tranche plus a $1B NVIDIA commitment due to fund in mid-November 2026.",
        "customers": "Microsoft, Anthropic",
        "rank": 4,
        "score": 53.1,
        "customer_count": 2
      },
      "sources": [
        {
          "url": "https://www.sec.gov/Archives/edgar/data/2110365/000119312526395475/ck0002110365-20260918.htm",
          "title": "Nscale S-1, filed 18 Sep 2026: prospectus summary, key metrics and H1 financials",
          "accessed": "2026-11-11",
          "fields": [
            "contracted_mw",
            "power_note",
            "backlog_usd_m",
            "backlog_note",
            "revenue_note",
            "customers"
          ]
        },
        {
          "url": "https://www.nscale.com/press-releases/execution-at-scale",
          "title": "Nscale: execution at scale",
          "accessed": "2026-11-11",
          "fields": [
            "gpus",
            "gpus_note"
          ]
        },
        {
          "url": "https://www.nscale.com/press-releases/pre-ipo-convertible-financing",
          "title": "Nscale pre-IPO convertible financing",
          "accessed": "2026-11-11",
          "fields": [
            "capital_usd_m",
            "capital_note"
          ]
        },
        {
          "url": "https://www.nscale.com/press-releases/nscale-files-initial-public-offering",
          "title": "Nscale files for initial public offering",
          "accessed": "2026-11-09",
          "fields": [
            "listing"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "tensorwave",
      "name": "TensorWave",
      "url": "https://tensorwave.com",
      "summary": "AMD-only GPU cloud renting Instinct MI300X, MI325X, MI355X and MI455X clusters. It has raised over $500M and rates Silver in ClusterMAX 3.0.",
      "hq": "US",
      "tags": [
        "Private",
        "US",
        "AMD"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Silver",
        "contracted_mw": 2000,
        "active_mw": null,
        "power_note": "More than 2 GW of long-term data center capacity secured (10 Jun 2026 Series B announcement). Secured capacity includes future deployments; active MW not stated.",
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": 8192,
        "gpus_note": "8,192 AMD Instinct MI325X GPUs online in a North American training cluster, stated 10 Jun 2026. This is a disclosed cluster count, not a claimed current total across all models.",
        "capital_usd_m": 500,
        "capital_note": "Over $500M raised in total (company about page).",
        "customers": "Moreh",
        "rank": 5,
        "score": 50.7,
        "customer_count": 1
      },
      "sources": [
        {
          "url": "https://tensorwave.com/about",
          "title": "About TensorWave",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "capital_usd_m",
            "capital_note"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "https://tensorwave.com/blog/tensorwave-raises-350-million-series-b-at-1-55b-valuation-to-expand-global-amd-powered-ai-infrastructure",
          "title": "TensorWave Series B announcement, 10 Jun 2026 (article body in page JSON)",
          "accessed": "2026-11-11",
          "fields": [
            "gpus",
            "gpus_note",
            "contracted_mw",
            "power_note"
          ]
        },
        {
          "url": "https://tensorwave.com/",
          "title": "TensorWave customer case study: Moreh",
          "accessed": "2026-11-11",
          "fields": [
            "customers"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "iren",
      "name": "IREN",
      "url": "https://iren.com",
      "summary": "Former bitcoin miner turned AI cloud, listed on Nasdaq, building GPU capacity for Microsoft and NVIDIA on its own powered sites. It reports a $20.3B revenue backlog and more than 5 GW of secured power.",
      "hq": "AU",
      "tags": [
        "Public",
        "US",
        "Bitcoin miner to AI"
      ],
      "fields": {
        "listing": "public",
        "ticker": "NASDAQ: IREN",
        "parent_company": null,
        "clustermax_tier": "Underperforming",
        "contracted_mw": 5000,
        "active_mw": null,
        "power_note": "More than 5 GW of secured power; 0.8 GW of data center capacity targeted by end of calendar 2027 (Q1 FY27 results, 5 Nov 2026).",
        "backlog_usd_m": 20300,
        "backlog_note": "Revenue backlog of $20.3B, up 22% in the quarter; about two-thirds with investment-grade counterparties (Q1 FY27).",
        "revenue_q_usd_m": 228,
        "arr_usd_m": null,
        "revenue_note": "Q1 FY27 total revenue, quarter to 30 Sep 2026; AI Cloud $200M of it. A $4B+ ARR target for Jan-Feb 2027 is fully contracted.",
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": 5360,
        "capital_note": "$3.0B of convertible note proceeds (quarter to 30 Jun 2026) plus $2.36B of financing-facility proceeds (quarter to 30 Sep 2026), per the cash flow statement.",
        "customers": "Microsoft, NVIDIA",
        "rank": 6,
        "score": 50.2,
        "customer_count": 2
      },
      "sources": [
        {
          "url": "https://www.sec.gov/Archives/edgar/data/1878848/000187884826000065/irenreportsq1-fy27result.htm",
          "title": "IREN Reports Q1 FY27 Results (SEC exhibit)",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "ticker",
            "contracted_mw",
            "power_note",
            "backlog_usd_m",
            "backlog_note",
            "revenue_q_usd_m",
            "revenue_note",
            "capital_usd_m",
            "capital_note",
            "customers"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "firmus",
      "name": "Firmus",
      "url": "https://firmus.co",
      "summary": "Australian AI factory builder and GPU cloud running NVIDIA capacity in Australia and Singapore, with sites in development across Asia-Pacific. It has more than 900 MW contracted, with OpenAI as anchor customer in Malaysia.",
      "hq": "AU",
      "tags": [
        "Private",
        "Asia-Pacific"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Silver",
        "contracted_mw": 900,
        "active_mw": null,
        "power_note": "More than 900 MW contracted across all customers; two sites live (Australia, Singapore) and five in development (8 Sep 2026).",
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": 2000,
        "capital_note": "USD 2B strategic equity round at a post-money valuation above USD 10.5B, 7 Aug 2026 (Coatue, NVIDIA, Blackstone, Jane Street).",
        "customers": "OpenAI",
        "rank": 7,
        "score": 47.6,
        "customer_count": 1
      },
      "sources": [
        {
          "url": "https://firmus.co/newsroom/firmus-surpasses-900-mw-contracted-capacity-adds-openai-as-anchor-customer-and-expands-into-malaysia",
          "title": "Firmus surpasses 900 MW contracted capacity",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "contracted_mw",
            "power_note",
            "customers"
          ]
        },
        {
          "url": "https://firmus.co/newsroom/firmus-announces-fully-subscribed-usdusd2-billion-strategic-equity-investment-to-accelerate-nvidia-ai-factory-expansion-across-australia-and-asia-pacific",
          "title": "Firmus announces USD 2 billion strategic equity investment",
          "accessed": "2026-11-11",
          "fields": [
            "capital_usd_m",
            "capital_note"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "gmi-cloud",
      "name": "GMI Cloud",
      "url": "https://www.gmicloud.ai",
      "summary": "Mountain View GPU cloud with NVIDIA capacity in the US and Asia, selling H100, H200 and Blackwell instances plus an inference engine. It raised a $668M equity and credit package in September 2026.",
