The Financialization of Compute contains 25 companies across
4 categories.
Compute is becoming financialized in three big ways: It is being priced like a commodity, traded like capacity and financed like infrastructure. Basically compute stops being only a technical resource and becomes a priced, tradable, financeable, optimizable asset.
Compute capacity & supply
Companies that own, operate, finance, or aggregate GPU capacity. In practice, this is mostly the neocloud layer. An important aspect is that neocloud does not always mean they own the data centers themselves. Some lease long-term capacity from datacenter partners, some build their own sites, and many do both.
CoreWeave, Inc.
Website: coreweave.com
CoreWeave is an AI-native cloud platform that provides GPU compute, storage, networking, and infrastructure control for training and inference workloads at scale.
Category: Compute capacity & supply.
Founded: 2017.
Country: United States.
Employees: 1,001-5,000 employees.
Landscape fit: It fits the category because it owns and operates GPU-heavy cloud infrastructure and sells access to that compute capacity for AI workloads, effectively aggregating and commercializing GPU supply.
Funding rounds
-
Debt financing
announced 2026-05-01
: 3100000000 USD
; investors: Morgan Stanley, Mitsubishi UFJ Financial Group
; source: CoreWeave Investor Relations
; https://investors.coreweave.com/news/news-details/2026/CoreWeave-Closes-3-1-Billion-Loan-Facility-Expanding-Access-to-Public-Market-for-GPU-Backed-Financing/default.aspx
-
Debt financing
announced 2026-03-01
: 8500000000 USD
; investors: Blackstone Credit & Insurance, MUFG, Morgan Stanley, Goldman Sachs, JPMorgan
; source: CoreWeave Investor Relations
; https://investors.coreweave.com/news/news-details/2026/CoreWeave-Closes-Landmark-8-5-Billion-Financing-Facility-Achieving-First-Investment-Grade-Rated-GPU-backed-Financing/default.aspx
-
Private Placement
announced 2026-01-01
: 2000000000 USD
; investors: NVIDIA
; source: CoreWeave investor relations results release
; https://investors.coreweave.com/news/news-details/2026/CoreWeave-Reports-Strong-First-Quarter-2026-Results/
-
Convertible senior notes
announced 2025-12-01
: 2250000000 USD
; investors: Institutional purchasers
; source: CoreWeave Investor Relations
; https://investors.coreweave.com/news/news-details/2025/CoreWeave-Prices-Upsized-2-25-Billion-Convertible-Senior-Notes-Offering/default.aspx
-
Debt financing
announced 2025-07-01
: 2600000000 USD
; investors: Morgan Stanley, Mitsubishi UFJ Financial Group
; source: CoreWeave Investor Relations
; https://investors.coreweave.com/news/news-details/2025/CoreWeave-Closes-2-6-Billion-Secured-Debt-Financing-Facility-Strengthening-Market-Position-as-AI-Cloud-Leader/default.aspx
Company timeline
-
2026-06-01:
Active hiring slate visible on CoreWeave jobs page
(hr)
. CoreWeave's LinkedIn jobs page shows current openings in sales, logistics, finance, and operations, including sales manager, logistics, accounts payable, and strategic finance roles.. Team size observed: 1,001-5,000 employees employees.
Source: LinkedIn Jobs.
https://www.linkedin.com/company/coreweave/jobs
-
2026-05-28:
Unified agentic AI capabilities launched
(product)
. CoreWeave launched unified agentic AI capabilities connecting training, inference, observability, and RL so agents can improve continuously in production.
Source: CoreWeave Newsroom.
https://www.coreweave.com/news/coreweave-closes-the-training-to-inference-gap-for-autonomous-agent-improvement
-
2026-05-28:
CoreWeave ARENA and agentic AI capabilities signal productized demand generation
(gtm)
. CoreWeave introduced ARENA and later unified agentic AI capabilities to let customers test workloads and improve agents continuously across training, inference, observability, and RL.
Source: CoreWeave press release / landing page.
https://www.coreweave.com/news/coreweave-closes-the-training-to-inference-gap-for-autonomous-agent-improvement
-
2026-05-14:
Sandboxes launched for RL, tool use, and evaluation
(product)
. CoreWeave launched Sandboxes as secure, isolated environments for reinforcement learning, agent tool use, and model evaluation.
Source: CoreWeave Investor Relations.
https://investors.coreweave.com/news/news-details/2026/CoreWeave-Sandboxes-Launches-to-Accelerate-Reinforcement-Learning-Agent-Tool-Use-and-Model-Evaluation/default.aspx
-
2026-05-01:
Debt financing
(funding)
. Amount: 3100000000 USD. Investors: Morgan Stanley, Mitsubishi UFJ Financial Group
Source: CoreWeave Investor Relations.
https://investors.coreweave.com/news/news-details/2026/CoreWeave-Closes-3-1-Billion-Loan-Facility-Expanding-Access-to-Public-Market-for-GPU-Backed-Financing/default.aspx
-
2026-04-30:
SUNK Self-Service and SUNK Anywhere expanded
(product)
. CoreWeave expanded SUNK with guided self-service and SUNK Anywhere to speed cluster setup and support multi-cloud environments.
Source: CoreWeave Newsroom.
https://coreweave.com/news/coreweave-sunk-expands-capabilities-to-bring-ai-workloads-online-faster-anywhere
Lambda
Website: lambda.ai
Lambda is an AI infrastructure company that builds and operates GPU cloud, supercomputers, and AI factory infrastructure for training and inference workloads.
Category: Compute capacity & supply.
Funding stage: Series D+.
Founded: 2012.
Country: United States.
Employees: 754 employees.
Landscape fit: It fits the compute capacity & supply category because it owns and operates large-scale GPU infrastructure, including AI factories and GPU cloud capacity, and finances continued expansion of that capacity through major equity and debt raises.
Funding rounds
-
Credit Facility
announced 2026-05-01
: 1000000000 USD
; investors: J.P. Morgan, Citigroup, MUFG Bank, Crédit Agricole
; source: Lambda
; https://lambda.ai/blog/lambda-closes-a-1-billion-senior-secured-credit-facility
-
Series E
announced 2025-11-01
: 1500000000 USD
; investors: TWG Global, US Innovative Technology Fund
; source: Lambda
; https://lambda.ai/blog/lambda-raises-over-1-5b-from-twg-global-usit-to-build-superintelligence-cloud-infrastructure
-
Credit Facility
announced 2025-08-01
: 275000000 USD
; investors: J.P. Morgan, Citigroup, MUFG Bank, Crédit Agricole
; source: Business Wire via Yahoo Finance
; https://finance.yahoo.com/news/lambda-closes-275m-senior-secured-130000830.html
-
Series D
announced 2025-02-01
: 480000000 USD
; investors: Andra Capital, SGW, Andrej Karpathy, ARK Invest, Fincadia Advisors, G Squared, In-Q-Tel, KHK & Partners, NVIDIA, Pegatron, Supermicro, Wistron, Wiwynn, 1517, Crescent Cove, USIT
; source: Lambda
; https://lambda.ai/blog/lambda-raises-480m-to-expand-ai-cloud-platform
-
Series C
announced 2024-02-01
: 320000000 USD
; investors: US Innovative Technology Fund, B Capital, SK Telecom, T. Rowe Price Associates, Inc., Crescent Cove, Mercato Partners, 1517 Fund, Bloomberg Beta, Gradient Ventures
; source: Lambda
; https://lambda.ai/blog/lambda-raises-320m-to-build-a-gpu-cloud-for-ai
Company timeline
-
2026-06-02:
Workspaces add account-level access boundaries
(product)
. Lambda launched free workspaces to organize resources and separate dev, staging, and production in shared accounts.
Source: Introducing workspaces for Lambda Cloud.
https://lambda.ai/blog/introducing-workspaces-for-lambda-cloud
-
2026-06-01:
Open roles page shows broad active hiring
(hr)
. Lambda’s careers page currently lists 38 open roles across data center business, G&A, and sales/marketing, including engineering, operations, finance, and intern positions.. Team size observed: 754 employees employees.
Source: Lambda careers page.
https://lambda.ai/careers/open-roles
-
2026-05-21:
Bare Metal Instances launch
(product)
. Lambda launched Bare Metal Instances with direct hardware access and API-driven lifecycle operations on Superclusters.
Source: Lambda Bare Metal Instances: full hardware control with API-driven operations.
https://lambda.ai/blog/lambda-bare-metal-instances
-
2026-05-05:
Leadership team expanded with CEO/CTO/chairman changes
(hr)
. Lambda announced Michel Combes as CEO, Stephen Balaban moving full-time to CTO, and John Donovan as Chairman of the Board.. Team size observed: 754 employees employees.
Source: Lambda blog.
https://lambda.ai/blog/lambda-assembles-leadership-team-to-power-gigawatt-scale-ai-infrastructure
-
2026-05-01:
Credit Facility
(funding)
. Amount: 1000000000 USD. Investors: J.P. Morgan, Citigroup, MUFG Bank, Crédit Agricole
Source: Lambda.
https://lambda.ai/blog/lambda-closes-a-1-billion-senior-secured-credit-facility
-
2026-03-16:
Expands into bare metal and rack-scale superclusters
(gtm)
. At GTC 2026, Lambda announced NVIDIA Vera CPUs, new Bare Metal Instances, and production-scale GB300 NVL72 superclusters, extending its offer toward rack-scale and agentic-AI infrastructure. ([lambda.ai](https://lambda.ai/blog/lambda-at-gtc-2026-building-the-superintelligence-cloud?utm_source=openai))
Source: Lambda blog / GTC page.
https://lambda.ai/blog/lambda-at-gtc-2026-building-the-superintelligence-cloud
Nebius Group N.V.
Website: nebius.com
Nebius is an AI cloud and AI infrastructure company that provides full-stack compute for the AI lifecycle, including large-scale GPU clusters, cloud platform services, training, tuning, inference, and deployment. ([nebius.com](https://nebius.com/company))
Category: Compute capacity & supply.
Founded: 2024.
Country: Netherlands.
Employees: 1,001-5,000 employees.
