AI data centers: which startup is ahead?
In our data center market deck, you will find everything you need to understand the market
SUMMARY
Crusoe is currently ahead in the AI data center startup race, with Nscale as the only challenger close enough to threaten that lead.
The decisive difference is delivered infrastructure. Crusoe already has more than 200 megawatts operating at Abilene for Oracle and OpenAI, while the largest rival projects are still mainly contracted, financed, or under construction.
Crusoe also pairs that live reference site with 4.9 gigawatts of contracted infrastructure. The combination matters more than either number alone because it shows both execution and a credible path beyond one flagship campus.
Nscale has the strongest chance of changing the ranking. Its Microsoft agreements cover roughly 200,000 GB300 GPUs, its later Rubin plans are huge, and its recent financing gives it enough firepower to build at hyperscale across several countries.
Nscale’s weakness is timing. A GPU commitment, a financed site, and an operating data center are three different stages, and much of Nscale’s case still depends on installations being energized and accepted on schedule.
Fluidstack has the most explosive single-customer opportunity through Anthropic and the strongest independent performance validation. Its position is held back by limited disclosure on completed megawatts, project economics, and dependence on one anchor customer.
Lambda leads the conventional buying experience. It has the most mature GPU cloud, transparent public pricing, no data-egress fees, and a large Microsoft relationship, but its disclosed physical campus footprint remains smaller than the bets being made by Crusoe, Nscale, and Fluidstack.
The market is easier to understand when megawatts, GPUs, contract values, and financing are kept separate. Adding them into one headline total would reward the companies that announce the most rather than the ones that have built the most.
Power is becoming the real moat. Crusoe’s ability to combine site development, on-site generation, batteries, cooling, hardware integration, and cloud software is harder to copy than a standard GPU rental product.
The ranking today is Crusoe first, Nscale second, Fluidstack third, and Lambda fourth. Nscale can close the gap through delivery, Fluidstack through completed Anthropic campuses and customer diversification, while Lambda needs a much larger disclosed physical footprint.
This market map, featured in our data center market deck, highlights top companies and startups in the data center market
Which AI data center startups belong in this race?
Crusoe, Nscale, Fluidstack, and Lambda are currently the four private companies with the strongest claim to being full AI data center startups.
All four now operate beyond basic GPU rental. They secure power or data center space, install dense computing systems, manage the software layer, and serve large AI customers. The comparison remains deliberately narrow: we want to know which startup can deliver a complete AI data center, rather than which company can simply offer access to a few hundred GPUs.
CoreWeave and Nebius provide useful public-market benchmarks, but neither still belongs in a startup ranking. Applied Digital, IREN, TeraWulf, Hut 8, and several former crypto miners are also public companies. Equinix, Digital Realty, Vantage, and QTS are established data center operators. Smaller GPU clouds such as TensorWave and Voltage Park remain credible, but their publicly documented physical infrastructure is still far below the scale now being attempted by the four companies below.
Funding needs a careful reading. Equity, corporate debt, GPU-backed loans, and project financing serve different purposes. We separate them wherever the disclosures allow it instead of adding every dollar into one flattering total.
| Startup | What the company does | Completed or clearly disclosed funding | Additional infrastructure financing |
|---|---|---|---|
| Crusoe | Develops large AI campuses, builds modular data centers, secures energy, and operates Crusoe Cloud | Approximately $3.9 billion across reported equity and corporate debt rounds after its $1.375 billion Series E | Large projects also use separate structures, including a previously announced $15 billion Abilene joint venture |
| Nscale | Develops sovereign and hyperscale AI campuses while operating a managed AI cloud | About $3.69 billion across its Series A, Series B, pre-Series C SAFE, and Series C rounds | A $1.4 billion GPU-backed loan, $790 million of Norway financing, and a recently closed $900 million revolving facility |
| Lambda | Operates a mature GPU cloud and is moving into dedicated AI factories | More than $2.3 billion of equity after its $1.5 billion-plus Series E | A $1 billion senior secured credit facility closed in 2026 |
| Fluidstack | Operates GPU clusters and develops custom AI infrastructure for frontier laboratories | The exact cumulative total remains unclear. A May 2026 SEC filing showed $842.5 million raised in one offering, while private databases place cumulative funding between roughly $1.05 billion and $1.78 billion | No complete company breakdown of corporate debt and project financing is publicly available |
Which AI data center startup is ahead today?
