AI Startups to Watch (Sep 5): Gimlet, Crusoe & FluidStack

AI infrastructure startups just had their biggest week yet. Gimlet Labs hit $3 billion on a multi-chip bet, Crusoe tripled to $30 billion, and FluidStack raised $1.5 billion from Jane Street. The money is flowing to compute, not models.

For more startup coverage, see yesterday's AI Startups roundup and our full AI Startups archive.

Gimlet Labs: $300M to Make AI Work Across Every Chip

Gimlet Labs raised a $300 million Series B led by Andreessen Horowitz, bringing its valuation to $3 billion. New investors include Arm Holdings and Microsoft's M12 venture fund. The round comes just six months after Gimlet raised $80 million in its Series A at a $400 million valuation, making this an unusually fast climb. Source: Bloomberg

Gimlet Labs is an AI inference orchestration company that routes workloads across chips from six different manufacturers: Nvidia, AMD, Intel, Arm, Cerebras, and d-Matrix. The company calls this the industry's first "multi-silicon inference cloud." The core pitch is that AI workloads do not all need to run on Nvidia GPUs. By splitting inference across chip types, customers can reduce costs and avoid supply constraints.

The Arm partnership is strategically significant. Gimlet is building compatibility with multiple Arm chip architectures, potentially giving AI customers a path to lower-cost compute outside the Nvidia ecosystem. The company has raised at least $380 million total. Source: Gimlet Labs Blog

Crusoe Energy: $3 Billion at $30 Billion Valuation

Crusoe closed a $3 billion Series F, tripling its valuation to $30 billion in under nine months. The round was co-led by Atreides Management and Valor Equity Partners, with Mubadala Capital participating. The Denver-based company has now raised $7.2 billion total. Source: Reuters

Crusoe builds data centers powered by renewable energy and provides AI cloud compute to companies like Meta and Oracle. The company signed a $13 billion, five-year contract with Jane Street, the quantitative trading firm, to supply GPUs and infrastructure. Crusoe says its total power pipeline exceeds 40 gigawatts across multiple data center projects, including a $15 billion joint venture with Blue Owl Capital for facilities in Abilene, Texas.

The Jane Street deal signals that demand for AI compute is expanding beyond tech companies into finance, where quantitative trading firms need massive GPU clusters for model training and real-time inference. Source: Silicon Republic

FluidStack: $1.5 Billion From Jane Street at $18 Billion Valuation

FluidStack raised $1.5 billion in a private equity round led by Jane Street Capital, pushing its valuation to $18 billion. The New York-based company provides large-scale GPU and data center infrastructure for demanding AI workloads. This brings FluidStack's total funding to just over $2.6 billion.

FluidStack is part of a growing wave of companies spending heavily to build out GPU infrastructure. The company operates a distributed cloud model, connecting data centers across multiple locations to serve AI workloads. The Jane Street connection, coinciding with Crusoe's separate $13 billion Jane Street deal, underscores how aggressively quantitative finance is moving into AI infrastructure. Source: Crunchbase

Wayve and Uber Launch UK's First Autonomous Rides in London

British AI startup Wayve and Uber launched supervised autonomous rides in London on September 3, marking the first time autonomous trips are available in the United Kingdom. Londoners requesting an UberX, Uber Electric, or Uber Comfort can now be matched with a Wayve vehicle at no extra cost.

The initial fleet covers all of London except airports, with safety drivers behind the wheel. The service will expand to fully driverless (L4) operations as Wayve completes trials under a new UK government fast-track program. Wayve, which raised $1 billion from SoftBank last year, develops end-to-end AI driving models that learn from real-world driving data rather than hand-coded rules. Source: Wayve

Upwind Hits $3.8 Billion on Cloud Security Bet

Cloud security startup Upwind raised a new round that values the company at $3.8 billion, bringing total funding to $730 million. The round included Greylock, Craft Ventures, and Cyberstarts. Upwind builds runtime cloud security that protects AI workloads in production, a category that has grown rapidly as companies deploy more AI systems. Source: Crunchbase

Related reading

Deeper context from the archive on the companies and technologies in this story:

• AI Startups to Watch (Sep 18): Crusoe, Emulate & Arcee
• AI Startups to Watch (Sep 11): Positron, Cognition & Fluidstack
• AI Startups to Watch August 9, 2026
• AI Startups to Watch (Sep 4): Gimlet, Baseten & Rilo

Building an AI agent, automation, or backend workflow? AsaanKaroobar builds custom AI solutions, apps, and digital experiences. Get in touch ->

Frequently Asked Questions

What are the top AI startups to watch today?

The top AI startups to watch on September 5, 2026 are Gimlet Labs ($300M at $3B for multi-chip AI inference), Crusoe Energy ($3B at $30B for AI data centers), FluidStack ($1.5B at $18B for GPU infrastructure), Wayve (UK's first autonomous rides with Uber), and Upwind ($3.8B valuation for cloud security). The overwhelming theme is AI infrastructure and compute, not model development.

Why is Gimlet Labs valued at $3 billion after only six months?

Gimlet Labs raised $300 million in a Series B just six months after its $80M Series A, tripling its valuation from roughly $400 million to $3 billion. The rapid jump reflects strong investor conviction that AI workloads will increasingly run across multiple chip architectures, not just Nvidia GPUs. Gimlet's multi-silicon inference cloud routes workloads across chips from Nvidia, AMD, Intel, Arm, Cerebras, and d-Matrix.

What makes Wayve's London launch significant for autonomous vehicles?

Wayve and Uber launched the United Kingdom's first autonomous ride-hailing service in London on September 3, 2026. The launch is significant because it uses Wayve's end-to-end AI driving models, which learn from real-world data rather than hand-coded rules, and it operates under a new UK government fast-track regulatory program designed to accelerate autonomous vehicle deployment.