Landscape fit: It fits the compute capacity & supply category because Nebius owns and operates GPU infrastructure and AI cloud capacity, including large-scale GPU clusters and data centers that it uses to sell compute to AI builders and enterprises. ([nebius.com](https://nebius.com/company))
Funding rounds
-
convertible note offering
announced 2026-03-01
: 4000000000 USD
; source: Nebius Group announces pricing of upsized private offering of $4.0 billion of convertible senior notes
; https://nebius.com/newsroom/nebius-group-announces-pricing-of-upsized-private-offering-of-4-0-billion-of-convertible-senior-notes
-
strategic investment / private placement
announced 2026-03-01
: 2000000000 USD
; investors: NVIDIA
; source: NVIDIA and Nebius partner to scale full-stack AI cloud
; https://nebius.com/newsroom/nvidia-and-nebius-partner-to-scale-full-stack-ai-cloud
-
Convertible senior notes
announced 2025-09-01
: 3160000000 USD
; source: Nebius press release
; https://nebius.com/newsroom/nebius-group-announces-closings-of-its-public-offering-of-class-a-ordinary-shares-and-concurrent-private-offering-of-convertible-senior-notes-with-aggregate-gross-proceeds-to-date-of-approximately-4-2-billion
-
convertible note offering
announced 2025-09-01
: 2750000000 USD
; source: Nebius Group announces pricing of upsized private offering of $2.75 billion of convertible senior notes
; https://nebius.com/newsroom/nebius-group-announces-pricing-of-upsized-private-offering-of-2-75-billion-of-convertible-senior-notes
-
Public offering
announced 2025-09-01
: 1000000000 USD
; source: Nebius press release
; https://nebius.com/newsroom/nebius-group-announces-closings-of-its-public-offering-of-class-a-ordinary-shares-and-concurrent-private-offering-of-convertible-senior-notes-with-aggregate-gross-proceeds-to-date-of-approximately-4-2-billion
Company timeline
-
2026-06-01:
Current jobs page shows active hiring
(hr)
. Nebius LinkedIn jobs page shows multiple current openings, including data center, security, sales, and solutions architecture roles.. Team size observed: 1,001-5,000 employees employees.
Source: Nebius jobs on LinkedIn.
https://www.linkedin.com/company/nebius/jobs
-
2026-06-01:
Careers page highlights open roles and Early Talent Program
(hr)
. Nebius careers page is actively recruiting and promotes an Early Talent Program alongside open positions.. Team size observed: 1,001-5,000 employees employees.
Source: Nebius careers page.
https://nebius.com/careers
-
2026-05-20:
Bloom Energy power partnership
(gtm)
. Nebius partnered with Bloom Energy to power its AI infrastructure build-out, supporting its US expansion with modular fuel-cell power for the customer-facing cloud footprint.
Source: Nebius newsroom.
https://nebius.com/newsroom/nebius-and-bloom-energy-partner-to-power-ai-infrastructure-build-out
-
2026-03-26:
Serverless AI platform update
(gtm)
. Nebius AI Cloud 3.5 added serverless AI, making compute more frictionless for prototyping, experimentation, and production inference while reducing operational overhead.
Source: Nebius newsroom.
https://nebius.com/newsroom/nebius-ai-cloud-3-5-introduces-serverless-ai-to-give-developers-frictionless-compute-for-real-world-ai
-
2026-03-26:
Aether 3.5 serverless AI release
(product)
. Nebius AI Cloud 3.5 added serverless AI, the NVIDIA RTX PRO 6000 Blackwell Server Edition, a new Data Transfer Service, redesigned marketplace workflows, and updated Kubernetes and Soperator controls. ([nebius.com](https://nebius.com/newsroom/nebius-ai-cloud-3-5-introduces-serverless-ai-to-give-developers-frictionless-compute-for-real-world-ai))
Source: Nebius newsroom.
https://nebius.com/newsroom/nebius-ai-cloud-3-5-introduces-serverless-ai-to-give-developers-frictionless-compute-for-real-world-ai
-
2026-03-16:
Physical AI cloud collaboration with NVIDIA
(gtm)
. Nebius announced an end-to-end cloud for robotics and physical AI, integrating NVIDIA’s Physical AI Data Factory Blueprint and targeting developers building autonomous systems.
Source: Nebius newsroom.
https://nebius.com/newsroom/nebius-teams-with-nvidia-to-build-cloud-for-robotics-and-physical-ai
Fluidstack
Website: fluidstack.io
Fluidstack is an AI cloud platform that builds and operates high-performance GPU clusters and data-center infrastructure for AI model training and inference. ([fluidstack.io](https://www.fluidstack.io/about-us/about))
Category: Compute capacity & supply.
Founded: 2017.
Country: United Kingdom.
Landscape fit: It fits the compute capacity & supply category because the company acquires power, designs/builds data centers, and operates large GPU clusters, i.e. it directly owns/operates and aggregates compute capacity for AI workloads. ([fluidstack.io](https://fluidstack.io/))
Funding rounds
-
Equity financing
announced 2026-01-01
: 450000000 USD
; source: Reuters via TradingView
; https://www.tradingview.com/news/reuters.com%2C2026%3Anewsml_FWN3YO0OF%3A0-fluidstack-raises-450-million-in-equity-financing/
-
Series A
announced 2025-02-01
: 200000000 USD
; investors: Cacti, Episode 1 Ventures, Mana Ventures, 7 Global Capital, InnoWhale Ventures, Armyn Capital, Thesiger Capital
; source: Sacra
; https://sacra.com/research/fluidstack/
-
Debt financing
announced 2024-01-01
: 37500000 USD
; source: Sacra
; https://sacra.com/research/fluidstack/
-
SAFE
announced 2024-01-01
: 24700000 USD
; source: Sacra
; https://sacra.com/research/fluidstack/
-
Seed
announced 2019-03-01
: 3000000 USD
; investors: Seedcamp, Mercuri, 7 Global Capital
; source: LegalClarity
; https://legalclarity.org/who-owns-fluidstack-investors-and-ownership-structure/
Company timeline
-
2026-06-09:
Broad hiring cluster visible on careers page
(hr)
. Fluidstack's careers and job pages currently show open roles spanning supply chain, infrastructure deployment, security, finance, strategic transactions, product, leasing, and brand/content.
Source: Fluidstack careers and job pages.
https://careers.fluidstack.io/
-
2026-01-01:
Equity financing
(funding)
. Amount: 450000000 USD
Source: Reuters via TradingView.
https://www.tradingview.com/news/reuters.com%2C2026%3Anewsml_FWN3YO0OF%3A0-fluidstack-raises-450-million-in-equity-financing/
-
2025-12-04:
New York HQ relocation reinforced U.S.-centered commercial posture
(gtm)
. Fluidstack announced its global headquarters move to New York City, tying the relocation to scaling American infrastructure and the Anthropic-related buildout.
Source: Fluidstack blog.
https://www.fluidstack.io/about-us/blog/fluidstack-expands-u-s-investment-by-relocating-global-headquarters-to-new-york-city
-
2025-11-12:
Anthropic deal moved Fluidstack into bespoke U.S. data-center delivery
(gtm)
. Fluidstack announced a $50 billion infrastructure deal with Anthropic to build custom data centers in New York and Texas, with more sites planned.
Source: Fluidstack blog.
https://www.fluidstack.io/about-us/blog/fluidstack-selected-by-anthropic-to-deliver-custom-data-centers-in-the-us
-
2025-11-12:
Anthropic custom data-center deployment announced
(product)
. Fluidstack announced a $50 billion deal with Anthropic to build custom data centers in New York and Texas, with more sites to come. ([fluidstack.io](https://www.fluidstack.io/about-us/blog/fluidstack-selected-by-anthropic-to-deliver-custom-data-centers-in-the-us))
Source: Fluidstack blog.
https://www.fluidstack.io/about-us/blog/fluidstack-selected-by-anthropic-to-deliver-custom-data-centers-in-the-us
-
2025-11-06:
ClusterMAX Gold validated Fluidstack’s managed-cluster quality
(gtm)
. SemiAnalysis awarded Fluidstack Gold in ClusterMAX, giving the company third-party validation for its GPU cloud and forward-deployed engineering model.
Source: Fluidstack blog.
https://www.fluidstack.io/resources/blog/fluidstack-awarded-clustermax-gold-by-semianalysis
RunPod
Website: runpod.io
Runpod is an AI developer cloud that provides on-demand GPU compute, serverless inference, and infrastructure for building, training, and scaling AI workloads.
Category: Compute capacity & supply.
Funding stage: Seed.
Founded: 2022.
Country: United States.
Employees: 51-200 employees.
Landscape fit: It fits this category because it owns/aggregates GPU capacity and sells access to cloud GPUs and serverless compute for AI workloads, directly monetizing compute supply.
Funding rounds
-
Seed
announced 2024-05-01
: 20000000 USD
; investors: Intel Capital, Dell Technologies Capital, Julien Chaumond, Nat Friedman, Adam Lewis
; source: Runpod's $20MM Milestone: Fueling Our Vision, Empowering Our Team
; https://www.runpod.io/blog/runpod-raises-20mm
Company timeline
-
2026-06-09:
Current careers page shows 20 open roles
(hr)
. Runpod's Greenhouse careers page lists 20 current openings across Engineering, Infrastructure, Marketing, People, Product, Sales, and Customer Support.. Team size observed: 51-200 employees employees.
Source: Greenhouse.
https://job-boards.greenhouse.io/runpod
-
2026-06-01:
Marketing team hiring push
(hr)
. Runpod publicly highlighted multiple marketing roles, including senior PLG growth marketing, developer relations, technical content QA, and partnerships leadership.. Team size observed: 51-200 employees employees.
Source: LinkedIn.
https://www.linkedin.com/posts/danstotts_we-are-hiring-across-the-runpod-marketing-activity-7459697813787086849-fkVx
-
2026-05-22:
Multi-Instance GPUs roll out on Serverless
(product)
. Runpod began rolling out MIG on RTX 6000 Pro cards, allowing Serverless workloads to use isolated 24 GB GPU slices more efficiently. ([runpod.io](https://www.runpod.io/blog/multi-instance-gpu-on-runpod))
Source: Runpod Blog.
https://www.runpod.io/blog/multi-instance-gpu-on-runpod
-
2026-04-30:
Flash reaches GA
(product)
. Runpod promoted Flash to GA as a production-ready way to run serverless GPU and CPU workloads in pure Python without Docker. ([runpod.io](https://www.runpod.io/blog/flash-is-ga?utm_source=openai))
Source: Runpod Blog.
https://www.runpod.io/blog/flash-is-ga
-
2026-04-01:
Technical Program Manager opening announced
(hr)
. A Runpod-linked post said the company was hiring a Technical Program Manager to support engineering operations.. Team size observed: 51-200 employees employees.
Source: LinkedIn.
https://www.linkedin.com/posts/pjain1596_were-hiring-at-runpod-were-looking-for-activity-7438322541196177408-49EQ
-
2026-03-31:
Runpod Assistant launches in the console
(product)
. Runpod shipped a built-in Assistant that lets users manage pods, endpoints, and infrastructure through natural language in the console. ([runpod.io](https://www.runpod.io/blog/introducing-the-runpod-assistant-manage-your-cloud-gpu-resources-with-natural-language/))
Source: Runpod Blog.
https://www.runpod.io/blog/introducing-the-runpod-assistant-manage-your-cloud-gpu-resources-with-natural-language
SPAN
Website: span.io
SPAN develops smart electrical panels and related energy-management infrastructure, and its XFRA initiative turns underutilized residential power capacity into distributed AI inference compute.
Category: Compute capacity & supply.
Funding stage: Series C.
Founded: 2018.