      "hq": "US",
      "founded": 2021,
      "tags": [
        "Private",
        "US",
        "Asia-Pacific"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Silver",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": 30000,
        "gpus_note": "30,000+ GPUs deployed, as stated on the current About page; no fleet as-of date supplied (read 11 Nov 2026).",
        "capital_usd_m": 668,
        "capital_note": "$223M Series B led by ARCHIV with NVIDIA participating, plus a $445M credit facility led by CTBC, 30 Sep 2026.",
        "customers": "Higgsfield, Nous Research, OpenRouter, Reflection, Cartesia, Trend Micro, Utopai Studios, Fireworks AI",
        "gpu_models": "NVIDIA H100, H200 and Blackwell (Blackwell shown as pre-order on the home page).",
        "regions": "US, Europe and Asia-Pacific (company regional footprint; GPU availability varies).",
        "rank": 8,
        "score": 45.6,
        "customer_count": 8
      },
      "sources": [
        {
          "url": "https://pulse2.com/gmi-cloud-raises-668-million-to-expand-global-ai-infrastructure/amp/",
          "title": "GMI Cloud raises $668 million (Pulse 2.0)",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "capital_usd_m",
            "capital_note",
            "customers"
          ]
        },
        {
          "url": "https://www.gmicloud.ai/",
          "title": "GMI Cloud",
          "accessed": "2026-11-11"
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "https://www.gmicloud.ai/en/company/about",
          "title": "GMI Cloud About: GPUs deployed and regional footprint",
          "accessed": "2026-11-11",
          "fields": [
            "gpus",
            "gpus_note",
            "regions"
          ]
        },
        {
          "url": "https://www.gmicloud.ai/en",
          "title": "GMI Cloud hardware availability",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "together-ai",
      "name": "Together AI",
      "url": "https://www.together.ai",
      "summary": "AI cloud that sells GPU clusters from 8 to 4,000+ GPUs alongside its inference and fine-tuning platform. It states a 2 GW+ US power portfolio and raised an $800M Series C in July 2026.",
      "tags": [
        "Private",
        "US",
        "Europe"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Bronze",
        "contracted_mw": 2000,
        "active_mw": null,
        "power_note": "2 GW+ in the US portfolio with 600 MW of near-term US capacity, plus 150 MW+ available in Europe (GPU clusters page, Nov 2026).",
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": "Reserved clusters up to 4,000+ GPUs; no fleet total published.",
        "capital_usd_m": 800,
        "capital_note": "$800M Series C at an $8.3B post-money valuation led by Aramco Ventures, 1 Jul 2026.",
        "customers": "Pika, Krea",
        "rank": 9,
        "score": 44.7,
        "customer_count": 2
      },
      "sources": [
        {
          "url": "https://www.together.ai/gpu-clusters",
          "title": "Together AI GPU clusters",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "contracted_mw",
            "power_note",
            "gpus_note",
            "customers"
          ]
        },
        {
          "url": "https://pulse2.com/together-ai-raises-800-million-series-c-at-8-3-billion-valuation/",
          "title": "Together AI raises $800 million Series C (Pulse 2.0)",
          "accessed": "2026-11-11",
          "fields": [
            "capital_usd_m",
            "capital_note"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "core42",
      "name": "Core42",
      "url": "https://www.core42.ai",
      "summary": "G42 subsidiary in Abu Dhabi running sovereign AI cloud capacity across nine data centers. It states 76K GPUs deployed or planned and more than 150 MW of deployments managed.",
      "hq": "AE",
      "tags": [
        "Subsidiary",
        "Middle East",
        "Sovereign"
      ],
      "fields": {
        "listing": "subsidiary",
        "ticker": null,
        "parent_company": "G42",
        "clustermax_tier": "Participation Ribbon",
        "contracted_mw": null,
        "active_mw": 150,
        "power_note": "150 MW+ of deployments managed; 9 global data centers (AI Cloud page, Nov 2026).",
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": 76000,
        "gpus_note": "76K GPUs deployed or planned (stats panel); the same page also says 86K+ GPUs are accessible across sovereign data centers.",
        "capital_usd_m": 550,
        "capital_note": "USD 550M in two HSBC structured trade finance facilities (USD 240M completed Feb 2026, USD 310M completed May 2026), announced 21 May 2026.",
        "customers": null,
        "rank": 10,
        "score": 37.7,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://www.core42.ai/products/ai-cloud",
          "title": "Core42 AI Cloud",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "parent_company",
            "active_mw",
            "power_note",
            "gpus",
            "gpus_note"
          ]
        },
        {
          "url": "https://www.core42.ai/resources/news/core42-raises-usd-550-million-from-hsbc-to-scale-global-ai-infrastructure?hsLang=en",
          "title": "Core42 raises USD 550 million from HSBC",
          "accessed": "2026-11-11",
          "fields": [
            "capital_usd_m",
            "capital_note"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "lambda",
      "name": "Lambda",
      "url": "https://lambda.com",
      "summary": "San Francisco GPU cloud selling NVIDIA clusters, on-demand instances and private cloud to AI labs and enterprises, with a multibillion-dollar Microsoft agreement. It has raised over $1.5B in equity and about $2.9B in debt.",
      "hq": "US",
      "founded": 2012,
      "tags": [
        "Private",
        "US"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Silver",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": "Builds distributed 5-75 MW sites and 75 MW+ campuses (investor page); no total MW published.",
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": 4426,
        "capital_note": "Over $1.5B Series E equity (Nov 2025) plus debt listed on the investor page: a $1B ($1.008B) senior secured fixed-rate delayed-draw term loan (1 Oct 2026), a $926M term loan B (27 Aug 2026) and a $1B senior secured credit facility (7 May 2026).",
        "customers": "Microsoft, Hudson River Trading",
        "rank": 11,
        "score": 36.3,
        "customer_count": 2
      },
      "sources": [
        {
          "url": "https://lambda.com/investors",
          "title": "Lambda investor information",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "power_note",
            "capital_usd_m",
            "capital_note",
            "customers"
          ]
        },
        {
          "url": "https://www.businesswire.com/news/home/20260520944576/en/Lambda-Partners-with-Hudson-River-Trading-to-Power-Quantitative-Research-and-Development",
          "title": "Lambda Partners with Hudson River Trading",
          "accessed": "2026-11-11",
          "fields": [
            "customers"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "verda",
      "name": "Verda",
      "url": "https://verda.com",
      "summary": "Helsinki-based GPU cloud, formerly DataCrunch, running NVIDIA GB300, B300, B200 and H200 capacity in Europe. It has raised over $450M.",
      "hq": "FI",
      "founded": 2020,
      "tags": [
        "Private",
        "Europe"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Bronze",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": 450,
