• Live
  • 65 verified
  • $2.48M MRR
  • 58 ranked
  • Updated Nov 5
  • Revenue verified via Stripe on TrustMRR

Home / Blog

Applied Compute: $100M a year in under 18 months (reported, not verified)

Data checked Nov 5, 2026 · How we verify

Short answer

Applied Compute says it makes $100 million a year (annualized run rate), less than 18 months after it was founded, and it raised $350M at a $3.25B valuation the same day (Applied Compute blog, Nov 2, 2026). Seven months earlier it was valued at $1.3B. Reported, not verified

Small AI businesses with real Stripe numbers, Mon/Wed/Fri.

Free. Plus the full spreadsheet of every verified business now.

Your CSV: 65 verified AI businesses. Download it · Preview

Coming soon · free newsletter

Revenue is the easy part. Profit is coming soon.

Real profit and margins for verified AI businesses, where founders published them. Free newsletter. Be first when they land.

Subscribe free to get the profit numbers first →

You're on the list for the profit numbers. See what's coming →

Nothing on this page is processor-verified. Applied Compute is private, so these are the company's own posts and press reports, each linked with its date. We did not estimate any revenue number; where we do math on reported figures, we say so. For AI businesses whose payment data we can check, see the verified AI revenue leaderboard.

How much revenue does Applied Compute make?

Applied Compute trains custom AI models for big companies and then runs them. On Nov 2, 2026 its three founders wrote: "Less than 18 months after founding Applied Compute, we've grown from zero to $100 million in annualized revenue run rate" (company blog). That is the only revenue figure it has published.

WhenReported figureSource
Jun 2025Pre-launch round of $20M led by Benchmark's Victor Lazarte, at a $100M valuation (sources to Upstarts)Upstarts Media, Jun 27, 2025
Oct 2025Launches publicly with $80M raised in total (Benchmark, Sequoia, Lux); customers named: DoorDash, Cognition, MercorSiliconANGLE, Oct 30, 2025
Apr 15, 2026$80M round led by Kleiner Perkins at a $1.3B post-money valuation; $160M raised in totalLatham & Watkins (company counsel); Kleiner Perkins, May 14, 2026
Nov 2, 2026$100M annualized revenue run rate, "less than 18 months after founding"; raises $350M at a $3.25B valuation, led by Elad GilApplied Compute blog (Yash Patil, Rhythm Garg, Linden Li)

Run rate = the current pace of revenue multiplied out over 12 months. It is not revenue already collected, and Applied Compute did not say what period it is based on, how much is recurring, or whether it is profitable. We found no earlier revenue figure from the company.

How fast is Applied Compute growing?

By its own account, from zero to $100M in under 18 months. The other markers it gave on Nov 2, 2026 (all company-reported):

  • More than doubled the number of enterprises it serves since its previous round (April 2026).
  • Customers have run more than 1,700 training runs on its AC2 platform since July.
  • AC Inference, launched about 12 weeks earlier, serves more than 1 trillion tokens a day.
  • 40% of customers use it for both training and inference.
  • The team has tripled since January; 85% are researchers and engineers.

Our math on the valuations: $100M (Jun 2025) to $1.3B (Apr 2026) to $3.25B (Nov 2026), about 32x in 17 months.

How much is Applied Compute worth?

ValuationRoundSource
$100M$20M pre-launch round led by Benchmark (Jun 2025, per sources)Upstarts Media, Jun 27, 2025
$1.3B$80M, led by Kleiner Perkins, post-money (Apr 2026); $160M raised in totalLatham & Watkins, Apr 15, 2026; Kleiner Perkins
$3.25B$350M, led by Elad Gil, with Kleiner Perkins, Lux, NVIDIA, Blackstone, Shopify Ventures, HubSpot Ventures and others (Nov 2026)Applied Compute blog, Nov 2, 2026

Our math: about $510M raised in total ($160M before + $350M now), and $3.25B is about 32x the reported $100M run rate. The company did not say whether $3.25B is pre- or post-money.

How does Applied Compute make money?

It sells companies their own AI. Instead of renting a general model from OpenAI or Anthropic, a customer gives Applied Compute its data and workflows, and Applied Compute post-trains an open-weight model for that one job, then hosts it. It charges for the training work and for inference (running the model). It publishes no prices.