Country: United States.
Employees: 201-500 employees.
Landscape fit: It fits this category because XFRA is explicitly positioned as a distributed data center / compute network that aggregates GPU capacity across many small installations to supply AI compute demand.
Funding rounds
-
Series C
announced 2026-01-01
: 163000000 USD
; investors: Kyber Knight Capital, undisclosed investors
; source: CB Insights
; https://www.cbinsights.com/company/spanio/financials
-
Series B2
announced 2023-04-01
: 96500000 USD
; investors: Wellington Management, Congruent Ventures, Capricorn Investment Group, Qualcomm Ventures, Fifth Wall, Munich Re Ventures, A/O PropTech, Amazon Alexa Fund
; source: SPAN
; https://www.span.io/blog/b2-funding
-
Series B
announced 2022-03-01
: 90000000 USD
; investors: Fifth Wall, Wellington Management, Angeleno Group, FootPrint Coalition, Obsidian Investment Partners, A/O PropTech, previous board investors
; source: Business Wire
; https://www.businesswire.com/news/home/20220322005900/en/Home-Electrification-Innovator-SPAN-Raises-%2490-Million-in-Series-B-Funding-Round-to-Accelerate-Next-Generation-of-EV-Charging-and-Electrification-Products
-
Series A
announced 2021-01-01
: 20000000 USD
; investors: Munich Re Ventures, Amazon Alexa Fund, existing investors
; source: PR Newswire
; https://www.prnewswire.com/news-releases/span-accelerates-with-20m-in-funding-and-amazon-alexa-integration-301214948.html
-
Series A
announced 2020-05-01
: 10200000 USD
; investors: ArcTern Ventures, Capricorn Investment Group, Incite Ventures, Wells Fargo Strategic Capital, Wireframe Ventures, Ulu Ventures, Energy Foundry, Hardware Club, 1/0 Capital
; source: PR Newswire
; https://www.prnewswire.com/news-releases/span-raises-10-2-million-for-reinvention-of-home-electrical-panel-to-optimize-solar-energy-storage-and-electric-vehicle-charging-301057255.html
Company timeline
-
2026-06-01:
LinkedIn jobs page shows a multi-function hiring cluster
(hr)
. SPAN’s LinkedIn jobs page shows recently posted roles spanning territory sales, staff software engineering, embedded security, hardware design, utility channel marketing, and HR internship positions.. Team size observed: 201-500 employees employees.
Source: SPAN Jobs on LinkedIn.
https://www.linkedin.com/company/spanio/jobs
-
2026-06-01:
Careers page shows open positions across global offices
(hr)
. SPAN’s mission/careers page states that the company has open positions across its global offices and invites applicants to apply now.. Team size observed: 201-500 employees employees.
Source: SPAN Mission & Careers.
https://www.span.io/mission-careers
-
2026-04-13:
XFRA extends SPAN into distributed AI inference compute
(gtm)
. SPAN launched XFRA, a distributed data center solution that uses residential and small commercial grid capacity for AI inference and other compute workloads.
Source: SPAN Blog.
https://www.span.io/blog/span-announces-xfra-a-distributed-data-center-solution-to-close-the-speed-to-power-gap-for-ai-compute-demand
-
2026-04-13:
Launched XFRA distributed data center solution
(product)
. SPAN launched XFRA, a distributed data center solution that uses residential and small commercial spaces to deliver compute capacity for AI workloads.
Source: SPAN Blog.
https://www.span.io/blog/span-announces-xfra-a-distributed-data-center-solution-to-close-the-speed-to-power-gap-for-ai-compute-demand
-
2026-03-10:
Industry coalition campaign reframes brand around grid utilization
(gtm)
. SPAN joined Carrier, Tesla, and others to launch Utilize, a campaign focused on lowering electricity costs by using underused grid capacity.
Source: SPAN Blog.
https://www.span.io/blog/span-carrier-tesla-and-others-launch-campaign-to-lower-electricity-costs-by-unlocking-underused-grid-capacity
-
2026-03-09:
Eaton partnership expands distribution and co-development
(gtm)
. SPAN announced a strategic partnership with Eaton, including investment, co-developed products, and broader distribution through Eaton’s channels.
Source: SPAN Blog.
https://www.span.io/blog/announcing-our-collaboration-with-eaton
PaleBlueDot AI
Website: palebluedot.ai
PaleBlueDot AI is an AI infrastructure company that provides a token-routing platform plus reserved GPU clusters and a marketplace for sourcing AI compute. ([palebluedot.ai](https://www.palebluedot.ai/?utm_source=openai))
Category: Compute capacity & supply.
Funding stage: Series B.
Founded: 2024.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits the compute capacity & supply category because its core products are GPU clusters, reserved capacity, and a marketplace that aggregates and routes AI compute supply across a global network. ([palebluedot.ai](https://www.palebluedot.ai/?utm_source=openai))
Funding rounds
-
Series B
announced 2026-01-01
: 150000000 USD
; investors: B Capital
; source: PR Newswire
; https://www.prnewswire.com/news-releases/palebluedot-ai-raises-150m-series-b-to-scale-global-ai-compute-infrastructure-302672782.html
Company timeline
-
2026-05-01:
Hiring cluster across supply, growth, legal, and solutions roles
(hr)
. LinkedIn job posts showed multiple open roles at PaleBlueDot AI, including Global Sourcing and Supply (Compute/GPU), User Growth & Community Growth (To B), Senior Legal Counsel Manager (AI Computing/SaaS), and Solutions Architect (GPU/Neocloud/Data Center).. Team size observed: 11-50 employees employees.
Source: LinkedIn.
https://www.linkedin.com/company/palebluedot-ai
-
2026-04-21:
Launched PBD TokenRouter and premium credit program
(gtm)
. PaleBlueDot AI launched PBD TokenRouter as a unified platform for AI model access, and introduced a monthly premium credit program plus hackathons and partner-led events to drive adoption.
Source: PaleBlueDot AI Newsroom / PR Newswire.
https://www.palebluedot.ai/newsRoom
-
2026-04-21:
TokenRouter launches as a unified AI model access platform
(product)
. PaleBlueDot AI launched PBD TokenRouter as a single API and control layer for accessing and managing multiple AI models, with smart routing, failover, cost governance, smart caching, and a premium token credit program.
Source: PaleBlueDot AI / PR Newswire.
https://www.prnewswire.com/news-releases/palebluedot-ai-launches-pbd-tokenrouter-a-unified-platform-for-accessing-ai-models-302749017.html
-
2026-01-23:
CEO appointment announced
(hr)
. PaleBlueDot AI announced Stephen Watts as CEO; the post says he had joined the company two years earlier as VP of Go-to-Market.. Team size observed: 11-50 employees employees.
Source: PaleBlueDot AI Newsroom.
https://www.palebluedot.ai/newsRoom/stephen-watts-appointed-ceo
-
2026-01-23:
Stephen Watts appointed CEO to lead customer-first GTM
(gtm)
. PaleBlueDot AI appointed Stephen Watts as CEO after he had served as VP of Go-to-Market, with the company saying he would focus on customer-first go-to-market execution and global expansion.
Source: PaleBlueDot AI Newsroom.
https://www.palebluedot.ai/newsRoom/stephen-watts-appointed-ceo
-
2026-01-01:
Series B
(funding)
. Amount: 150000000 USD. Investors: B Capital
Source: PR Newswire.
https://www.prnewswire.com/news-releases/palebluedot-ai-raises-150m-series-b-to-scale-global-ai-compute-infrastructure-302672782.html
Hyperbolic
Website: hyperbolic.ai
Hyperbolic is an open-access AI cloud platform that provides on-demand GPU clusters, serverless inference, and reserved compute for developers and AI teams.
Category: Compute capacity & supply.
Funding stage: Series A.
Founded: 2022.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits the compute capacity & supply category because it owns and aggregates GPU capacity through an on-demand GPU marketplace and reserved clusters, letting customers rent compute and suppliers monetize idle GPUs.
Funding rounds
-
series-a
announced 2024-12-01
: 12000000 USD
; investors: Variant Fund, Polychain Capital, Chapter One, Lightspeed Faction, Bankless Ventures, IOSG, Vertex, GSR, Wintermute Ventures, Blockchain Builders Fund, Alumni Ventures, Ambush
; source: Hyperbolic official blog
; https://hyperbolic.xyz/blog/hyperbolic-secures-20m-total-funding-with-latest-series-a
-
seed
announced 2024-07-01
: 7000000 USD
; investors: Polychain Capital, Lightspeed Faction, Chapter One, LongHash, Bankless Ventures, Republic Digital, Nomad Capital, CoinSummer Labs, Third Earth Capital, Balaji Srinivasan, Illia Polosukhin, Sandeep Nailwal, Casey Caruso, Tekin Salimi, Santiago Santos
; source: Hyperbolic official blog
; https://www.hyperbolic.ai/blog/hyperbolic-secures-7-million-seed-funding-round
-
pre-seed
announced 2022-11-01
: 725000 USD
; investors: Chapter One, Modular Capital, Samsung Next, Blizzard Fund by Avalanche, IMO Ventures, SNZ Holding, Ausvic Capital
; source: Hyperbolic official blog
; https://www.hyperbolic.ai/blog/hyperbolic-secures-7-million-seed-funding-round
Company timeline
-
2026-05-01:
Broad live hiring slate appears on Ashby careers page
(hr)
. Hyperbolic's Ashby jobs page shows active openings including Head of Talent, Head of Operations, Head of Supply, Head of Finance, Senior Platform Engineer, Senior Data Analytics Engineer, Forward Deployed Infrastructure Engineer, Member of Technical Staff - Full Stack, Head of Marketing, Technical Writer, and Business Operations Specialist.. Team size observed: 11-50 employees employees.
Source: Ashby.
https://jobs.ashbyhq.com/hyperbolic
-
2026-04-01:
Founder announces rapid hiring ramp and careers page expansion
(hr)
. Jasper Zhang said Hyperbolic was ramping up hiring quickly, planning 10+ roles across engineering, infra, supply, finance, and go-to-market, and looking to bring on a full-time recruiter; he also noted the careers page was adding more roles soon.. Team size observed: 11-50 employees employees.
Source: LinkedIn.
https://www.linkedin.com/posts/yuezhang95_hyperbolic-labs-jobs-activity-7446304995106467841-XgvA
-
2026-02-25:
Introduced short-term GPU reservations
(gtm)
. Hyperbolic launched GPU Reservations with 7-, 14-, and 30-day terms, guaranteed capacity, and dedicated support for long-running workloads.
Source: Hyperbolic blog.
https://www.hyperbolic.ai/blog/introducing-on-demand-gpu-reservations-up-to-30-days
-
2026-02-25:
GPU Reservations up to 30 days launched
(product)
. Hyperbolic added short-term GPU reservations with 7, 14, and 30-day options to guarantee capacity for longer workloads.