Crusoe is ahead in AI data centers today, and the lead is large enough to be meaningful.
Crusoe has already brought more than 200 megawatts of AI capacity online in Abilene for Oracle Cloud Infrastructure and OpenAI. The company also reports 4.9 gigawatts of contracted infrastructure across five American campuses and Crusoe Cloud. Those two figures give Crusoe something no private rival currently matches: a large operating reference site and a multi-campus contracted pipeline.
Nscale is now the closest challenger. Its agreements cover approximately 200,000 Nvidia GB300 GPUs for Microsoft, followed by major Rubin deployments planned in Norway and Portugal. Nscale could eventually reach or exceed Crusoe’s computing scale, but most of those systems still need to be installed, energized, and accepted by the customer.
Fluidstack has the boldest single-customer opportunity through Anthropic, while Lambda has the most mature conventional GPU cloud. Neither currently matches Crusoe across live megawatts, power development, campus construction, and customer diversification.
The structure is clear rather than a four-way tie. Crusoe leads overall, Nscale forms the second tier by itself, and Fluidstack and Lambda lead narrower parts of the market.
If you want more recent data on this point, please see our latest data center market report.
As this chart shows, and as featured in our data center market deck, search interest in data centers has increased significantly
Who has the most AI data center capacity running now?
Crusoe has the clearest lead in AI data center capacity that customers can use today.
The first two Abilene buildings provide more than 200 megawatts of IT capacity. Oracle is already using the site for OpenAI workloads, which gives us an unusually solid chain of evidence: completed buildings, an identified cloud tenant, an identified end customer, installed Blackwell systems, and live computing activity.
Nscale’s clearest operational reference is Glomfjord in Norway. Nscale and InfraPartners described 30 megawatts as operating, with an expansion toward 60 megawatts. On the directly comparable disclosed figure, Crusoe’s live site is already more than six times larger.
Fluidstack says it has more than 100,000 GPUs under management, but that figure includes systems hosted in third-party data centers. GPU management shows commercial and operational reach, although it tells us little about how much power Fluidstack has personally developed or how many complete campuses it has delivered.
Lambda operates production clusters across several locations, yet it does not publish one consolidated figure for live megawatts. The company has announced 30-plus megawatts with EdgeConneX across Chicago and Atlanta and a 24-megawatt opening phase in Kansas City, but announced capacity cannot be treated as operating capacity without a confirmed launch.
| Startup | Best disclosed operating evidence | Comparable live-capacity index | What remains unclear |
|---|---|---|---|
| Crusoe | More than 200 MW operating at the first Abilene phase | 100 | How much of the remaining contracted portfolio is energized |
| Nscale | 30 MW operating at Glomfjord, expanding toward 60 MW | At least 15 | Whether the expansion has reached full operation |
| Fluidstack | More than 100,000 GPUs under management | Not comparable | Owned power, live campus capacity, and the split between direct and partner infrastructure |
| Lambda | Several production cloud regions, without a consolidated MW total | Not comparable | Total energized power across owned and partner sites |
Who has the biggest AI data center pipeline?
Crusoe leads on contracted power, while Nscale leads on the number of next-generation GPUs publicly attached to customer agreements.
Crusoe reported 4.9 gigawatts of contracted AI infrastructure and a development pipeline above 40 gigawatts. We give far more weight to the first number. The 4.9-gigawatt total covers projects backed by customer commitments, whereas the wider pipeline also includes sites under negotiation or advanced development.