        "capital_note": "Over $450M (EUR 390M) raised in total as of Sep 2026, including a $189M round led by Emergence Capital (22 Sep 2026).",
        "customers": null,
        "rank": 12,
        "score": 25.1,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://verda.com/newsroom",
          "title": "Verda newsroom",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "capital_usd_m",
            "capital_note"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "hyperbolic",
      "name": "Hyperbolic",
      "url": "https://hyperbolic.ai",
      "summary": "GPU cloud and inference platform renting on-demand and reserved H100, H200 and B200 capacity to 250,000+ AI builders. It raised a $30M Series A1 in October 2026.",
      "tags": [
        "Private",
        "US"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Participation Ribbon",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": 30,
        "capital_note": "$30M Series A1 led by Walden Catalyst, 27 Oct 2026.",
        "customers": "Modal, Higgsfield, Daytona, HeyGen, Pika, Voio, RadixArk",
        "rank": 13,
        "score": 20.5,
        "customer_count": 7
      },
      "sources": [
        {
          "url": "https://hyperbolic.ai/blog/a1-funding",
          "title": "Hyperbolic Series A1",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "capital_usd_m",
            "capital_note",
            "customers"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "lightning-ai",
      "name": "Lightning AI",
      "url": "https://lightning.ai",
      "summary": "AI development platform that merged with Voltage Park in 2026, combining its software with 36,000+ owned NVIDIA GPUs in seven US data centers. 450,000+ developers and AI teams use it.",
      "tags": [
        "Private",
        "US"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Unavailable",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": 36000,
        "gpus_note": "More than 36,000 NVIDIA GPUs across seven Tier 3+ US data centers (the Voltage Park fleet).",
        "capital_usd_m": 108,
        "capital_note": "$108M from venture investors including Index Ventures, Coatue, Bain Capital Ventures and NVIDIA (about page); Voltage Park merger terms not disclosed.",
        "customers": "NVIDIA, Cursor, Cisco, Runway, Supercell, Stability AI, Black Forest Labs",
        "rank": 14,
        "score": 20.4,
        "customer_count": 7
      },
      "sources": [
        {
          "url": "https://lightning.ai/about",
          "title": "About Lightning AI",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "gpus",
            "gpus_note",
            "capital_usd_m",
            "capital_note",
            "customers"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "latitude",
      "name": "Latitude.sh",
      "url": "https://www.latitude.sh",
      "summary": "Global bare-metal cloud offering dedicated GPU infrastructure and a managed Kubernetes service.",
      "tags": [
        "US",
        "Europe",
        "Asia-Pacific"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Participation Ribbon",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": 594,
        "capital_note": "Company homepage states $594M secured to build its AI inference cloud; funding composition and dates are not specified there.",
        "customers": null,
        "regions": "26 metro regions worldwide (cloud footprint, not a GPU availability guarantee).",
        "orchestration": "Bare metal and Latitude Kubernetes Service (LKS).",
        "rank": 15,
        "score": 19.6,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://www.latitude.sh",
          "title": "Latitude.sh product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "regions",
            "orchestration",
            "capital_usd_m",
            "capital_note"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "whitefiber",
      "name": "WhiteFiber",
      "url": "https://www.whitefiber.com",
      "summary": "Listed AI cloud and colocation provider reporting revenue obligations separately from future colocation lease payments. Its NC-1 contracted IT capacity is now online.",
      "tags": [
        "Public",
        "US"
      ],
      "fields": {
        "listing": "public",
        "ticker": "NASDAQ: WYFI",
        "parent_company": null,
        "clustermax_tier": "Underperforming",
        "contracted_mw": 40,
        "active_mw": 40,
        "power_note": "NC-1: all 40 MW of contracted IT capacity online and billing by the 10 Nov 2026 release; this is a site subtotal, not a group-wide power total. NC-2/NC-3 development potential excluded.",
        "backlog_usd_m": 729,
        "backlog_note": "ASC 606 RPO of $729.0M at 30 Sep 2026. Separate ASC 842 future lease payments of $421.5M are excluded; those payments also exclude NC-1’s remaining 20 MW because that lease had not commenced at quarter-end.",
        "revenue_q_usd_m": 24,
        "arr_usd_m": null,
        "revenue_note": "Q3 2026 group revenue $24.0M, including Cloud Services $18.1M. The approximately $365M annualized portfolio revenue is a mid-2027 deployment target, not current ARR.",
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": 310,
        "capital_note": "Completed $310M gross convertible notes due 2032 in Aug 2026; approximately $298.5M net proceeds. This is one offering, not cumulative funding. New $300M NC-1 facility excluded because initial funding was still conditional at release.",
        "customers": null,
        "rank": 16,
        "score": 19.3,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://www.prnewswire.com/news-releases/whitefiber-inc-reports-third-quarter-2026-results-302934441.html",
          "title": "WhiteFiber-issued Q3 2026 results, 10 Nov 2026 (PR Newswire)",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "ticker",
            "active_mw",
            "contracted_mw",
            "power_note",
            "backlog_usd_m",
            "backlog_note",
            "revenue_q_usd_m",
            "revenue_note",
            "capital_usd_m",
            "capital_note"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "digitalocean",
      "name": "DigitalOcean",
      "url": "https://www.digitalocean.com/products/gpu-droplets",
      "summary": "Developer cloud listed on the NYSE that sells NVIDIA and AMD GPU Droplets and bare-metal GPUs next to its general cloud. AI customer ARR reached $376M in Q3 2026, up 288% year on year.",
      "hq": "US",
      "founded": 2011,
      "tags": [
        "Public",
        "US",
        "Europe"
      ],
      "fields": {
        "listing": "public",
        "ticker": "NYSE: DOCN",
        "parent_company": null,
        "clustermax_tier": "Bronze",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": 321,
        "arr_usd_m": 376,
        "revenue_note": "Q3 2026 total company revenue $321M (whole company, not GPU-only); AI customer ARR $376M, up 288% year on year.",
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "rank": 17,
        "score": 16.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://investors.digitalocean.com/",
          "title": "DigitalOcean investor relations",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "ticker",
            "revenue_q_usd_m",
            "arr_usd_m",
            "revenue_note"
          ]
        },
        {
          "url": "https://investors.digitalocean.com/news/news-details/2026/DigitalOcean-Announces-Third-Quarter-2026-Financial-Results/default.aspx",
          "title": "DigitalOcean Announces Third Quarter 2026 Financial Results",
          "accessed": "2026-11-11"
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "gcore",
      "name": "Gcore",
      "url": "https://gcore.com",