Source: Hyperbolic blog.
https://www.hyperbolic.ai/blog/Introducing-On-Demand-GPU-Reservations-Up-to-30-Days
-
2026-01-21:
Launched dedicated model hosting for production inference
(gtm)
. Hyperbolic introduced single-tenant dedicated model hosting with reserved GPU capacity, isolated networking, and private customer endpoints.
Source: Hyperbolic blog.
https://www.hyperbolic.ai/blog/dedicated-inference-hosting
-
2026-01-21:
Dedicated Model Hosting launched
(product)
. Hyperbolic launched single-tenant dedicated model hosting with isolated networking, reserved GPUs, and private customer-only endpoints.
Source: Hyperbolic blog.
https://www.hyperbolic.ai/blog/dedicated-inference-hosting
TensorWave
Website: tensorwave.com
TensorWave is an AMD-exclusive AI cloud provider that supplies GPU-based infrastructure for training, fine-tuning, inference, and other large-scale AI workloads.
Category: Compute capacity & supply.
Funding stage: Series B.
Founded: 2023.
Country: United States.
Employees: 51-200 employees.
Landscape fit: It fits this category because it owns and operates GPU cloud capacity and has raised capital to expand large-scale AMD Instinct GPU infrastructure, effectively aggregating and monetizing compute supply for AI customers.
Funding rounds
-
Series B
announced 2026-06-01
: 350000000 USD
; investors: Magnetar, AMD Ventures, Maverick Silicon, Nexus Venture Partners, Western Frontier
; source: Business Wire
; https://www.businesswire.com/news/home/20260610650010/en/TensorWave-Raises-%24350-Million-Series-B-at-%241.55B-Valuation-to-Expand-Global-AMD-Powered-AI-Infrastructure
-
Series A
announced 2025-05-01
: 100000000 USD
; investors: Magnetar, AMD Ventures, Maverick Silicon, Nexus Venture Partners, Prosperity7
; source: Business Wire
; https://www.businesswire.com/news/home/20250514340458/en/TensorWave-Secures-%24100-Million-Series-A-Funding-Co-Led-by-Magnetar-and-AMD-Ventures
-
SAFE
announced 2024-10-01
: 43000000 USD
; investors: Nexus VP, Maverick Capital, Translink Capital, Javelin Venture Partners, StartupNV, AMD Ventures
; source: Business Wire
; https://www.businesswire.com/news/home/20241008734926/en/TensorWave-Raises-%2443M-in-SAFE-Funding-the-Largest-in-Nevada-Startup-History-to-Advance-AI-Compute-Solutions
Company timeline
-
2026-06-01:
Series B
(funding)
. Amount: 350000000 USD. Investors: Magnetar, AMD Ventures, Maverick Silicon, Nexus Venture Partners, Western Frontier
Source: Business Wire.
https://www.businesswire.com/news/home/20260610650010/en/TensorWave-Raises-%24350-Million-Series-B-at-%241.55B-Valuation-to-Expand-Global-AMD-Powered-AI-Infrastructure
-
2026-06-01:
Current hiring cluster in data center, networking, engineering, operations, and recruiting
(hr)
. TensorWave’s live jobs board shows multiple open roles, including data center operations, infrastructure engineering, network engineering, head of global operations, technical recruiter, and related positions.. Team size observed: 51-200 employees employees.
Source: TensorWave Ashby jobs and LinkedIn hiring posts.
([jobs.ashbyhq.com](https://jobs.ashbyhq.com/tensorwave/5f3b2d7b-efac-4444-8c4c-a76dbc7a19ed?utm_source=openai))
-
2026-05-14:
Web Summit appearance reinforced AMD-open-ecosystem positioning
(gtm)
. TensorWave used Web Summit Vancouver 2026 to frame the market around open ecosystems, sovereign AI, distributed deployment, and real-time inference.
Source: TensorWave blog.
https://tensorwave.com/blog/the-future-of-ai-infrastructure-7-key-industry-takeaways-from-web-summit-vancouver-2026
-
2026-04-06:
Modular partnership sharpened the enterprise inference motion
(gtm)
. TensorWave and Modular published a joint post positioning TensorWave as the AMD-exclusive infrastructure layer for Modular's inference stack and partner solutions.
Source: TensorWave blog.
https://tensorwave.com/blog/tensorwave-and-modular-breaking-ai-inference-barriers-beyond-nvidia
-
2026-04-06:
Inference partnership with Modular broadened the application stack
(product)
. TensorWave published a partnership post with Modular that positioned MAX for AMD-based inference workloads, including image generation and image editing support in Modular 26.2. ([tensorwave.com](https://tensorwave.com/blog/tensorwave-and-modular-breaking-ai-inference-barriers-beyond-nvidia?utm_source=openai))
Source: TensorWave blog.
([tensorwave.com](https://tensorwave.com/blog/tensorwave-and-modular-breaking-ai-inference-barriers-beyond-nvidia?utm_source=openai))
-
2026-02-04:
Moreh customer proof showed inference performance gains on MI355X
(gtm)
. TensorWave highlighted Moreh running frontier LLM inference on MI355X and reducing model weight download time by about 75%.
Source: TensorWave blog.
https://tensorwave.com/blog/tensorwave-powers-morehs-frontier-llm-inference-on-mi355x-and-reduces-model-weight-download-time-by-75
Nscale
Website: nscale.com
Nscale is a UK-based AI infrastructure company that builds and operates large-scale GPU cloud platforms, data centers, and the underlying compute stack for training, fine-tuning, and inference.
Category: Compute capacity & supply.
Funding stage: Series C.
Founded: 2024.
Country: United Kingdom.
Employees: 201-500 employees.
Landscape fit: It fits this category because it owns and operates AI compute infrastructure—GPU cloud capacity and data centers—and aggregates large-scale GPU supply for customers. Nscale’s site describes it as building the advanced infrastructure and systems for AI, and its recent Series C funding supports further expansion of that compute capacity. ([nscale.com](https://www.nscale.com/about?utm_source=openai))
Funding rounds
-
Project Financing
announced 2026-05-01
: 790000000 USD
; investors: ABN AMRO, DNB, Eksfin, Nordea, SEB
; source: Nscale
; https://www.nscale.com/press-releases/nscale-secures-790-million-norway
-
Series C
announced 2026-03-01
: 2000000000 USD
; investors: Aker ASA, 8090 Industries, Astra Capital Management, Citadel, Dell, Jane Street, Lenovo, Linden Advisors, Nokia, NVIDIA, Point72
; source: Nscale
; https://www.nscale.com/press-releases/nscale-series-c
-
Delayed Draw Term Loan
announced 2026-02-01
: 1400000000 USD
; investors: PIMCO, Blue Owl, LuminArx Capital Management
; source: Nscale
; https://www.nscale.com/press-releases/nscale-signs-1-4bn-delayed-draw-term-loan
-
Pre-Series C SAFE
announced 2025-10-01
: 433000000 USD
; investors: Blue Owl Managed Funds, Dell, NVIDIA, Nokia
; source: Nscale
; https://www.nscale.com/press-releases/nscale-pre-series-c-safe
-
Series B
announced 2025-09-01
: 1100000000 USD
; investors: Aker ASA, Sandton Capital, Blue Owl, Dell, Fidelity Management & Research Company, G Squared, Nokia, NVIDIA, Point72, T.Capital
; source: Nscale
; https://www.nscale.com/press-releases/nscale-series-b
Company timeline
-
2026-06-01:
Broad open-roles slate across infrastructure and corporate functions
(hr)
. Nscale's open positions page shows many live openings across AI infrastructure engineering, security, data center delivery, legal, finance, operations, people, and corporate affairs.. Team size observed: 201-500 employees employees.
Source: Nscale open positions page.
https://www.nscale.com/open-positions
-
2026-05-01:
Project Financing
(funding)
. Amount: 790000000 USD. Investors: ABN AMRO, DNB, Eksfin, Nordea, SEB
Source: Nscale.
https://www.nscale.com/press-releases/nscale-secures-790-million-norway
-
2026-04-28:
President of Data Centers appointed
(hr)
. Nscale appointed Sam Huckaby as President of Data Centers to lead global build, operations, and capital deployment.. Team size observed: 201-500 employees employees.
Source: Nscale press release.
https://www.nscale.com/press-releases/nscale-appoints-sam-huckaby
-
2026-03-16:
Acquired AIPCorp and formed Nscale Energy & Power
(gtm)
. Nscale agreed to acquire American Intelligence & Power Corporation, adding a power-and-compute platform and a new energy division in the U.S. ([nscale.com](https://www.nscale.com/press-releases/nscale-acquires-american-intelligence-power-corporation?utm_source=openai))
Source: Nscale press release.
https://www.nscale.com/press-releases/nscale-acquires-american-intelligence-power-corporation
-
2026-03-16:
Full-stack AI infrastructure posture expands with AIPCorp acquisition
(product)
. Nscale agreed to acquire American Intelligence & Power Corporation, adding the Monarch Compute Campus and creating a new energy-and-power division.
Source: Nscale.
https://www.nscale.com/press-releases/nscale-acquires-american-intelligence-power-corporation
-
2026-03-01:
Series C
(funding)
. Amount: 2000000000 USD. Investors: Aker ASA, 8090 Industries, Astra Capital Management, Citadel, Dell, Jane Street, Lenovo, Linden Advisors, Nokia, NVIDIA, Point72
Source: Nscale.
https://www.nscale.com/press-releases/nscale-series-c
Verda
Website: verda.com
Verda is a Helsinki-based AI cloud infrastructure provider that offers on-demand GPU instances, instant and bare-metal clusters, and serverless inference for training and serving AI workloads.
Category: Compute capacity & supply.
Founded: 2020.
Country: Finland.
Employees: 51-200 employees (LinkedIn; 128 employees visible).
Landscape fit: It fits this category because Verda owns/operates GPU cloud capacity and data-center infrastructure, sells reserved and on-demand compute, and positions itself as a sovereign AI cloud / neocloud provider for GPU-heavy workloads.
Funding rounds
-
Debt and Equity
announced 2026-04-01
: 100000000 EUR
; investors: Lifeline Ventures, byFounders, Tesi, Varma
; source: Verda
; https://verda.com/blog/verda-raises-117m-funding
-
Series A
announced 2025-09-01
: 55000000 EUR
; investors: byFounders, Skaala, Varma, Tesi, J12 Ventures, Nordea, Armada Credit Partners, Danske Bank, Norion Bank, Local Tapiola
; source: Verda
; https://verda.com/blog/datacrunch-raises-usd64m-in-series-a
-
Seed
announced 2024-10-01
: 13000000 USD
; investors: byFounders, J12 Ventures, Lasse Espeholt, Nal Kalchbrenner, Oskari Saarenmaa, Henrik Rosendahl, Anders Bo Pedersen, Maaike Bryon, Ari Tulla, Tuomo Riekki, Tapio Tolvanen, Local Tapiola, Nordea
; source: Verda
; https://verda.com/blog/datacrunch-io-secures-usd13-million-seed-round-to-transform-ai-computing
-
Pre-Seed
announced 2021-11-01
: 1100000 USD
; source: Seedtable
; https://www.seedtable.com/funding-round/DataCrunch.io_November_2021_Pre-Seed_Funding_Round-D5J9ANG
Company timeline
-
2026-06-02:
Next-wave GPU rollout across Europe, the US, and APAC
(gtm)
. Verda announced early deployments of NVIDIA Vera Rubin systems and said its first-availability window now spans Europe, the US, and APAC.