Nscale’s strongest comparable figure is approximately 200,000 GB300 GPUs contracted for Microsoft. The plan includes around 104,000 GPUs in Texas and an initial 12,600-GPU deployment in Portugal, alongside other locations. Nscale later announced more than 30,000 Rubin GPUs in Norway and more than 66,000 in Portugal. Adding every announcement together would be misleading because some newer Rubin plans may expand or update earlier hardware commitments.
Fluidstack’s Anthropic relationship is enormous in dollar terms, but Anthropic’s $50 billion refers to a broad American computing-infrastructure program. Public disclosures do not tell us how much of that spending becomes Fluidstack revenue, how many megawatts Fluidstack must deliver, or what minimum payments Anthropic has guaranteed.
Lambda’s Microsoft agreement covers tens of thousands of Nvidia GPUs and was described as multibillion-dollar and multiyear. That clearly places Lambda in the hyperscale market, although the private contract gives us too little detail to compare its exact size with Nscale or Crusoe.
| Startup | Strongest pipeline evidence | How we interpret it |
|---|---|---|
| Crusoe | 4.9 GW contracted across data centers and Crusoe Cloud | The strongest directly disclosed measure of customer-backed power capacity |
| Nscale | About 200,000 GB300 GPUs for Microsoft, with later Rubin deployments announced | Potentially comparable computing scale, but most delivery remains ahead |
| Fluidstack | Selected for Anthropic’s $50 billion American infrastructure program | Exceptional upside, with too few disclosed commercial and capacity details |
| Lambda | Multibillion-dollar Microsoft agreement covering tens of thousands of GPUs | A major contract, but exact capacity and timing remain confidential |
This chart, featured in our data center market deck, illustrates yearly venture capital funding for data center startups
Which startup has the strongest AI data center customers?
Crusoe has the strongest customer book because major buyers have already trusted the company with separate, large-scale projects.
The operating Abilene phase connects Crusoe with Oracle and OpenAI. Microsoft then selected Crusoe for an additional 900-megawatt campus with its own power plant. The Microsoft project sits next to the Oracle and OpenAI development, but it represents a separate customer commitment rather than another brand attached to the same contract.
Nscale has built an unusually deep relationship with Microsoft. The customer returned across Texas, Portugal, Norway, and the UK, which carries more weight than a single pilot. Nscale also participates in Stargate Norway with OpenAI and Aker. The customer quality is excellent, although Microsoft currently accounts for a large share of the publicly disclosed demand.
Fluidstack’s selection by Anthropic is equally serious. Anthropic confirmed that Fluidstack would design and build custom facilities in Texas and New York, with more sites expected. Fluidstack has also served Mistral AI, Poolside, Character.AI, Black Forest Labs, and other advanced AI developers. The concern comes from concentration: Anthropic now dominates Fluidstack’s physical infrastructure story.
Lambda has worked with Microsoft for years and now has a multibillion-dollar infrastructure agreement with the company. Lambda also serves a wider base of researchers, startups, enterprises, and AI laboratories through its cloud. Crusoe still ranks first because its largest relationships already span both an operating campus and a second major development.
If you want more recent data on this point, please see our latest data center market report.
Which AI data center startup is growing fastest now?
Nscale is growing fastest on announced scale, while Crusoe is growing fastest on infrastructure that has already passed a real delivery test.
Nscale raised $155 million in its Series A at the end of 2024, followed by a $1.1 billion Series B, $433 million of SAFE financing, and a $2 billion Series C. The Series C valued Nscale at $14.6 billion. The company has since added $790 million for Norway and a $900 million revolving credit facility, giving Nscale far more financial reach than it had one year earlier.