      "summary": "GPU cloud offering NVIDIA virtual machines and bare-metal infrastructure, with container orchestration and high-speed networking.",
      "tags": [
        "Europe"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Bronze",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA L40S, A100, H100, H200, B300 and GB300 (B300/GB300 reservable).",
        "orchestration": "Virtual machines, bare metal, Docker and Kubernetes.",
        "interconnect": "InfiniBand available for supported GPU configurations.",
        "rank": 17,
        "score": 16.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://gcore.com/gpu-cloud",
          "title": "Gcore product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "orchestration",
            "interconnect"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "gmo-gpu-cloud",
      "name": "GMO GPU Cloud",
      "url": "https://gpucloud.gmo",
      "summary": "Japanese GPU cloud supplying managed HPC clusters and bare-metal NVIDIA infrastructure.",
      "tags": [
        "Asia-Pacific"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Bronze",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA HGX B300 and H200.",
        "regions": "Japan (domestic data centers).",
        "orchestration": "Managed HPC clusters with Slurm; bare-metal service.",
        "interconnect": "NVIDIA Spectrum-X Ethernet; DDN distributed storage.",
        "rank": 17,
        "score": 16.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://gpucloud.gmo",
          "title": "GMO GPU Cloud product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "regions",
            "orchestration",
            "interconnect"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "hyperstack",
      "name": "Hyperstack",
      "url": "https://www.hyperstack.cloud",
      "summary": "NexGen Cloud’s GPU-as-a-Service platform, launched in 2023, renting NVIDIA H100, H200, B200, B300 and GB300 capacity on renewable power. Head office in London.",
      "hq": "GB",
      "tags": [
        "Subsidiary",
        "Europe"
      ],
      "fields": {
        "listing": "subsidiary",
        "ticker": null,
        "parent_company": "NexGen Cloud",
        "clustermax_tier": "Bronze",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "rank": 17,
        "score": 16.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://www.hyperstack.cloud/about-us",
          "title": "About Hyperstack",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "parent_company"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "moonlite",
      "name": "Moonlite",
      "url": "https://moonlite.ai",
      "summary": "Managed GPU infrastructure provider supplying training clusters with an operations team responsible for monitoring and recovery.",
      "tags": [
        "US"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Bronze",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "H100 tested in ClusterMAX 3.0; report also describes initial B300 and GB300 handovers.",
        "orchestration": "Kubernetes and Slinky tested in ClusterMAX 3.0.",
        "rank": 17,
        "score": 16.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0 provider review",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "orchestration"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "prime-intellect",
      "name": "Prime Intellect",
      "url": "https://www.primeintellect.ai",
      "summary": "Integrated compute, training and inference platform offering GPU capacity with managed orchestration.",
      "tags": [
        "US"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Bronze",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA H100, H200 and B300 listed for compute.",
        "orchestration": "Slurm and Kubernetes for dedicated clusters.",
        "interconnect": "InfiniBand for distributed training.",
        "rank": 17,
        "score": 16.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://www.primeintellect.ai",
          "title": "Prime Intellect product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "orchestration",
            "interconnect"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "vast-ai",
      "name": "Vast.ai",
      "url": "https://vast.ai",
      "summary": "Los Angeles GPU marketplace where independent hosts and certified data centers rent out NVIDIA GPUs by the hour. It passed 17,000 GPUs across 350+ hosts in 2024.",
      "hq": "US",
      "founded": 2016,
      "tags": [
        "Private",
        "US",
        "Marketplace"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Participation Ribbon",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": 17000,
        "gpus_note": "17,000+ GPUs across 350+ hosts (2024 milestone on the About page; no newer count published).",
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "rank": 23,
        "score": 16.1,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://vast.ai/about",
          "title": "About Vast.ai",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "gpus",
            "gpus_note"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "runpod",
      "name": "Runpod",
      "url": "https://www.runpod.io",
      "summary": "Self-serve GPU cloud with on-demand pods and serverless GPUs for more than one million developers in 186 countries. It passed $325M ARR in October 2026.",
      "tags": [
        "Private",
        "US"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Participation Ribbon",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": 325,
        "revenue_note": "Surpassed $325M ARR (13 Oct 2026), up from $120M in January 2026.",
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": 100,
        "capital_note": "$100M round led by Summit Partners at a $1.0B valuation, 24 Jun 2026.",
        "customers": null,
        "rank": 24,
        "score": 16,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://www.runpod.io/press",
          "title": "Runpod press",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "arr_usd_m",
            "revenue_note",
            "capital_usd_m",
            "capital_note"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "neysa",
      "name": "Neysa",
      "url": "https://neysa.ai",
      "summary": "Indian AI infrastructure provider operating the Velocis platform for dedicated compute and managed AI services.",
      "tags": [
        "Asia-Pacific",
        "Sovereign"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Participation Ribbon",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": "Indian Institute of Science, Jocata",
        "regions": "India; platform positioned for Indian data residency.",
        "orchestration": "GPU-as-a-Service and managed AI PaaS via Velocis.",
        "rank": 25,
        "score": 12.5,
        "customer_count": 2
      },
      "sources": [
        {
          "url": "https://neysa.ai",
          "title": "Neysa product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "regions",
            "orchestration",
            "customers"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "ovhcloud",
      "name": "OVHcloud",
      "url": "https://www.ovhcloud.com",
      "summary": "French listed cloud and hosting group offering GPU instances alongside general-purpose infrastructure. Its power disclosure covers the whole group.",
      "tags": [
        "Public",
        "Europe"
      ],
      "fields": {
        "listing": "public",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Underperforming",
        "contracted_mw": null,
        "active_mw": 140,
        "power_note": "FY2026 release: 250 MW connected, of which 140 MW in operation across the group. Group-wide cloud/hosting capacity, not GPU-only MW; connected spare capacity is not labelled contracted.",