Source: Verda Blog.
https://verda.com/blog/nvidia-vr200-r200-early-deployments
-
2026-06-02:
Vera Rubin early-deployment plan announced
(product)
. Verda said it would be among the first to deploy NVIDIA Vera Rubin NVL72 and HGX Rubin NVL8 at scale in H2 2026 across Europe, the US, and APAC.
Source: Verda blog.
https://verda.com/blog/nvidia-vr200-r200-early-deployments
-
2026-06-01:
Current hiring cluster appears on careers page
(hr)
. Verda’s careers site currently lists multiple open roles, including Partner Marketing Lead, Product Marketing Lead, Talent Acquisition Partner, Data Center Deployment Engineer - GPU & HPC, and Executive Assistant.. Team size observed: 51-200 employees (LinkedIn; 128 employees visible) employees.
Source: Verda careers page.
https://jobs.verda.com/
-
2026-05-01:
Founder post signals team growth and ongoing hiring
(hr)
. Founder Ruben Bryon said Verda expects to employ close to 100 people in Finland and noted that the company is hiring AI researcher and engineer teams.. Team size observed: 51-200 employees (LinkedIn; 128 employees visible) employees.
Source: Ruben Bryon LinkedIn post.
https://www.linkedin.com/company/verda-cloud/
-
2026-04-30:
Arm-native agentic infrastructure direction
(gtm)
. Verda introduced an Arm AGI CPU deployment alongside NVIDIA GB300 and future Vera Rubin systems, framing the stack around agentic AI workloads.
Source: Verda Blog.
https://verda.com/blog/arm-agi-cpu-launch-partner
-
2026-04-30:
Arm AGI CPU added to next-generation infrastructure
(product)
. Verda said it was deploying the Arm AGI CPU across its infrastructure, paired with NVIDIA GB300 today and upcoming Vera Rubin systems for agentic AI workloads.
Source: Verda blog.
https://verda.com/blog/arm-agi-cpu-launch-partner
Compute procurement, resale & routing
Platforms that help buyers access, compare, reserve, resell, or route compute across providers.
Compute Exchange
Website: compute.exchange
A GPU compute marketplace that lets buyers compare, reserve, and auction cloud compute capacity from verified providers, with transparent pricing and standardized contracts.
Category: Compute procurement, resale & routing.
Founded: 2024.
Country: United States.
Employees: 2-10 employees.
Landscape fit: It fits because the platform directly supports compute procurement and resale/routing by matching buyers and sellers, enabling real-time price discovery, reserved capacity access, and monetization of unused compute across multiple providers.
Company timeline
-
2026-06-05:
SOC 2 Type II compliance announced
(gtm)
. Compute Exchange announced that it is now SOC 2 Type II compliant, adding a formal trust and security credential to the marketplace.
Source: Compute Exchange blog.
https://compute.exchange/blogs
-
2026-06-05:
SOC 2 Type II compliance announced
(product)
. Compute Exchange published that it is now SOC 2 Type II compliant. ([compute.exchange](https://compute.exchange/blogs?utm_source=openai))
Source: Compute Exchange blog.
https://compute.exchange/blogs
-
2026-06-01:
Support expands to more GPU families and condition tiers
(product)
. The product pages show coverage extending across reserved, used, and refurbished inventory for models including H100, H200, B200, A100, L40S, V100, and AMD MI300X/MI325X. ([compute.exchange](https://compute.exchange/hardware-market?utm_source=openai))
Source: Compute Exchange product pages.
https://compute.exchange/gpus
-
2026-06-01:
Hardware Market adds used and refurbished GPU trading
(product)
. Compute Exchange introduced a hardware-market surface for used and refurbished data-center GPUs, with indicative price ranges, condition tiers, and counterparty verification for institutional buyers and sellers. ([compute.exchange](https://compute.exchange/hardware-market?utm_source=openai))
Source: Compute Exchange hardware market.
https://compute.exchange/hardware-market
-
2026-03-01:
Reserved GPU marketplace becomes a named product surface
(product)
. By early 2026, Compute Exchange was publishing reserved-GPU pages that emphasized verified providers, side-by-side comparison, live quotes within 24 hours, and a single contract / unified SLA experience. ([compute.exchange](https://compute.exchange/gpus?utm_source=openai))
Source: Compute Exchange product pages.
https://compute.exchange/gpus
-
2026-02-01:
Partner/referral channel launched to source buyers through warm introductions
(gtm)
. Compute Exchange introduced a referral program that offers revenue share to partners who introduce compute buyers, with CX handling onboarding and deal execution.
Source: Compute Exchange referral program.
https://compute.exchange/referral-program
Shadeform
Website: shadeform.com
Shadeform is a GPU cloud marketplace and unified control plane that helps AI/ML teams compare, reserve, provision, and manage GPU compute across 30+ cloud providers.
Category: Compute procurement, resale & routing.
Funding stage: Pre Seed.
Founded: 2023.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits this category because it helps buyers access and compare compute across multiple providers, reserve capacity, and route workloads to the best available GPU infrastructure through a single platform and API.
Funding rounds
-
pre-seed
announced 2023-09-01
: 500000 USD
; investors: Y Combinator, Unpopular Ventures, Soma Capital, MDSV Capital, Charlotte Fund, Gravity Fund, Dig Ventures
; source: Caplight
; https://www.caplight.com/company/shadeform
Company timeline
-
2026-06-01:
Hiring cluster: marketing and operations roles
(hr)
. Shadeform posted that it is hiring a marketing leader and an operations specialist, while the careers page still presents the company as actively hiring.. Team size observed: 11-50 employees employees.
Source: LinkedIn.
https://www.linkedin.com/company/shadeformai
-
2026-06-01:
Jeffrey Jewett joined as COO
(hr)
. Shadeform publicly welcomed Jeffrey Jewett as Chief Operating Officer.. Team size observed: 11-50 employees employees.
Source: LinkedIn.
https://www.linkedin.com/company/shadeformai
-
2026-04-09:
Customer proof broadened to infrastructure platforms
(gtm)
. Shadeform highlighted Parasail and Tinfoil, showing the platform being used to source hard-to-find GPU configurations quickly for an AI inference platform and a private AI platform with secure-hardware requirements. ([shadeform.com](https://shadeform.com/resources/customer-stories/how-parasail-turns-around-any-customer-gpu-requirement-in-under-an-hour-using-shadeform?utm_source=openai))
Source: Shadeform.
https://shadeform.com/resources/customer-stories
-
2026-02-12:
Blackwell Ultra launch with inquiry-based reserved clusters
(gtm)
. Shadeform announced Blackwell Ultra B300 availability, including on-demand access plus reserved HGX B300 and GB300 clusters available via inquiry, extending the platform into the newest scarce accelerator supply. ([shadeform.com](https://shadeform.com/resources/articles/nvidia-blackwell-ultra-b300s-now-available-on-shadeform?utm_source=openai))
Source: Shadeform.
https://shadeform.com/resources/articles/nvidia-blackwell-ultra-b300s-now-available-on-shadeform
-
2026-02-12:
Formalized reserved-capacity motion for larger deployments
(gtm)
. Shadeform’s reserved commitments flow made long-term GPU sourcing inquiry-led, with quote-based reservation requests, flexible terms, and customer proof from teams like Fal, Parasail, and Tinfoil that need scale, speed, and scarce configurations. ([shadeform.com](https://shadeform.com/features/reserved-commitments?utm_source=openai))
Source: Shadeform.
https://shadeform.com/features/reserved-commitments
-
2026-02-12:
Blackwell Ultra B300 availability on Shadeform
(product)
. Shadeform announced on-demand access to NVIDIA Blackwell Ultra B300 GPUs, with reserved HGX B300 and GB300 clusters available via inquiry. ([shadeform.com](https://shadeform.com/resources/articles/nvidia-blackwell-ultra-b300s-now-available-on-shadeform?utm_source=openai))
Source: Shadeform.
https://shadeform.com/resources/articles/nvidia-blackwell-ultra-b300s-now-available-on-shadeform
San Francisco Compute Company
Website: sfcompute.com
San Francisco Compute operates large-scale GPU clusters and a marketplace that lets customers buy, sell back, and resell compute time so they can avoid being locked into long-term infrastructure contracts.
Category: Compute procurement, resale & routing.
Funding stage: Series A.
Founded: 2023.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits this category because it helps buyers access and reserve compute across providers, and its marketplace/docs explicitly support selling back or reselling unused compute capacity.
Funding rounds
-
Series A
announced 2025-11-01
: 40000000 USD
; investors: DCVC, Wing Venture Capital, Electric Capital, Alt Capital
; source: Data Center Dynamics
; https://www.datacenterdynamics.com/en/news/sf-compute-raises-40m-for-ai-compute-marketplace-offering/
-
seed
announced 2024-07-01
: 12000000 USD
; investors: Alt Capital
; source: PYMNTS / Dealroom summary of 36kr report
; https://www.datacenterdynamics.com/en/news/sf-compute-raises-40m-for-ai-compute-marketplace-offering/
Company timeline
-
2026-06-09:
Dashboard navigation redesign and migration tools
(product)
. The dashboard was redesigned around Capacities and Instances, and SF Compute added node-to-instance migration, image revoke/reinstate controls, and auto-reload for billing. ([sfcompute.com](https://sfcompute.com/changelog))
Source: Changelog.
https://sfcompute.com/changelog
-
2026-06-09:
Redesigned dashboard around capacities and instances
(gtm)
. SF Compute redesigned the dashboard around capacities and instances and provided migration tooling from v1 nodes to v2 instances. ([sfcompute.com](https://sfcompute.com/changelog))
Source: SF Compute changelog.
https://sfcompute.com/changelog
-
2026-06-01:
New engineering and people roles posted on LinkedIn
(hr)
. SF Compute’s LinkedIn company page showed a run of new posts for Software Engineer - Systems, Software Engineer - Product, Principal/Staff HPC Network Engineer, Software Engineer - Cloud, People & Talent Aquisition Leader, and Software Engineer, Hermetic Build.. Team size observed: 11-50 employees employees.
