Five startups that defined Silicon Valley's last two years — and why crypto barely made the list

Over the past two years, five companies have emerged from Silicon Valley with a combined valuation today approaching $100 billion. All five are about artificial intelligence. Not one is a crypto project. That's not a coincidence — to understand where the Valley's money and attention actually went during this window, it's worth breaking down each company individually, then honestly looking at what happened to crypto startups from the same region in the same period.

The selection criteria are simple: a company founded in the Bay Area (San Francisco, Palo Alto, Stanford and surroundings) roughly between mid-2024 and mid-2026, with a publicly known valuation and on-record statements from its founders about their mission.

Safe Superintelligence — $32B and Not a Single Product

Safe Superintelligence Inc. (SSI) was founded in June 2024 by Ilya Sutskever — OpenAI's former chief scientist, who left the company a month earlier — along with Daniel Gross (former head of AI at Apple) and Daniel Levy (a former OpenAI researcher). Its headquarters are split between Palo Alto and Tel Aviv, with roughly 20-30 employees.

Valuation trajectory:

• September 2024 — $1B raised at a $5B valuation
• April 2025 — another $2B at a $32B valuation

SSI has raised roughly $6B in total since founding — while still having no commercially available product.

“one goal and one product: a safe superintelligence”

— Ilya Sutskever, TechCrunch, April 2025

Quote source: TechCrunch, April 12, 2025

Thinking Machines Lab — $12B After a Failed Run at $60B

Thinking Machines Lab was founded in February 2025 by Mira Murati, OpenAI's former CTO, who left the company in 2024. Headquartered in San Francisco, the company has around 140-169 employees.

Its valuation trajectory is a rare case where the number didn't just go up:

• July 2025 — $2B seed round at a $12B valuation (led by a16z, with Nvidia, Accel, ServiceNow, Cisco, AMD and Jane Street participating)
• November 2025 — talks of a new round at a $50-60B valuation
• January 2026 — the talks collapsed without a deal
• Mid-2026 — the company settled back at its original $12B valuation, with two shipped products and an NVIDIA partnership for one gigawatt of Vera Rubin compute capacity (signed March 2026)

Cognition AI — From $10B to $26B in 8 Months

Cognition AI, the maker of autonomous AI software engineer Devin, was founded in late 2023 by Scott Wu, Steven Hao and Walden Yan — all three former competitive programmers and gold medalists at the International Olympiad in Informatics. It's headquartered in San Francisco.

• September 2025 — $400M round at a $10.2B valuation
• May 2026 — over $1B at a $26B valuation (led by Lux Capital, General Catalyst and 8VC)

Cognition's annualized revenue has reached $492M, with customers including Mercedes-Benz, NASA, Goldman Sachs and Santander. Meanwhile, 89% of all code committed inside Cognition itself is now written by its own product, Devin.

“We've never thought about it as replacing humans. I know it's like a scenario, folks have said these things. It has never been our view. We are all programmers ourselves. I started coding when I was nine.”

— Scott Wu, TechCrunch, May 2026

Quote source: TechCrunch, May 29, 2026

Sierra — From $4.5B to $15.8B in Eighteen Months

Sierra was founded in 2023 by Bret Taylor (former Salesforce co-CEO and OpenAI board chairman) and Clay Bavor (a former senior Google executive). The company builds AI agents for customer service, headquartered in San Francisco.

• October 2024 — $175M at a $4.5B valuation
• Fall 2025 — valuation climbed to $10B
• May 2026 — a $950M Series E at a $15.8B valuation (led by Tiger Global and Google's GV)

The company's annualized revenue passed $150M in under two years since founding, with customers including Prudential, Cigna, Blue Cross Blue Shield and Rocket Mortgage. By Sierra's own account, it now serves more than 40% of the Fortune 50.

“Most companies don't want to make software. They want solutions to their problems.”

— Bret Taylor, TechCrunch, April 2026

Quote source: TechCrunch, April 9, 2026

Physical Intelligence — A Robot Version of ChatGPT, Valued at $11B

Physical Intelligence was founded in 2024 by Karol Hausman, Sergey Levine (a UC Berkeley professor) and Chelsea Finn (a Stanford professor) — coming out of academia and Google DeepMind. The company builds foundation AI models for robots, headquartered in San Francisco.

• Prior round — $600M at a $5.6B valuation
• March 2026 — talks of a $1B round at a valuation above $11B

The company has raised roughly $1.6B across two rounds in total.

“Think of it like ChatGPT, but for robots.”

— Sergey Levine, TechCrunch, office visit

Quote source: TechCrunch, April 16, 2026

What About Crypto?

Applying the same criteria — Bay Area, founded in the last two years, publicly known valuation — to crypto startups produces a noticeably smaller list:

• MegaETH — an Ethereum layer-2 network, founded in March 2024 by Shuyao Kong (formerly of ConsenSys) and Yilong Li (a Stanford CS PhD), headquartered near Stanford. Mainnet launched in February 2026; its token opened at a $1.56B fully diluted valuation, which had slipped to around $486M by mid-2026
• Mezo — a Bitcoin-backed lending platform, founded in 2024 by Matt Luongo, headquartered in Menlo Park. Raised a $21M Series A from Pantera Capital; its MEZO token, launched in January 2026, was trading at roughly a $1-1.4M market cap by mid-year — orders of magnitude below any of the AI startups above
• World Network (Tools for Humanity) — Sam Altman's iris-scanning digital-identity project, valued at $2.5B. It technically doesn't fit the window — founded back in 2019 as Worldcoin — but it's telling that by mid-2026 the company is cutting staff

The single largest crypto-native launch of the past two years, the Monad blockchain, technically isn't from Silicon Valley at all — the company is based in New York, and its MON token trades at a fully diluted valuation of roughly $3.48B.

A similar picture shows up in the news: in summer 2026, Coinbase — one of the few genuinely large Bay Area crypto companies — publicly reported that over 95% of its code is now AI-written — meaning even the region's biggest crypto players spent this period investing attention and resources into AI infrastructure rather than launching new crypto protocols.

What This Means in Practice

The scale gap here isn't a coincidence, and it isn't a sign that crypto startups are worse managed. It reflects where Bay Area venture capital actually went in 2024-2026: building foundation AI models and agents requires enormous compute and upfront capital, and that exact niche was claimed by alumni of OpenAI, DeepMind and Google — people whose names and track records alone open doors to billion-dollar rounds. Crypto startups from the same region had a structurally harder time competing for that same investor attention during this window — and that, more than the quality of any individual project, explains the order-of-magnitude gap in the numbers throughout this piece.

This material is for informational purposes only and is not investment advice. Startup valuations cited here are based on publicly reported funding-round data as of publication and may change.