Crusoe’s expansion has followed a different pattern. In March 2025, Crusoe reported more than 1.6 gigawatts operating or under construction. The company now reports 4.9 gigawatts contracted, and its latest campus announcement adds another one-gigawatt site in Childress, Texas. The definitions are slightly different, but the evidence still shows a rapid move from one flagship development toward a network of large campuses.
Fluidstack has made the sharpest strategic jump. The company moved from operating clusters in partner facilities to leading Anthropic’s custom American buildout. An amended SEC Form D also showed $842.5 million raised toward an $850 million offering. Fluidstack’s growth is obvious, although public information still gives us little detail on delivered megawatts or revenue.
Lambda’s growth is steadier. The company secured a $1 billion credit facility, signed a larger Microsoft agreement, and continues adding dedicated facilities and newer Nvidia systems. Nscale currently wins the speed contest, but Crusoe’s growth carries less delivery risk.
This chart, featured in our data center market deck, shows how Equinix is capturing share in data centers
Who can build and power AI data centers fastest?
Crusoe has the best evidence that it can build and energize a large AI data center quickly.
Crusoe constructed and energized two 100-megawatt buildings at Abilene in under one year. Delivering 200 megawatts at that speed required the company to coordinate the shell, electrical systems, cooling, networking, substations, backup power, and GPU installation rather than merely placing servers inside an available colocation hall.
The company has also built power procurement into its operating model. Recent agreements include roughly 750 megawatts of Bergen Engines generation, 12 gigawatt-hours of Form Energy storage, a 900-megawatt on-site plant for Microsoft, and expanded battery systems using Redwood Materials technology. Crusoe can therefore move ahead with grid power, on-site generation, batteries, or a combination of the three.
Nscale’s best power strategy currently sits in northern Europe. Glomfjord uses local hydropower, while the 230-megawatt Narvik campus is being built around Norway’s renewable electricity supply. Nscale’s latest partnership with Nordkraft adds local operating expertise, although the largest Microsoft and OpenAI deployments still need to prove their schedules in practice.
Fluidstack says its integrated system can deliver gigawatts in six months, compared with an industry timeline of 18 to 24 months. We treat that as a company target for now. Fluidstack has shown that it can provision more than 2,500 GPUs for Poolside within 48 hours, but rapid cluster setup remains a different job from building a complete gigawatt campus.
Which startup has the most mature AI data center product?
Lambda has the most mature AI cloud product, while Crusoe has the most mature product for customers that need the whole data center delivered.
Lambda gives customers several clear ways to buy computing capacity. A small team can launch on-demand GPUs, a growing company can reserve a cluster, and a frontier laboratory can commission a single-tenant supercluster. Lambda advertises production-ready clusters from 16 to more than 2,000 GPUs and dedicated architectures that can scale from 4,000 to more than 165,000 GPUs.
Crusoe covers more of the physical stack. The company can find or generate power, develop the site, manufacture modular infrastructure, install the GPU systems, and operate the workload through Crusoe Cloud. Crusoe has lately expanded the software side with managed inference, serverless fine-tuning, self-serve deployments, an operations command center, and modular edge zones.
Fluidstack’s product strength lies in managing large, dedicated clusters across infrastructure owned by different partners. Its managed Kubernetes and Slurm services include cluster-health monitoring and bare-metal orchestration. Nscale also offers GPU nodes, inference, fine-tuning, and sovereign AI factories, but many of its largest configurations remain tied to campuses scheduled for later delivery.
For cloud access today, Lambda is the easiest winner. For a customer asking one company to deliver the buildings, power, hardware, and cloud layer, Crusoe is further ahead.
This chart, featured in our data center market deck, illustrates yearly funding for data center startups
Which AI data center startup performs best?
Fluidstack has the strongest independent performance evidence today, although no startup has proved a universal speed advantage.
SemiAnalysis gave Fluidstack a Gold rating in its ClusterMAX 2.0 review after evaluating the company’s operating platform. The review covered around 80 providers and described Fluidstack as especially good at turning mixed third-party data center infrastructure into a dependable customer experience. That is useful validation because networking, cluster health, scheduling, and support often decide whether expensive GPUs remain productive.