Source: LinkedIn.
https://www.linkedin.com/company/sfcompute
-
2026-05-12:
Nodes renamed to Instances and orderbook API added
(product)
. The v2 update renamed nodes to instances, introduced capacity transfers for isolating resellable compute, and added an orderbook API for live buy/sell order visibility. ([sfcompute.com](https://sfcompute.com/changelog))
Source: Changelog.
https://sfcompute.com/changelog
-
2026-05-12:
Renamed nodes to instances and added capacity transfers
(gtm)
. SF Compute completed the v2 terminology shift from nodes to instances and introduced capacity transfers to isolate sellable compute time. ([sfcompute.com](https://sfcompute.com/changelog))
Source: SF Compute changelog.
https://sfcompute.com/changelog
-
2026-05-01:
Hiring cluster visible on careers page
(hr)
. SF Compute’s careers page and job mirrors show a broad set of open roles across engineering and operations, including software engineering, HPC/networking, people/talent, and accounting positions.. Team size observed: 11-50 employees employees.
Source: SF Compute / Scoutify / job board mirrors.
https://sfcompute.com/
Vast.ai
Website: vast.ai
Vast.ai is a GPU cloud marketplace and infrastructure platform that lets AI researchers, startups, and enterprises find, compare, reserve, and run compute across a distributed network of independent GPU hosts. ([vast.ai](https://vast.ai/press-kit?utm_source=openai))
Category: Compute procurement, resale & routing.
Founded: 2016.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits this category because its core product is a two-sided compute marketplace with real-time pricing and API/CLI access, designed to match compute supply and demand across providers and let customers launch, compare, and scale GPU workloads quickly. ([vast.ai](https://vast.ai/?utm_source=openai))
Funding rounds
-
Seed
announced 2024-07-01
: 4000000 USD
; investors: DRW Trading Group, Nazaré Ventures
; source: Caplight
; https://www.caplight.com/company/vast-ai
Company timeline
-
2026-06-09:
B200/B300 supply and serverless metrics improved
(product)
. Vast.ai added B200 and B300 GPU supply, revamped serverless metrics and UI performance, and shipped new Qwen3.6 and Gemma 4 templates.
Source: June 2026 Product Update.
https://vast.ai/article/june-2026-product-update
-
2026-06-09:
Marketplace supply expanded to B200 and B300 plus migration guides
(gtm)
. Vast.ai’s June 2026 update brought B200/B300 availability, serverless template additions, and new migration guides such as moving from Salad Cloud to Vast.ai.
Source: Vast.ai article.
https://vast.ai/article/june-2026-product-update
-
2026-06-01:
Security Engineer opening published
(hr)
. Vast.ai posted an application page for a Security Engineer role in Los Angeles, California.. Team size observed: 11-50 employees employees.
Source: Vast.ai Jobs.
https://vast.ai/jobs/apply/security-engineer
-
2026-05-12:
Benchmarking CLI and stronger auth controls shipped
(product)
. Vast.ai added a benchmarking CLI, upgraded 2FA to support authenticator apps and backup codes, merged the SDK and CLI repos, and introduced an All-in-One App Studio.
Source: May 2026 Product Update.
https://vast.ai/article/may-2026-product-update
-
2026-05-12:
Benchmarking CLI and creative app studio added
(gtm)
. Vast.ai’s May 2026 product update introduced a benchmarking CLI for GPU price-performance comparison and an all-in-one creative AI studio template.
Source: Vast.ai article.
https://vast.ai/article/may-2026-product-update
-
2026-02-13:
Startup Program and model-first packaging introduced
(product)
. Vast.ai launched the Startup Program with GPU credits and priority support, and added the Model Library plus orchestration support updates such as SkyPilot enhancements.
Source: January & February 2026 Product Updates.
https://vast.ai/article/january-february-2026-product-update
OpenRouter
Website: openrouter.ai
OpenRouter is a unified API and marketplace for large language models that lets developers access, compare, and route requests across many model providers through a single interface.
Category: Compute procurement, resale & routing.
Funding stage: Series B.
Founded: 2023.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits this category because it helps buyers of compute-like AI inference capacity route workloads across multiple providers, compare options, and switch between them to optimize price, performance, and reliability.
Funding rounds
-
Series B
announced 2026-05-01
: 113000000 USD
; investors: CapitalG, NVentures, ServiceNow Ventures, MongoDB Ventures, Snowflake Ventures, Databricks Ventures, AMP PBC, Pace Capital, Andreessen Horowitz, Menlo Ventures
; source: OpenRouter Raises $113M Series B
; https://openrouter.ai/announcements/series-b
-
Seed and Series A
announced 2025-06-01
: 40000000 USD
; investors: Andreessen Horowitz, Menlo Ventures, Sequoia, prominent industry angels
; source: OpenRouter raises $40 million to scale up multi-model inference for enterprise
; https://www.globenewswire.com/news-release/2025/06/25/3105125/0/en/openrouter-raises-40-million-to-scale-up-multi-model-inference-for-enterprise.html
Company timeline
-
2026-06-10:
Advisor and subagent server tools deepen agent platform
(gtm)
. OpenRouter launched advisor and subagent tools that let models consult stronger models or delegate tasks mid-generation. ([openrouter.ai](https://openrouter.ai/blog/announcements/advisor-server-tool/?utm_source=openai))
Source: Advisor: Give Any Model a Lifeline to a Smarter One.
https://openrouter.ai/blog/announcements/advisor-server-tool/
-
2026-06-01:
Current LinkedIn job post for Software Engineer, Product
(hr)
. OpenRouter has a live LinkedIn job post recruiting Software Engineers for the Product team, describing a small team owning dashboards, admin tooling, billing, analytics, and onboarding.. Team size observed: 11-50 employees employees.
Source: LinkedIn Jobs.
https://www.linkedin.com/jobs/view/software-engineer-product-at-openrouter-4413715984
-
2026-06-01:
OpenRouter careers page shows a multi-function hiring cluster
(hr)
. OpenRouter’s careers page lists multiple open positions across Engineering, Finance & Operations, Marketing, Product & Design, and Revenue, with remote US hiring emphasized.. Team size observed: 11-50 employees employees.
Source: OpenRouter Careers.
https://openrouter.helicone.ai/careers
-
2026-05-29:
Guardrails
(product)
. OpenRouter launched workspace guardrails for budget enforcement, zero data retention, provider restrictions, prompt injection defense, and data loss prevention.
Source: OpenRouter blog.
https://openrouter.ai/blog/announcements/guardrails/
-
2026-05-29:
Guardrails target enterprise governance
(gtm)
. OpenRouter launched configurable guardrails for budgets, zero data retention, model/provider restrictions, prompt injection defense, and data loss prevention. ([openrouter.ai](https://openrouter.ai/blog/announcements/))
Source: Guardrails: Protect your Agents, Data, and Costs.
https://openrouter.ai/blog/guardrails-protect-your-agents-data-and-costs/
-
2026-05-07:
Web search and fetch across every model
(product)
. OpenRouter introduced server-side web search and web fetch tools that any tool-calling model can invoke during a request.
Source: OpenRouter blog.
https://openrouter.ai/blog/announcements/agentic-web-tools/
Compute market data & benchmarks
Companies creating prices, indices, benchmarks, or cost/performance intelligence.
Silicon Data
Website: silicondata.com
Silicon Data builds pricing, benchmark, and market-intelligence tools for the GPU/compute economy, including daily GPU rental-price indices and performance benchmarks for AI infrastructure markets.
Category: Compute market data & benchmarks.
Funding stage: Seed.
Founded: 2024.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits this category because it creates compute market data products—daily GPU price indices, performance benchmarks, and transparency tools used for price discovery and cost/performance intelligence in AI infrastructure markets.
Funding rounds
-
Seed
announced 2025-03-01
: 4700000 USD
; investors: DRW, Jump Trading Group
; source: Business Wire
; https://www.businesswire.com/news/home/20250320674807/en/Silicon-Data-Secures-%244.7M-in-Seed-Funding-and-Expands-Compute-Industry-Data-Product-Suite
Company timeline
-
2026-06-01:
Team size observed
(hr)
. Public source shows a team size of 11-50 employees employees.. Team size observed: 11-50 employees employees.
Source: LinkedIn.
https://www.linkedin.com/company/silicon-data
-
2026-05-12:
Announced CME compute futures partnership
(gtm)
. Silicon Data and CME Group announced plans to launch compute futures later in 2026, using Silicon Data’s indices as the underlying benchmark. ([silicondata.com](https://www.silicondata.com/news-room/cme-group-and-silicon-data-partner-to-launch-first-compute-futures?utm_source=openai))
Source: Silicon Data newsroom.
https://www.silicondata.com/news-room/cme-group-and-silicon-data-partner-to-launch-first-compute-futures
-
2026-05-12:
Partnered with CME Group to launch compute futures
(product)
. Silicon Data and CME Group announced a partnership to launch the first compute futures, extending Silicon Data’s benchmarks into a derivatives-adjacent market workflow.
Source: Silicon Data News Room.
https://www.silicondata.com/news-room/cme-group-and-silicon-data-partner-to-launch-first-compute-futures
-
2026-04-20:
Launched GPU Forward Curve
(gtm)
. Silicon Data launched a GPU Forward Curve to show term structure and implied forward rates for AI compute over a 36-month horizon. ([silicondata.com](https://www.silicondata.com/news-room/silicon-data-unveils-first-gpu-forward-curve))
Source: Silicon Data newsroom.
https://www.silicondata.com/news-room/silicon-data-unveils-first-gpu-forward-curve
-
2026-04-20:
Launched GPU forward curve
(product)
. Silicon Data introduced a GPU Forward Curve for forward-looking rental pricing visibility in AI infrastructure markets.
Source: Silicon Data News Room.
https://www.silicondata.com/news-room/silicon-data-unveils-first-gpu-forward-curve-signaling-transition-of-ai-compute-into-a-tradable-commodity
-
2026-02-01:
Introduced the Silicon Data Portal subscription bundle
(gtm)
. Silicon Data’s pricing page shows a $499/month portal package with index dashboards, predictive modeling, daily values, and add-on modules such as forward curves, token pricing, and country-level data. ([silicondata.com](https://www.silicondata.com/pricing))
Source: Silicon Data pricing page.
https://www.silicondata.com/pricing
AIMultiple
Website: aimultiple.com
AIMultiple is an independent research and benchmarking company that publishes data-driven comparisons, benchmarks, and buying guides for enterprise software and AI tools.
Category: Compute market data & benchmarks.
Founded: 2019.
Country: Estonia.
Employees: 11-50 employees.
Landscape fit: It fits this category because it creates quantitative benchmarks, feature comparisons, and cost/performance intelligence for AI, infrastructure, and software products—exactly the kind of market data and benchmarking layer used to price and evaluate compute-related tools.
Company timeline
-
2026-06-16:
First independent AI code review benchmark and continued category expansion
(product)
. AIMultiple published the world’s first independent benchmark of AI code review platforms and continued releasing adjacent benchmarks such as agentic coding, long-context retrieval, and email archiving software in 2026.