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": "FY2026 revenue reported in EUR; no USD quarterly conversion estimated.",
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA H200, H100, A100, L40S and L4 (Public Cloud catalog).",
        "rank": 25,
        "score": 12.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://corporate.ovhcloud.com/en/newsroom/news/financial-results-fy26/",
          "title": "OVHcloud FY2026 results, 22 Oct 2026",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "active_mw",
            "power_note",
            "revenue_note"
          ]
        },
        {
          "url": "https://www.ovhcloud.com/en/public-cloud/prices/",
          "title": "OVHcloud Public Cloud GPU catalog",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "massed-compute",
      "name": "Massed Compute",
      "url": "https://massedcompute.com",
      "summary": "US GPU cloud renting NVIDIA instances and bare-metal servers on demand and on contract. It secured up to $300M of equity and revenue-share financing from Digital Alpha in 2025.",
      "hq": "US",
      "founded": 2021,
      "tags": [
        "Private",
        "US"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Underperforming",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": 300,
        "capital_note": "Up to $300M equity and revenue-share financing from Digital Alpha, Aug 2025 (as cited in ClusterMAX 3.0).",
        "customers": null,
        "rank": 27,
        "score": 12.2,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier",
            "listing",
            "capital_usd_m",
            "capital_note"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "shadeform",
      "name": "Shadeform",
      "url": "https://www.shadeform.ai",
      "summary": "GPU marketplace aggregating independent clouds through one provisioning console and billing relationship.",
      "tags": [
        "Marketplace"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Participation Ribbon",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": "Fal",
        "regions": "North America, Europe, Asia and Australia through partner clouds.",
        "capacity_note": "Access across 30+ clouds; this is aggregated marketplace supply, not an owned fleet.",
        "rank": 28,
        "score": 11.5,
        "customer_count": 1
      },
      "sources": [
        {
          "url": "https://www.shadeform.ai",
          "title": "Shadeform product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "regions",
            "capacity_note",
            "customers"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "stn",
      "name": "STN",
      "url": "https://www.stninc.com",
      "summary": "GPU One provides single-tenant reserved GPU clusters and private AI infrastructure operated for enterprise customers.",
      "tags": [
        "US"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Participation Ribbon",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": "Skild AI",
        "gpu_models": "B200, B300 and GB300 listed; Vera Rubin also advertised.",
        "orchestration": "Kubernetes on single-tenant reserved GPU clusters.",
        "rank": 28,
        "score": 11.5,
        "customer_count": 1
      },
      "sources": [
        {
          "url": "https://www.stninc.com",
          "title": "STN product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "orchestration",
            "customers"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "scaleway",
      "name": "Scaleway",
      "url": "https://www.scaleway.com",
      "summary": "European sovereign cloud and AI provider within iliad Group, offering GPU instances and dedicated AI clusters.",
      "tags": [
        "Subsidiary",
        "Europe",
        "Sovereign"
      ],
      "fields": {
        "listing": "subsidiary",
        "ticker": null,
        "parent_company": "iliad Group",
        "clustermax_tier": "Unavailable",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": 5000,
        "gpus_note": "5,000+ GPUs available to researchers and companies (About page, read 11 Nov 2026); no separate deployment date stated.",
        "capital_usd_m": null,
        "capital_note": null,
        "customers": "Hugging Face, Golem, France TV, Aternos, Photoroom, Université Paris 8, Veepee, Turbo VPN, Qubit",
        "regions": "France, Poland, Italy and the Netherlands; Sweden and Germany are upcoming, excluded from this footprint.",
        "rank": 30,
        "score": 11,
        "customer_count": 9
      },
      "sources": [
        {
          "url": "https://www.scaleway.com/en/about-us/",
          "title": "Scaleway About: parent, operating countries and GPU availability",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "parent_company",
            "gpus",
            "gpus_note",
            "regions",
            "customers"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "fluidstack",
      "name": "Fluidstack",
      "url": "https://www.fluidstack.io",
      "summary": "GPU cloud and data center developer that builds dedicated clusters for frontier AI labs, including custom data centers for Anthropic in New York and Texas. It raised an $830M Series A at a $7.5B valuation in January 2026.",
      "hq": "US",
      "tags": [
        "Private",
        "US",
        "Europe"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Unavailable",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": "No current MW figure published; the company states a goal of hundreds of gigawatts of compute.",
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": 830,
        "capital_note": "$830M Series A at a $7.5B valuation, January 2026 (company blog).",
        "customers": "Anthropic",
        "rank": 31,
        "score": 10.8,
        "customer_count": 1
      },
      "sources": [
        {
          "url": "https://www.fluidstack.io/news/series-a-announcement",
          "title": "Fluidstack Series A announcement",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "power_note",
            "capital_usd_m",
            "capital_note",
            "customers"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "bitdeer-ai",
      "name": "Bitdeer AI",
      "url": "https://www.bitdeer.ai",
      "summary": "GPU cloud platform providing compute, distributed training and inference services alongside AI data center infrastructure.",
      "tags": [
        "Asia-Pacific"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Participation Ribbon",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA GB200 NVL72 and GB300 NVL72.",
        "orchestration": "Virtual machines, bare metal, container services and distributed training jobs.",
        "rank": 32,
        "score": 10.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://www.bitdeer.ai",
          "title": "Bitdeer AI product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "orchestration"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "buzz-hpc",
      "name": "BUZZ HPC",
      "url": "https://buzzhpc.ai",
      "summary": "Sovereign AI cloud offering GPU clusters, managed schedulers and bare-metal compute in Canada and the Nordics.",
      "tags": [
        "Canada",
        "Europe",
        "Sovereign"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Participation Ribbon",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA GB200 NVL72, B200, H200 and H100.",
        "regions": "Canada and the Nordics.",
        "orchestration": "Bare metal, managed Slurm and managed Kubernetes.",
        "interconnect": "NVIDIA Quantum InfiniBand and NVLink.",
        "rank": 32,