Crusoe’s strongest evidence comes from operating high-density GB200 systems at Abilene and supporting direct-to-chip liquid cooling. Lambda provides immediately accessible B200 and H100 clusters and supports large single-tenant systems. Nscale has secured substantial allocations of GB300 and Rubin hardware, giving the company a strong future specification sheet.
These examples measure different things. SemiAnalysis tested operational quality, Crusoe demonstrated facility density, Lambda shows product availability, and Nscale emphasizes incoming hardware generations. We still lack a neutral benchmark comparing training throughput, failure rates, network congestion, utilization, and cost across equivalent clusters.
Fluidstack wins the narrow performance question. That result alone is not enough to crown it the overall data center leader.
If you want more recent data on this point, please see our latest data center market report.
Which AI data center startup offers the best prices?
Lambda currently offers the clearest value for customers buying standard GPU cloud capacity.
Lambda publishes prices without requiring an initial sales call. Its current on-demand rates start at $3.99 per GPU-hour for an eight-GPU H100 system and $6.69 for an eight-GPU B200 system. Smaller configurations cost slightly more per GPU, and Lambda does not charge data-egress fees. A customer can therefore estimate a real budget before beginning procurement.
Crusoe lists on-demand, reserved, and spot options for Nvidia and AMD systems, but much of the exact pricing still goes through sales. Crusoe has advertised spot capacity at discounts of up to 90% against hyperscaler on-demand rates. That can be attractive for interruptible jobs, although a spot discount should never be compared directly with guaranteed reserved capacity.
Nscale and Fluidstack negotiate their largest deployments privately. Those contracts may offer better economics than public cloud pricing, particularly when cheap Nordic electricity, dedicated capacity, or long commitments are involved. We cannot verify the claim without contract prices, utilization guarantees, power costs, and financing terms.
The answer changes for a custom campus. Crusoe may eventually deliver the lowest total cost through integrated power and construction, but no startup publishes enough customer-level information to prove that today. Lambda wins here because its economics are visible, comparable, and immediately usable.
This chart, featured in our data center market deck, compares the main business model options for hyperscale data center operators
Which startup can scale AI data centers across the most countries?
Nscale has the broadest international AI data center pipeline, although Crusoe has delivered more at hyperscale.
Nscale has operating or announced infrastructure in Norway, Portugal, the UK, Finland, Iceland, Texas, West Virginia, and other planned markets. The European portfolio is particularly valuable because governments and large companies increasingly want computing capacity, data storage, and operational control to remain inside their own jurisdictions.
Crusoe remains more concentrated in the United States. Its American footprint now extends beyond the original Abilene project, with major sites in Texas, Wyoming, and other disclosed locations. Crusoe also operates cloud capacity in Iceland and is using modular systems to pursue smaller sovereign and edge deployments.
Fluidstack already manages clusters through third-party facilities in several regions, including Nordic deployments with Borealis. That partner-led approach can expand quickly because Fluidstack does not need to own every building. Lambda follows a similar route through companies such as EdgeConneX and Cologix, while gradually taking more responsibility for dedicated facilities.
Nscale wins on geographic coverage. Crusoe remains ahead in proving that one company can repeat a very large physical development model.
What can Crusoe do that rivals cannot easily copy?
Crusoe has the strongest moat because competitors would need to copy several difficult businesses at the same time.
A conventional GPU cloud can buy Nvidia systems and lease data center space. A traditional developer can build a powered shell. Crusoe combines energy procurement, on-site generation, campus development, modular manufacturing, liquid cooling, hardware integration, and cloud software. Few startups currently control that entire route from electricity to a running AI workload.
Crusoe’s energy background is especially useful now. Power availability has become one of the main limits on AI expansion, and Crusoe spent years learning how to place computing infrastructure near underused or stranded energy. The company can now apply that experience to natural gas, grid connections, batteries, wind-rich regions, and future nuclear projects.