Source: AIMultiple About / AIMultiple Email Archiving Software Benchmark.
https://aimultiple.com/about
-
2026-06-05:
Agentic coding benchmark launched
(gtm)
. AIMultiple published A-CODE-LLM Bench, an agentic coding benchmark for large language models across software development tasks.
Source: A-CODE-LLM Bench: Agentic Coding Benchmark | AIMultiple.
https://aimultiple.com/pt/agentic-llm
-
2026-06-01:
Careers page signals active hiring
(hr)
. AIMultiple’s about and culture pages invite candidates to view open positions and describe the company as growing fast.. Team size observed: 11-50 employees employees.
Source: AIMultiple.
https://aimultiple.com/about
-
2026-05-07:
AI coding benchmark compares CLI tools vs IDE editors
(gtm)
. AIMultiple published a benchmark comparing Claude Code, Cursor, and other coding tools across full-stack tasks.
Source: AI Coding Benchmark: Claude Code vs Cursor | AIMultiple.
https://aimultiple.com/ai-coding-benchmark
-
2026-03-13:
AI code review benchmark launched
(gtm)
. AIMultiple published its first independent benchmark of AI code review platforms, positioning code review as a new evaluation category.
Source: AI Code Review Tools Benchmark | AIMultiple.
https://aimultiple.com/ai-code-review-tools
-
2025-01-01:
Benchmark portfolio expanded into enterprise AI and software
(gtm)
. AIMultiple said it launched over 100 benchmarks in 2025 and added benchmark-curation pages for AI models, AI GPUs, MCP servers, and MCP clients.
Source: About | AIMultiple.
https://aimultiple.com/about
gpuprice.fyi
Website: gpuprice.fyi
An independent, developer-facing cloud GPU price index that normalizes on-demand GPU rental prices across providers and exposes live pricing, history, methodology, and a public API.
Category: Compute market data & benchmarks.
Funding stage: Bootstrapped.
Founded: 2026.
Country: United States.
Landscape fit: It fits the compute market data & benchmarks category because it publishes normalized GPU cloud price indices, historical trends, and citable benchmarks that compare compute costs across providers.
Company timeline
-
2026-06-01:
Methodology-first credibility and citation layer
(gtm)
. The company formalized a methodology page that explains normalization, quality tiers, freshness rules, snapshot history, and explicit boundaries against treating the index as a transaction benchmark.
Source: gpuprice.fyi methodology and about pages.
gpuprice.fyi/methodology and gpuprice.fyi/about
-
2026-06-01:
Comparison pages added for citable GPU price lookups
(product)
. The product added SKU comparison pages that normalize and contrast cloud GPU rental prices across providers.
Source: gpuprice.fyi comparison page.
https://www.gpuprice.fyi/compare/a100-vs-rtx-5090
-
2026-06-01:
Public launch with live index, API, and snapshots
(gtm)
. The site is now live with current pricing, historical curves, a public API, downloadable snapshots, methodology pages, and citable SKU pages.
Source: gpuprice.fyi homepage and API.
gpuprice.fyi/ and gpuprice.fyi/api/v1/prices/h100-sxm
-
2026-06-01:
Live index launched with history, methodology, and public API
(product)
. The site is now live with an index median, historical charts built from snapshots, methodology documentation, downloadables, and a public API endpoint for SKU data.
Source: gpuprice.fyi homepage, methodology, about, and API.
https://www.gpuprice.fyi/
-
2026-05-01:
Pre-launch data pipeline and index design
(gtm)
. The company established its initial cloud GPU pricing pipeline, normalized all listings to USD per GPU per hour, and introduced quality tiers, exclusion reasons, and versioned schema fields before public launch.
Source: gpuprice.fyi changelog.
gpuprice.fyi/changelog
-
2026-05-01:
Pre-launch pricing pipeline and methodology v0.3 defined
(product)
. The product’s initial pipeline was set up with live sources, USD-per-GPU-per-hour normalization, quality tiers, explicit exclusion reasons, and a versioned snapshot schema.
Source: gpuprice.fyi Changelog.
https://www.gpuprice.fyi/changelog
SkyPilot
Website: skypilot.co
SkyPilot is an open-source framework and GPU/AI infrastructure platform that helps teams run, manage, and scale AI workloads across Kubernetes, Slurm, and 20+ clouds, with a GPU pricing dashboard (GPU Compass) that compares availability and cost across providers.
Category: Compute market data & benchmarks.
Founded: 2021.
Country: United States.
Employees: 1 employee.
Landscape fit: It fits this category because GPU Compass is explicitly a compute market data tool: it aggregates and updates GPU pricing and availability across many cloud providers, effectively acting as a price/comparison benchmark for compute markets.
Company timeline
-
2026-06-08:
SkyPilot Sandboxes adds BYOC execution for agent code
(gtm)
. SkyPilot launched sandboxes for running untrusted agent-generated code on a customer’s own Kubernetes clusters at scale.
Source: SkyPilot Blog.
https://blog.skypilot.co/sandboxes/
-
2026-06-08:
SkyPilot Sandboxes add BYOC agent code execution
(product)
. SkyPilot launched Sandboxes to run untrusted agent code on users' own Kubernetes clusters at scale, with warm pools and sub-second launches. ([blog.skypilot.co](https://blog.skypilot.co/sandboxes/))
Source: SkyPilot Sandboxes: Run Agent Code on Your Own Kubernetes, at Scale.
https://blog.skypilot.co/sandboxes/
-
2026-06-01:
Team size observed
(hr)
. Public source shows a team size of 1 employee employees.. Team size observed: 1 employee employees.
Source: LinkedIn.
https://www.linkedin.com/company/skypilot-oss
-
2026-04-21:
GPU Compass launches as a GPU pricing and availability dashboard
(gtm)
. SkyPilot launched GPU Compass, a browsable dashboard for comparing GPU availability and pricing across 20+ cloud providers and thousands of offerings.
Source: SkyPilot Blog.
https://blog.skypilot.co/gpu-compass/
-
2026-04-21:
GPU Compass launches as a pricing and availability dashboard
(product)
. SkyPilot launched GPU Compass, a real-time dashboard for comparing GPU pricing, availability, and historical spot trends across 20+ providers and 2,000+ offerings. ([blog.skypilot.co](https://blog.skypilot.co/gpu-compass/?utm_source=openai))
Source: GPU Compass: Navigate the GPU Frontier Across 20+ Clouds & 2K+ Offerings.
https://blog.skypilot.co/gpu-compass/
-
2026-04-09:
Agent Skill launches SkyPilot into coding-agent workflows
(gtm)
. SkyPilot introduced an Agent Skill so AI coding agents can launch clusters, run training jobs, and manage cloud resources with natural language.
Source: SkyPilot Blog.
https://blog.skypilot.co/agent-skill/
Artificial Analysis, Inc.
Website: artificialanalysis.ai
An independent AI benchmarking and analysis company that compares model quality, speed, and price across LLMs and other generative AI systems.
Category: Compute market data & benchmarks.
Funding stage: Seed.
Founded: 2024.
Country: United States.
Employees: 40+ employees.
Landscape fit: It fits the compute market data & benchmarks category because its core product is pricing, performance, and leaderboard-style benchmarking intelligence for AI models and API providers, giving buyers a cost/performance reference for compute-intensive AI workloads.
Funding rounds
-
Seed
announced 2024-01-01
: 250000 USD
; investors: AI Grant, Nat Friedman, Daniel Gross, Andrew Ng
; source: AI Grant
; https://aigrant.com/
Company timeline
-
2026-06-16:
Updated Intelligence Index to v4.1
(product)
. On 16 June 2026, Artificial Analysis updated Intelligence Index v4.1 to emphasize agentic workloads and to revise token and cost metrics, while replacing and reshaping several underlying evaluations.
Source: Intelligence Benchmarking | Artificial Analysis.
https://artificialanalysis.ai/methodology/intelligence-benchmarking
-
2026-06-16:
Intelligence Index shift toward agentic workloads
(gtm)
. Artificial Analysis updated its Intelligence Index to v4.1, reweighting toward agentic tasks and upgrading core benchmarks such as GDPval-AA, Terminal-Bench, and banking workflows.
Source: Artificial Analysis.
https://artificialanalysis.ai/articles/artificial-analysis-intelligence-index-v4-1
-
2026-06-12:
Launched AA-AgentPerf
(product)
. On 12 June 2026, Artificial Analysis introduced AA-AgentPerf, a hardware benchmark for the agent era.
Source: Articles | Artificial Analysis.
https://artificialanalysis.ai/articles
-
2026-06-12:
Agentic hardware benchmark launch
(gtm)
. Artificial Analysis launched AA-AgentPerf, a hardware benchmark for the agent era built around real coding-agent trajectories and service-level capacity targets.
Source: Artificial Analysis.
https://artificialanalysis.ai/articles/aa-agentperf
-
2026-06-11:
Guardrail benchmark analysis with NVIDIA
(gtm)
. Artificial Analysis published a guardrail-model comparison with NVIDIA, using open datasets to evaluate safety, refusal, and latency across safety classifiers and moderation APIs.
Source: Artificial Analysis.
https://artificialanalysis.ai/articles/guardrail-safety-benchmark
-
2026-06-01:
Launched AA-WER Streaming
(product)
. On 1 June 2026, Artificial Analysis announced AA-WER Streaming, a speech-to-text streaming benchmark focused on accuracy and latency for voice-agent use cases.
Source: Changelog | Artificial Analysis.
https://artificialanalysis.ai/changelog
Epoch AI
Website: epoch.ai
Epoch AI is a nonprofit research organization that curates AI data, builds benchmarks and indices, and publishes analysis on the trajectory, economics, and capabilities of AI systems. ([epoch.ai](https://epoch.ai/about))
Category: Compute market data & benchmarks.
Founded: 2022.
Country: United States.
Landscape fit: It fits the compute market data & benchmarks category because it tracks AI compute and hardware trends, maintains databases on AI companies and model training inputs, publishes capability indices, and develops benchmarks such as FrontierMath. ([epoch.ai](https://epoch.ai/topics/finances?utm_source=openai))
Funding rounds
-
Grant
announced 2025-12-01
: 100000 USD
; investors: Girish Sastry
; source: Epoch AI Transparency
; https://epoch.ai/about/transparency
-
Grant
announced 2025-09-01
: 85000 USD
; investors: Sentinel Bio
; source: Epoch AI Transparency
; https://epoch.ai/about/transparency
-
Grant
announced 2025-04-01
: 8500000 USD
; investors: Coefficient Giving
; source: Epoch AI Transparency
; https://epoch.ai/about/transparency
-
Grant
announced 2025-03-01
: 70000 USD
; investors: Coefficient Giving
; source: Epoch AI Transparency
; https://epoch.ai/about/transparency
-
Grant
announced 2025-01-01
: 600000 USD
; investors: Jaan Tallinn
; source: Epoch AI Transparency
; https://epoch.ai/about/transparency
Company timeline
-
2026-06-12:
FrontierMath Tier 4 v2 major update
(product)
. Epoch AI released a major update to FrontierMath Tier 4 and reported fixes to a large share of problems.