        "score": 10.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://buzzhpc.ai",
          "title": "BUZZ HPC product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "regions",
            "orchestration",
            "interconnect"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "radiant",
      "name": "Radiant",
      "url": "https://radiant.co",
      "summary": "AI infrastructure platform formed by merging Ori Industries with Brookfield’s AI infrastructure business; the Ori cloud now runs as Radiant AI Cloud. It plans up to 1 GW of UK compute at Drax.",
      "tags": [
        "Private",
        "Europe"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": "Brookfield-backed (merger of Ori Industries and Brookfield AI infrastructure)",
        "clustermax_tier": "Participation Ribbon",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": "Up to 1 GW of UK compute planned at Drax; builds AI factories from 10K to 100K+ GPUs (company site, Nov 2026).",
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "rank": 32,
        "score": 10.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://radiant.co/ori",
          "title": "Radiant and Ori",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "parent_company",
            "power_note"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "vessl",
      "name": "VESSL",
      "url": "https://vessl.ai",
      "summary": "AI cloud providing GPU workspaces, jobs and dedicated clusters for training and inference.",
      "tags": [
        "Asia-Pacific"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Participation Ribbon",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA B300, B200, GB300, H200, H100, A100 and L40S (newer models reservable).",
        "interconnect": "400 Gbps NDR InfiniBand on B200 and H200 multi-node offerings.",
        "orchestration": "Root access and dedicated clusters; managed Kubernetes and Slurm marked coming soon.",
        "rank": 32,
        "score": 10.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://vessl.ai",
          "title": "VESSL product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "interconnect",
            "orchestration"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "vultr",
      "name": "Vultr",
      "url": "https://www.vultr.com",
      "summary": "Independent cloud with 34 regions that rents NVIDIA and AMD Instinct GPUs next to general compute. It launched a 20 MW AMD MI355X supercluster in Chennai in October 2026.",
      "hq": "US",
      "founded": 2014,
      "tags": [
        "Private",
        "US",
        "Europe",
        "Asia-Pacific",
        "AMD"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Participation Ribbon",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": "Chennai supercluster starts at 20 MW of AMD Instinct MI355X, MI455X expansion planned for 2027; Chennai is the 34th region (21 Oct 2026).",
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "rank": 32,
        "score": 10.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://blogs.vultr.com/chennai-cloud-data-center-region",
          "title": "Vultr Chennai cloud data center region",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "power_note"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "sesterce",
      "name": "Sesterce",
      "url": "https://www.sesterce.com",
      "summary": "French AI infrastructure group developing GPU campuses in France and Finland, including a 1 GW campus at Bosquel with SoftBank. NVIDIA, Société Générale and BNP Paribas appear among its customers.",
      "tags": [
        "Private",
        "Europe",
        "Sovereign"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Unavailable",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": "In development: 1 GW Bosquel campus with SoftBank; 200 MW first phase at Jämsä, Finland (600 MW campus, EUR 10B+ planned investment); a confirmed 40 MW cap at Rovaltain. No operational MW stated.",
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": "NVIDIA, SoftBank, Société Générale, Crédit Agricole, BNP Paribas, F5, UMA Robotics, Patriarche, Flex AI, VAST Data",
        "rank": 37,
        "score": 10,
        "customer_count": 10
      },
      "sources": [
        {
          "url": "https://www.sesterce.com/",
          "title": "Sesterce",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "power_note",
            "customers"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "shakti-cloud",
      "name": "Shakti Cloud",
      "url": "https://shakticloud.ai",
      "summary": "Yotta Data Services’ sovereign GPU cloud in India, running 8,000+ NVIDIA H100s from Navi Mumbai and Greater Noida. Sarvam AI and IIT Madras train on it.",
      "hq": "IN",
      "tags": [
        "Subsidiary",
        "Asia-Pacific",
        "Sovereign"
      ],
      "fields": {
        "listing": "subsidiary",
        "ticker": null,
        "parent_company": "Yotta Data Services",
        "clustermax_tier": "Unavailable",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": 8000,
        "gpus_note": "8,000+ NVIDIA H100 GPUs, B200 clusters coming (Shakti Cloud site, Nov 2026). ClusterMAX lists the parent, Yotta.",
        "capital_usd_m": null,
        "capital_note": null,
        "customers": "Sarvam AI, IIT Madras, Fractal, RenderNet AI",
        "rank": 38,
        "score": 7.4,
        "customer_count": 4
      },
      "sources": [
        {
          "url": "https://shakticloud.ai/",
          "title": "Shakti Cloud",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "parent_company",
            "gpus",
            "gpus_note",
            "customers"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "chronoscale",
      "name": "ChronoScale",
      "url": "https://www.chronoscale.com",
      "summary": "Nasdaq-listed AI cloud spun out of Applied Digital in 2026, deploying NVIDIA GB300 NVL72 capacity for Microsoft and Baseten in North America. It targets a $1B annualized run-rate by calendar Q3 2027.",
      "hq": "US",
      "tags": [
        "Public",
        "US"
      ],
      "fields": {
        "listing": "public",
        "ticker": "NASDAQ: CHRN",
        "parent_company": "Applied Digital (NASDAQ: APLD), majority shareholder",
        "clustermax_tier": null,
        "contracted_mw": 54,
        "active_mw": null,
        "power_note": "50 MW GB300 NVL72 deployment for Microsoft plus about 4 MW for Baseten due calendar Q1 2027 (fiscal Q1 2027 results, 15 Oct 2026).",
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": 41.5,
        "arr_usd_m": null,
        "revenue_note": "Fiscal Q1 2027 revenue, quarter to 31 Aug 2026 (up 148%). Agreements support $1B annualized run-rate revenue by calendar Q3 2027 (company target).",
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": "Microsoft, Baseten",
        "rank": 39,
        "score": 7,
        "customer_count": 2
      },
      "sources": [
        {
          "url": "https://www.chronoscale.com/press-release/chronoscale-reports-fiscal-first-quarter-2027-results",
          "title": "ChronoScale Reports Fiscal First Quarter 2027 Results",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "ticker",
            "contracted_mw",
            "power_note",
            "revenue_q_usd_m",
            "revenue_note",
            "customers"
          ]
        },
        {
          "url": "https://www.chronoscale.com/press-release/chronoscale-announces-agreements-bringing-total-company-contracted-run-rate-revenue-to-1-billion-and-sale-of-ekso-bionics-business-unit",
          "title": "ChronoScale: contracted run-rate revenue to $1 billion",
          "accessed": "2026-11-11"
        },
        {