Nscale’s moat comes from sovereign positioning, access to financing, European power relationships, and repeated Microsoft commitments. Fluidstack’s operating software gives the company a way to make inconsistent partner infrastructure behave like one reliable cloud. Lambda benefits from a mature developer product, transparent pricing, and more than a decade of GPU experience.
Those advantages should attract customers. Crusoe’s advantage reaches further into the physical bottlenecks that currently decide whether a project gets built at all.
If you want more recent data on this point, please see our latest data center market report.
This chart, featured in our data center market deck, shows the revenue mix across customer segments in the data center market
Which AI data center startup has the strongest financing position?
Crusoe has the strongest financing model for building campuses, while Nscale has raised the most aggressive recent pool of corporate capital.
Crusoe combines venture equity, corporate credit, project debt, joint ventures, and long-term customer demand. That structure lets outside capital finance individual campuses instead of forcing Crusoe’s shareholders to fund every building and GPU directly. The operating Abilene phase also gives lenders and investors a completed project against which to judge future developments.
Nscale has accumulated about $3.69 billion in equity and SAFE financing, followed by more than $3 billion in disclosed loans, project financing, and revolving credit. The amount is remarkable for a company created in its present form only recently. The open question concerns conversion: Nscale must turn that financing into completed Microsoft and OpenAI capacity across several countries.
Lambda’s $2.3 billion-plus equity base and $1 billion credit facility look appropriate for a company expanding from cloud operations into larger dedicated facilities. Fluidstack has also raised substantial capital, but conflicting database totals and limited project-financing disclosure make the balance sheet harder to evaluate from outside.
Crusoe gets the edge because its capital has already produced a large working asset. Nscale may have greater available firepower now, but the spending program remains several steps ahead of the completed infrastructure.
What could knock Crusoe out of first place?
Nscale can overtake Crusoe if its huge Microsoft and OpenAI deployments arrive on schedule, while Fluidstack needs to prove that the Anthropic buildout can survive permitting and customer-concentration risks.
Nscale’s route to first place is clear. The company needs to energize the Texas, Portugal, Norway, and UK projects, show reliable operation at that scale, and convert preliminary sites into firm customer-backed capacity. A 2026 Guardian investigation found that Nscale’s highly promoted Loughton site remained at an early stage, illustrating the distance that can exist between an investment announcement and a working data center.
Fluidstack faces a different problem. Anthropic gives Fluidstack access to extraordinary demand, but the relationship also concentrates a large part of Fluidstack’s future around one customer. New York has recently paused environmental permits for new hyperscale data centers for one year. Public information does not yet establish whether Fluidstack’s New York project has every permit needed to avoid delays, so we treat the effect as unresolved rather than assuming the site is blocked.
Crusoe still carries serious risk. Its contracted portfolio requires several campuses and power systems to move forward at the same time. The company is also relying heavily on gas generation for some developments, which can trigger environmental opposition, fuel-price exposure, and permitting disputes. Microsoft replaced an abandoned OpenAI expansion at Abilene, showing that even major AI customers can change where they place future capacity.
Lambda’s main threat comes from competition. GPU clouds are becoming crowded, and Microsoft, Google, Amazon, Oracle, CoreWeave, Nebius, and newer providers can all pressure pricing. Lambda’s mature cloud makes the company resilient, but the physical data center ranking will remain difficult to win without much larger disclosed campuses.
This chart, featured in our data center market deck, shows how hyperscale AI-ready campus technology has evolved over time
Which AI data center startups are actually ahead?
Crusoe is currently the clear AI data center leader, Nscale is the only close challenger, and Fluidstack and Lambda win narrower parts of the market.