Source: Epoch AI benchmark page.
https://epoch.ai/benchmarks/frontiermath-tier-4-v2
-
2026-05-01:
Active hiring cluster on careers page
(hr)
. Epoch AI’s careers page lists open roles for Designer, Researcher / Senior Researcher, and Data Scientist (Contract).
Source: Epoch AI Careers.
https://epoch.ai/careers
-
2026-01-27:
Released FrontierMath Open Problems
(gtm)
. Epoch AI launched FrontierMath: Open Problems, a pilot benchmark for unsolved math research problems.
Source: Epoch AI.
https://epoch.ai/frontiermath/open-problems/about
-
2026-01-27:
FrontierMath Open Problems released
(product)
. Epoch AI released a pilot benchmark of unsolved mathematics problems designed to test AI on research-level tasks.
Source: Epoch AI FrontierMath page.
https://epoch.ai/frontiermath/open-problems/about
-
2026-01-13:
AI Chip Sales Data Explorer launched
(product)
. Epoch AI launched a data explorer estimating AI chip sales, compute capacity, power usage, and spending over time for major chip vendors.
Source: Epoch AI blog.
https://epoch.ai/blog/introducing-the-ai-chip-sales-data-explorer
-
2025-12-01:
Grant
(funding)
. Amount: 100000 USD. Investors: Girish Sastry
Source: Epoch AI Transparency.
https://epoch.ai/about/transparency
Compute financial products & assetization
Companies turning compute exposure, GPU assets, or future cash flows into pure financial products.
CME Group
Website: cmegroup.com
CME Group is the world’s leading derivatives marketplace, offering futures, options, cash, OTC markets, market data, and clearing services across major asset classes. ([cmegroup.com](https://www.cmegroup.com/company/about-us.html))
Category: Compute financial products & assetization.
Country: United States.
Employees: 1,001-5,000 employees.
Landscape fit: It fits the category because CME Group is actively packaging compute into a tradable financial product: it announced the launch of first-in-class compute futures to let traders, financial institutions, AI builders, and cloud providers hedge price risk in the compute market. ([cmegroup.com](https://www.cmegroup.com/media-room/press-releases/2026/5/12/cme_group_and_silicondatapartnertolaunchfirstcomputefutures.html?utm_source=openai))
Company timeline
-
2026-06-29:
Launched financially settled Micro E-mini S&P 500 and Nasdaq-100 options
(product)
. CME Group announced new Micro E-mini equity index options with Monday-through-Friday expiries and cash settlement.
Source: CME Group press release.
https://www.cmegroup.com/media-room/press-releases/2026/6/11/cme_group_to_launchnewfinancially-settledmicroe-minisp500andnasd.html
-
2026-06-17:
CEO succession announced for 2027
(hr)
. CME Group announced that Terry Duffy will transition to Executive Chairman on March 1, 2027, and Lynne Fitzpatrick, then President and CFO, will become CEO and join the Board.. Team size observed: 1,001-5,000 employees employees.
Source: CME Group.
https://www.cmegroup.com/media-room/press-releases/2026/6/17/cme_group_terry_duffywillstepdownaschiefexecutiveofficerandtrans.html
-
2026-06-11:
Expanded WTI and gold with 24/7 trading and smaller contract sizes
(product)
. CME Group announced 24/7 trading for 1-Ounce Gold futures and a new smaller WTI crude oil contract.
Source: CME Group press release.
https://www.cmegroup.com/media-room/press-releases/2026/6/11/cme_group_to_expand247tradingforwticrudeoilandgold.html
-
2026-05-29:
Launched 24/7 trading for cryptocurrency futures and options
(product)
. CME Group launched continuous trading for its regulated crypto futures and options suite.
Source: CME Group press release.
https://www.cmegroup.com/media-room/press-releases/2026/5/29/cme_group_announceslaunchof247cryptocurrencyfuturesandoptionstra.html
-
2026-02-13:
Event contracts reached 100 million traded
(gtm)
. CME Group said event contracts reached 100 million contracts traded since launch, showing rapid early adoption.
Source: CME Group.
https://investor.cmegroup.com/news-releases/news-release-details/cme-group-announces-100-million-event-contracts-traded
-
2026-02-09:
Added Cardano, Chainlink, and Stellar futures
(product)
. CME Group announced first trades for ADA, LINK, and XLM futures, broadening the listed crypto lineup beyond Bitcoin, Ether, and Solana.
Source: CME Group press release.
https://www.cmegroup.com/media-room/press-releases/2026/2/11/cme_group_announcesfirsttradesfornewcardanochainlinkandstellarcr.html
Compute Labs
Website: computelabs.ai
Compute Labs is an AI infrastructure investment platform that tokenizes GPU/compute assets and packages them into yield-bearing financial products for investors.
Category: Compute financial products & assetization.
Funding stage: Pre Seed.
Founded: 2024.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits this category because its core product is financialization of compute: it turns GPUs and AI infrastructure into investable, yield-generating assets via tokenization, structured products, and GPU financing.
Funding rounds
-
Pre-Seed
announced 2024-07-01
: 3000000 USD
; investors: Protocol Labs, MH Ventures, Blockchain Coinvestors, OKX Ventures, CMS Holdings, HashKey Capital, Amber Group, P2 Ventures, Polygon Ventures, Symbolic Capital, Sandeep Nailwal, Illia Polosukhin, Austin Federa
; source: The Block
; https://www.theblock.co/amp/post/303398/solana-based-gpu-tokenization-protocol-compute-labs-funding-token-valuation
Company timeline
-
2026-06-01:
Team size observed
(hr)
. Public source shows a team size of 11-50 employees employees.. Team size observed: 11-50 employees employees.
Source: LinkedIn.
https://www.linkedin.com/company/computelabs-ai
-
2026-05-21:
Private capital positioning sharpened around compute as an investable asset
(gtm)
. By May 2026, Compute Labs was framing GPU infrastructure as the fastest-growing asset class in AI and describing private capital as locked out, reinforcing a direct-to-investor private placement strategy.
Source: Compute Labs Press page.
https://www.computelabs.ai/en/press
-
2025-12-28:
Family office thesis published for GPU infrastructure
(gtm)
. Compute Labs added a whitepaper positioning GPU-based infrastructure as an emerging asset class for family offices, widening its investor narrative beyond web3-native buyers.
Source: Compute Labs About page.
https://www.computelabs.ai/en/about
-
2025-08-05:
Mainstream press validation for the GPU vault model
(gtm)
. Compute Labs highlighted Forbes and CoinDesk coverage of its yield-bearing GPU vault, reinforcing credibility for its investor-facing AI infrastructure thesis.
Source: Compute Labs Newsletter.
https://news.computelabs.ai/compute-labs-featured-in-forbes-and-coindesk/
-
2025-08-05:
Compute Labs positioned itself as a broader AI infrastructure financial layer
(product)
. The company described expanding the compute ecosystem with financial modeling and derivatives such as GPU ETF, market index, perp, staking, and lending concepts.
Source: Compute Labs Newsletter.
https://news.computelabs.ai/compute-labs-featured-in-forbes-and-coindesk/
-
2025-07-17:
First $1M GPU vault went live with NVIDIA H200s
(product)
. Compute Labs said it had launched its first $1M GPU vault backed by NVIDIA H200s running live AI workloads.
Source: Compute Labs Press.
https://www.computelabs.ai/en/press
GAIB
Website: gaib.ai
GAIB is an AI infrastructure finance platform that tokenizes enterprise GPUs and related AI compute assets into onchain, yield-bearing financial products, including its synthetic dollar AID and staked AID (sAID). ([docs.gaib.ai](https://docs.gaib.ai/?utm_source=openai))
Category: Compute financial products & assetization.
Founded: 2024.
Country: Singapore.
Employees: 11-50 employees.
Landscape fit: It fits the category because its core product turns compute exposure into pure financial instruments: tokenized GPUs, GPU-backed yields, a synthetic dollar, and DeFi products built on compute asset cash flows. ([blog.gaib.ai](https://blog.gaib.ai/about-gaib-part-1/))
Funding rounds
-
strategic investment
announced 2025-08-01
: 10000000 USD
; investors: Amber Group
; source: GlobeNewswire / GAIB press release
; https://www.globenewswire.com/news-release/2025/08/01/3125672/0/en/gaib-secures-10m-to-accelerate-ai-infrastructure-growth-in-strategic-investment-round-led-by-amber-group.html
-
pre-seed
announced 2024-12-01
: 5000000 USD
; investors: Hack VC, Faction, Hashed
; source: GAIB Blog
; https://blog.gaib.ai/gaib-raises-5m-pre-seed-to-build-the-first-economic-layer-for-ai-compute/
Company timeline
-
2026-06-05:
Gonka opportunity opened to the market
(gtm)
. GAIB moved a previously limited collaboration with Gonka into open-market access for GNK rewards and mining resources.
Source: GAIB Blog - GAIB × Gonka: GNK Rewards Now Open.
https://blog.gaib.ai/
-
2026-06-05:
GNK rewards opened with Gonka collaboration
(product)
. GAIB opened access to GNK rewards after working with Gonka since late 2025 to deploy mining resources into the network.
Source: GAIB blog.
https://blog.gaib.ai/gaib-x-gonka-gnk-rewards-now-open
-
2026-03-05:
Yield Boost added a second rewards layer
(gtm)
. GAIB introduced monthly $GAIB rewards on top of sAID's native yield, with eligibility tied to holding AID, sAID, or Pendle assets.
Source: GAIB Blog - Introducing Yield Boost.
https://blog.gaib.ai/introducing-yield-boost-earn-monthly-gaib-on-top-of-saids-native-yield/
-
2026-03-05:
Yield Boost added on top of sAID yield
(product)
. GAIB introduced Yield Boost, a monthly $GAIB reward program layered on top of native sAID yield with real-time calculations and optional lockups.
Source: GAIB blog.
https://blog.gaib.ai/introducing-yield-boost-earn-monthly-gaib-on-top-of-saids-native-yield/
-
2025-11-21:
AID Alpha migration finalized with reserves and unstaking features
(product)
. GAIB published the November 21 upgrade, adding proof-of-reserves, restricting AID Alpha redemptions to migration, and enabling sAID unstaking with a 30-day cooldown.
Source: GAIB blog.
https://blog.gaib.ai/aid-and-said-november-21st-upgrade/
-
2025-10-31:
AID and sAID mainnet launched
(gtm)
. GAIB launched AID and sAID on Ethereum mainnet and multiple chains, converting the pre-deposit campaign into live products with an onchain distribution footprint.
Source: GAIB Blog - The AI Dollar is live.
https://blog.gaib.ai/the-ai-dollar-is-live/