          "url": "https://letsdatascience.com/news/chronoscale-completes-spinout-begins-nasdaq-trading-0983ab20",
          "title": "ChronoScale completes spinout, begins Nasdaq trading",
          "accessed": "2026-11-09",
          "fields": [
            "parent_company"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "taiga-cloud",
      "name": "Taiga Cloud",
      "url": "https://northerndata.de/taiga-cloud-by-northern-data-group",
      "summary": "Northern Data Group’s GPU cloud, now part of Quake AI, with clusters in the UK, Portugal, Sweden and Norway. It states 24,000 GPUs and 970 PFlop/s of AI Linpack performance.",
      "hq": "DE",
      "tags": [
        "Subsidiary",
        "Europe"
      ],
      "fields": {
        "listing": "subsidiary",
        "ticker": null,
        "parent_company": "Northern Data Group, part of Quake AI",
        "clustermax_tier": null,
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": 24000,
        "gpus_note": "24,000 GPUs; 970 PFlop/s AI Linpack Rmax (company site, Nov 2026).",
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "rank": 40,
        "score": 6.6,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://northerndata.de/taiga-cloud-by-northern-data-group",
          "title": "Taiga Cloud by Northern Data Group",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "parent_company",
            "gpus",
            "gpus_note"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "hydra-host",
      "name": "Hydra Host",
      "url": "https://hydrahost.com",
      "summary": "Bare-metal GPU infrastructure provider with Brokkr provisioning and management across partner data centers.",
      "tags": [
        "US"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Underperforming",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA H200 and GH200 listed in the product catalog.",
        "interconnect": "InfiniBand for multi-node training clusters.",
        "orchestration": "Brokkr for bare-metal provisioning; interruptible, on-demand and reserved terms.",
        "capacity_note": "40+ data centers advertised; this is a facility footprint, not GPU inventory.",
        "rank": 41,
        "score": 4.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://hydrahost.com",
          "title": "Hydra Host product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "interconnect",
            "orchestration",
            "capacity_note"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "mithril",
      "name": "Mithril",
      "url": "https://mithril.ai",
      "summary": "GPU virtual-machine marketplace selling spot capacity and flexible reservations, with hosted Kubernetes control planes.",
      "tags": [
        "Marketplace"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Underperforming",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA H100 and H200.",
        "orchestration": "Hosted Kubernetes control plane; spot and reserved instances can join the same cluster.",
        "rank": 41,
        "score": 4.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://mithril.ai",
          "title": "Mithril product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "orchestration"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "palebluedot",
      "name": "PaleBlueDot AI",
      "url": "https://www.palebluedot.ai",
      "summary": "AI infrastructure platform sourcing reserved clusters and custom compute deployments across a partner network.",
      "tags": [
        "Marketplace"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Underperforming",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "regions": "60+ regions through its compute network.",
        "capacity_note": "220,000+ GPUs connected and 25+ supply partners; network access is not an owned or deployed fleet count.",
        "orchestration": "Reserved GPU clusters and custom token-factory deployments.",
        "rank": 41,
        "score": 4.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://www.palebluedot.ai",
          "title": "PaleBlueDot AI product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "regions",
            "capacity_note",
            "orchestration"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "sharon-ai",
      "name": "SHARON AI",
      "url": "https://sharonai.com",
      "summary": "Australian sovereign GPU cloud with dedicated clusters and flexible virtualized deployments from Melbourne infrastructure.",
      "tags": [
        "Asia-Pacific",
        "Sovereign"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Underperforming",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA H200, H100, L40S and A40; H200/B200 supercluster described separately.",
        "regions": "NEXTDC Tier IV facility, Melbourne, Australia.",
        "interconnect": "800 Gbps non-blocking InfiniBand; BlueField 200GbE fabric.",
        "orchestration": "Bare metal, virtualization, Docker and Kubernetes clusters.",
        "rank": 41,
        "score": 4.5,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://sharonai.com",
          "title": "SHARON AI product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "regions",
            "interconnect",
            "orchestration"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "tensordock",
      "name": "TensorDock",
      "url": "https://www.tensordock.com",
      "summary": "GPU marketplace supplying virtual machines through independent hosts. Partner availability is kept separate from an owned or contracted GPU fleet.",
      "tags": [
        "Marketplace"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": null,
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": "airgpu, ELBO, Creavite",
        "gpu_models": "NVIDIA H100 SXM5, A100 SXM4 and RTX 4090 (among the models listed).",
        "regions": "100+ partner locations in more than 20 countries.",
        "orchestration": "KVM virtual machines with dedicated GPU passthrough and root access.",
        "capacity_note": "Up to 30,000 GPUs available through partners; only hundreds described as immediately deployable. Partner supply is excluded from the deployed/owned/contracted GPU score.",
        "rank": 45,
        "score": 3,
        "customer_count": 3
      },
      "sources": [
        {
          "url": "https://www.tensordock.com/",
          "title": "TensorDock hardware, marketplace coverage and customer stories",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "regions",
            "orchestration",
            "capacity_note",
            "customers"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "boostrun",
      "name": "Boost Run",
      "url": "https://boostrun.com",
      "summary": "Listed AI infrastructure provider offering GPU servers, shared storage and managed Kubernetes through its own console and API.",
      "tags": [
        "Public",
        "US"
      ],
      "fields": {
        "listing": "public",
        "ticker": "NASDAQ: BRUN",
        "parent_company": null,
        "clustermax_tier": "Unavailable",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA B300, B200, H200, H100 and A100 (listed as available now).",
        "orchestration": "Managed Kubernetes, GPU/CPU compute and shared storage.",
        "rank": 46,
        "score": 0,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://boostrun.com",
          "title": "Boost Run product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "ticker",
            "gpu_models",
            "orchestration"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "cirrascale",
      "name": "Cirrascale",
      "url": "https://www.cirrascale.com",
      "summary": "US cloud provider offering dedicated multi-accelerator infrastructure for training and inference, with hardware and networking configurations published in its catalog.",