Its pitch is cost and control. Two examples it gives (company-reported): a model for Harvey's Review Table that improved answer and citation quality "at less than half the inference cost of a frontier alternative," and a custom router for Clay that cut production inference costs "by up to 47%" (Nov 2, 2026).

Who are Applied Compute's customers?

Large companies and well-funded AI startups. It names Microsoft, Shopify, Mercor, Harvey, DoorDash, Cognition, Clay and Decagon, across life sciences, cybersecurity, hardware design, legal, financial services and logistics (Nov 2, 2026). Several are AI companies themselves: Mercor and Cognition both have revenue breakdowns on this site.

What does Applied Compute mean for making money with AI?

It is a "picks and shovels" business: it gets paid when other companies build with AI, whoever wins. You don't need $350M to sell to AI builders. Small businesses doing the same thing at a much smaller scale have Stripe-verified revenue in our data:

BusinessVerified revenueChecked
Apiframe
One API for AI image, video and music models, sold to developers.
$35.4k MRRTrustMRR, Nov 5 2026
Audn.AI
Autonomous AI "ethical hacker" that pentests AI agents and apps around the clock, sold to teams shipping AI and to security teams.
$29.3k MRRTrustMRR, Nov 5 2026
SocialCrawl
One API key for public data from 69 social platforms, built for AI agents and developers and sold as pay-per-call credits, not subscriptions.
$45.2k / 30 daysTrustMRR, Nov 5 2026

All five indie developer tools in our data, compared: how much AI developer tools make. More: B2B AI SaaS ideas that charge $100+ a month, how to automate your business with AI and profitable AI startups.

How does Applied Compute compare with other AI companies?

All reported, not verified: fal, which runs AI image and video models for developers, reached about $800M in annualized revenue in 2026 (Bloomberg); Sierra passed $400M ARR in October 2026; Meshy passed $100M ARR in September 2026, up from $1M at the end of 2024; Cognition passed $1B a year. Applied Compute says it got to $100M less than 18 months after founding. All of them side by side: how much do AI companies make?

More AI revenue breakdowns

Big AI companies: OpenAI, Anthropic, Cursor. Small AI businesses with Stripe-verified revenue: GoMotion AI, Glossa (live AI translation for churches) and the October 2026 AI revenue report.

Compare every AI company we cover: how much do AI companies make? (verified and reported revenue, side by side).

FAQ

How much does Applied Compute make?

Applied Compute says it reached a $100 million annualized revenue run rate on Nov 2, 2026, less than 18 months after it was founded (Applied Compute blog). A run rate is the current pace multiplied out over a year, not revenue already collected. Reported by the company, not verified: Applied Compute is private and publishes no financial statements.

What is Applied Compute valued at?

$3.25 billion, after a $350 million round led by Elad Gil announced Nov 2, 2026. In April 2026 it was valued at $1.3 billion post-money (Kleiner Perkins-led $80M round), and in June 2025 at $100 million (pre-launch round, per Upstarts Media sources). Our math: $3.25B is about 32x its reported $100M run rate.

What does Applied Compute do?

It trains and runs custom AI models for companies on their own data, mostly by post-training open-weight models. Its AC2 platform covers training, inference and continual improvement; it says AC Inference serves more than 1 trillion tokens a day and 40% of customers use it for both training and inference (company blog, Nov 2, 2026).

Who are Applied Compute's customers?

The company names Microsoft, Shopify, Mercor, Harvey, DoorDash, Cognition, Clay and Decagon, and says it more than doubled the number of enterprises it serves since its previous round (Nov 2, 2026). It reports a model for Harvey at less than half the inference cost of a frontier alternative, and a router for Clay that cut inference costs by up to 47% (company-reported).

Who founded Applied Compute?

Yash Patil (CEO), Rhythm Garg and Linden Li, three former OpenAI researchers who met at Stanford. Patil worked on Codex, Garg on the o1 reasoning model and Li on training infrastructure (SiliconANGLE, Oct 30, 2025).

Can a small business make money selling tools to AI builders?

Yes, at a much smaller scale. The AI Receipts tracks small "picks and shovels" businesses with Stripe-verified revenue: Apiframe (an API for AI image and video models), Audn.AI (security testing for AI agents) and SocialCrawl (social media data for AI agents). See their case studies for current figures.

Keep reading