Running models (inference)

Makes the GPUs that most AI inference runs on, and paid about $20B in late 2025 to license Groq's inference chip technology.

$89.0B data centre revenue in the quarter to July 2026 (fiscal Q2 2027), out of $96.2B total

Designs its own inference chips and runs a cloud that serves open models quickly.

$3.5B in a 2026 round that followed its Nvidia licensing deal

Makes very large AI chips and now sells hosted model access as well as the hardware itself.

$510M in 2025, with $87.9M net income

Rents Nvidia GPU cloud capacity to AI labs and companies, and sells GPU servers.

About $760M annualised in 2025 (Sacra estimate)

fast open-model hosting

Company says more than $1B annual recurring revenue (Jul 2026). Sacra estimates about $800M annualised by May 2026

open-model hosting

About $1B annualised (Sacra estimate, Feb 2026)

Hosts the main library of open AI models and datasets, and sells inference, storage and subscriptions to companies.

Annualised revenue above $150M, from one unofficial report. The company does not disclose revenue

pay-as-you-go GPUs

About $300M annualised (CEO, 2026, reported)

Lets developers run AI models through an API. Now part of Cloudflare, feeding its Workers AI platform.

Bought by Cloudflare for $57.4M cash consideration, completed 1 Dec 2025

runs production models for companies

About $600M annualised run-rate, reported in mid 2026 (not audited)

Hosts open AI models through an API at low prices and is building its own data centre capacity.

Company says annual recurring revenue above $100M (2026). Not audited

Serves image, video and audio models through an API for app builders.

Reported to have doubled from $200M to $400M annualised in about five months (one outlet, 2026)

Makes its own AI chips and sells inference cloud access and on-premise systems to large organisations.

About $11B after a round of about $1B (July 2026)

marketplace of hundreds of models behind one API

Sacra estimate of about $50M annualised in early 2026. Company does not publish revenue

Runs a global network and an AI Gateway that routes and caches AI calls for apps, and bought Replicate in Dec 2025.

Hosts web apps and runs AI Gateway, one endpoint for many model providers with no token markup.

About $340M run-rate (Apr 2026, reported)

AI gateway with monitoring and guardrails

No figure found

Open-source logging and gateway for AI calls. Now part of Mintlify and in maintenance mode.

Last disclosed raise of $1.5M (Apr 2023, per CB Insights). Acquisition price not disclosed

open-source router with a paid enterprise tier

Seed round co-led by Y Combinator (reported)

Model router that picks the best model for each request to balance answer quality against cost.

$9M seed led by NEA (Nov 2023). No 2026 funding found

Model router that predicts which AI model will give the best answer for a prompt at the lowest cost.

About $2.3M raised in 2024, plus an undisclosed 2025 round from strategic investors including IBM and SAP.iO

Sells NIM microservices and NVIDIA AI Enterprise software, which run AI models inside a company's own data centre.

$89.0B data centre revenue in the quarter to July 2026 (fiscal Q2 2027), out of $96.2B total

Hosts private model deployments inside a customer's Google Cloud account, and runs its own Gemini models.

25% of AI chip supply (Epoch AI)

Hosts private model deployments inside a customer's Azure account, and sells its own Copilot and model services.

20M+ users (2025)

Hosts private and dedicated copies of AI models inside a customer's own AWS account, through Bedrock and related services.

5% of AI chip supply (Epoch AI)

sells on-premise deployments

About $14B valuation (Sep 2025)

Sells AI servers and the Dell AI Factory package, built with Nvidia chips, for companies that run AI in their own sites.

Full-year fiscal 2027 guidance raised to $192B (Sep 2026)

Canadian company that builds enterprise AI models and mostly sells them for private or on-premise use.

About $240M annualised recurring revenue at end 2025 (reported)

Desktop app that runs open AI models on your own computer. Free for work and personal use since July 2025.

About $1.8M annual revenue in 2025 (reported)

the popular tool for running models on a laptop or server

Not disclosed. Company declined to discuss revenue at its Series B

Sells Private Cloud AI, a bundled on-premise AI system with Nvidia chips, for companies that keep data in-house.

Sells OpenShift AI, a platform for running and managing AI models on a company's own servers or cloud. Owned by IBM.

Networking giant that agreed to buy Galileo in April 2026 to add AI monitoring to its Splunk observability products.

About $4B of AI infrastructure revenue in fiscal 2026 (year to July 2026)

open-source monitoring

Bought by ClickHouse (Jan 2026)

Runs a fast open-source database used for analytics. Bought Langfuse to add AI tracing to its stack.

Completed a $400M Series D alongside the Langfuse deal (Jan 2026). Valuation not in the source

Tests AI models for hallucinations, errors and safety problems, and monitors them once they are live.

Total funding about $68M. Its $45M Series B was led by Scale Venture Partners (dated Oct 2024 by CB Insights)

tracing for LangChain apps

LangSmith annual recurring revenue reported at $12M to $16M (TechCrunch, 2025). Not confirmed by the company

monitoring

$70M Series C (2025)

testing and monitoring

$80M at about $800M valuation (Feb 2026)

Evaluates LLMs and AI agents for accuracy and safety, and builds simulated worlds to test agents.

Company says revenue grew more than 15 times in the past year. Not audited