      "tags": [
        "US",
        "AMD"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Unavailable",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA GB300, B300, B200, RTX PRO 6000; AMD MI355X and MI350; Qualcomm AI 100 Ultra",
        "interconnect": "3,200 Gb/s listed for AMD MI355X/MI350; available for NVIDIA B300/B200 configurations.",
        "regions": "US-West, US-East and US-South.",
        "orchestration": "Dedicated multi-accelerator cloud servers; monthly billing and term blocks.",
        "rank": 46,
        "score": 0,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://www.cirrascale.com/pricing",
          "title": "Cirrascale server configurations and billing",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "interconnect",
            "orchestration"
          ]
        },
        {
          "url": "https://www.cirrascale.com/about",
          "title": "Cirrascale operating locations",
          "accessed": "2026-11-11",
          "fields": [
            "regions"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "cudo-compute",
      "name": "CUDO Compute",
      "url": "https://www.cudocompute.com",
      "summary": "AI infrastructure operator designing, commissioning and managing GPU clusters across partner facilities. Its pipeline targets and team delivery track record are recorded separately from current capacity.",
      "tags": [
        "Private",
        "Europe",
        "US"
      ],
      "fields": {
        "listing": "private",
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Unavailable",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": "Capacity pipeline: 250 MW+ contracted by end 2026 and 750 MW+ targeted by end 2027. Neither is a current contracted-MW disclosure at the 11 Nov cutoff, so both are excluded from scoring.",
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": "The 40,000+ GPUs statement describes the engineering team’s collective track record across partnerships and direct operations, not a verified current CUDO fleet.",
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA H100 and H200 deployed in the company case study; newer architectures advertised separately.",
        "interconnect": "InfiniBand and Ethernet cluster design.",
        "capacity_note": "Team delivery and operating experience across 40,000+ GPUs is excluded from the current fleet total.",
        "rank": 46,
        "score": 0,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://www.cudocompute.com/",
          "title": "CUDO Compute",
          "accessed": "2026-11-11",
          "fields": [
            "listing",
            "contracted_mw",
            "power_note",
            "gpus",
            "gpus_note",
            "gpu_models",
            "interconnect",
            "capacity_note"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "ClusterMAX 3.0 (SemiAnalysis, 23 Sep 2026)",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "denvr-dataworks",
      "name": "Denvr Dataworks",
      "url": "https://www.denvr.com",
      "summary": "Canadian AI platform offering GPU compute, inference and data services from its own facilities. Cluster sizing is disclosed separately from fleet size.",
      "tags": [
        "US",
        "Canada",
        "Sovereign"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Unavailable",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": "Site advertises clusters up to 1,024 GPUs. That is a cluster size limit, not a disclosed fleet total.",
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA H200, H100, A100; Intel Gaudi2",
        "regions": "Canada and the USA (owned and operated data centers).",
        "interconnect": "Up to 3,200 Gbps InfiniBand for H100 clusters.",
        "orchestration": "Bare metal, virtual machines and containers; self-service cluster provisioning.",
        "rank": 46,
        "score": 0,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://www.denvr.com/",
          "title": "Denvr AI platform and infrastructure",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "regions",
            "interconnect",
            "orchestration",
            "gpus_note"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "hot-aisle",
      "name": "Hot Aisle",
      "url": "https://hotaisle.xyz",
      "summary": "AMD-focused infrastructure provider offering isolated compute on Dell servers from the Switch Pyramid facility in Michigan.",
      "tags": [
        "US",
        "AMD"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": null,
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "AMD Instinct MI300X (eight GPUs per Dell PowerEdge XE9680 server).",
        "interconnect": "Eight 400G links, 3,200 Gbps RoCEv2 Ethernet per server.",
        "regions": "Switch Pyramid, Grand Rapids, Michigan.",
        "orchestration": "Isolated on-demand compute, provisioned through an automated platform; customer Kubernetes runs above the platform.",
        "rank": 46,
        "score": 0,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://hotaisle.xyz/compute",
          "title": "Hot Aisle supercomputer specifications",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "interconnect"
          ]
        },
        {
          "url": "https://hotaisle.xyz/datacenter",
          "title": "Hot Aisle data center location",
          "accessed": "2026-11-11",
          "fields": [
            "regions"
          ]
        },
        {
          "url": "https://hotaisle.xyz/",
          "title": "Hot Aisle provisioning model",
          "accessed": "2026-11-11",
          "fields": [
            "orchestration"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    },
    {
      "id": "qumulusai",
      "name": "QumulusAI",
      "url": "https://qumulusai.com",
      "summary": "Dedicated AI infrastructure provider designing and operating single-tenant GPU environments for customer-controlled workloads.",
      "tags": [
        "US"
      ],
      "fields": {
        "listing": null,
        "ticker": null,
        "parent_company": null,
        "clustermax_tier": "Unavailable",
        "contracted_mw": null,
        "active_mw": null,
        "power_note": null,
        "backlog_usd_m": null,
        "backlog_note": null,
        "revenue_q_usd_m": null,
        "arr_usd_m": null,
        "revenue_note": null,
        "gpus": null,
        "gpus_note": null,
        "capital_usd_m": null,
        "capital_note": null,
        "customers": null,
        "gpu_models": "NVIDIA B300, B200, RTX PRO 6000, H200, H100 SXM and H100 NVL.",
        "orchestration": "Managed Kubernetes and infrastructure lifecycle management around single-tenant compute.",
        "rank": 46,
        "score": 0,
        "customer_count": null
      },
      "sources": [
        {
          "url": "https://qumulusai.com",
          "title": "QumulusAI product and infrastructure disclosures",
          "accessed": "2026-11-11",
          "fields": [
            "gpu_models",
            "orchestration"
          ]
        },
        {
          "url": "https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard",
          "title": "SemiAnalysis ClusterMAX 3.0, 23 Sep 2026",
          "accessed": "2026-11-11",
          "fields": [
            "clustermax_tier"
          ]
        },
        {
          "url": "/methodology/",
          "title": "Computed rank, score and named-customer count",
          "accessed": "2026-11-11",
          "fields": [
            "rank",
            "score",
            "customer_count"
          ]
        }
      ],
      "verifiedAt": "2026-11-11",
      "featured": false
    }
  ]
}