Crusoe takes first place because it has already completed the hardest step: a large AI campus is operating for serious customers. Crusoe then added contracted power, new campuses, dedicated generation, storage, manufacturing, and cloud software around that reference project. The lead covers more than fundraising or promised GPUs.
Nscale deserves second place rather than a shared first. The company has raised enough capital, signed enough demand, and secured enough hardware to become enormous. Delivery will decide the argument. Should Nscale bring its largest Microsoft and OpenAI projects online without major delays, the gap with Crusoe could close quickly.
Fluidstack ranks third because Anthropic has handed the company one of the largest opportunities in the market, and independent testing supports the quality of Fluidstack’s operating platform. Fluidstack would move higher once it discloses completed campus capacity, firmer project economics, and progress across more than one anchor customer.
Lambda ranks fourth in this specific comparison, even though Lambda currently offers the strongest standard cloud-buying experience. The company has real customers, transparent pricing, a mature product, and a large Microsoft contract. Crusoe, Nscale, and Fluidstack are simply making bigger bets on developing complete physical campuses.
| Rank | Startup | Why the startup holds this position | What could change the rank |
|---|---|---|---|
| 1 | Crusoe | The strongest mix of operating hyperscale capacity, contracted power, construction speed, energy development, customer quality, and full-stack delivery | Delays across its expanding campus portfolio, power opposition, or weaker capital discipline |
| 2 | Nscale | The largest challenger by announced GPU deployments, recent financing, international reach, and repeated Microsoft demand | Successful delivery across Texas and Europe could move Nscale into first place |
| 3 | Fluidstack | A major Anthropic mandate, more than 100,000 GPUs under management, and the best independent cloud-operations rating among the four | Completed Anthropic campuses and a second large infrastructure customer could push Fluidstack above Nscale |
| 4 | Lambda | The most mature GPU cloud, the clearest pricing, and a multibillion-dollar Microsoft relationship | Lambda needs to disclose and deliver a much larger physical data center footprint |
If you want more recent data on this point, please see our latest data center market report.
OUR METHODOLOGY
This analysis compares Crusoe, Nscale, Fluidstack, and Lambda across the dimensions that best reveal whether an AI data center startup is genuinely ahead: operating capacity, contracted pipeline, customer strength, delivery speed, power strategy, financing, product maturity, performance, pricing, and international reach.
We kept megawatts, GPU commitments, capital raised, contract values, and GPUs under management separate because they measure different things. A financed campus is not an operating campus, and a broad customer spending program is not automatically revenue committed to one infrastructure provider.
The ranking gives the greatest weight to completed infrastructure, identified customers, firm commitments, and demonstrated delivery. Announced developments and wider pipelines still count, but they receive less weight until the capacity is installed, energized, and accepted by the customer.
We first reached a conclusion for each analytical dimension and then combined those findings into the final ranking. This stops one spectacular funding round, customer announcement, or hardware allocation from deciding the entire answer.
Source selection prioritized direct company announcements, customer confirmations, regulatory filings, independent technical assessments, and tier-1 reporting. We used company claims for disclosed capacity, financing, products, and plans, then looked for outside confirmation whenever the difference between announced and operating infrastructure affected the ranking.
Key sources include Crusoe’s disclosure of 4.9 GW of contracted infrastructure, Crusoe’s initial 200 MW Abilene development, Associated Press reporting on Abilene, Microsoft, OpenAI, and Oracle, Nscale’s Microsoft GPU agreements, Nscale’s Series C disclosure, and the Financial Times on Nscale’s financing and execution risks.
We also used Anthropic’s confirmation of its Fluidstack infrastructure program, Fluidstack’s SEC Form D filing, Lambda’s Microsoft agreement, and Microsoft’s GB300 cluster confirmation for customer, financing, and hardware context.
In our data center market deck, we identify pain points entrepreneurs should prioritize
Related blog posts
- AI cloud: which startup is ahead?
- Who has the edge in modular data centers?
- Which startup is redefining how data centers are powered?
