Source: Gimlet Labs raised its $300M round after telling investors OpenAI may spend $100M+ annually on its services; OpenAI says it's not a paying customer yet
THE SO WHAT
Valuations are now trading on hypothetical hyperscaler contracts, not booked revenue — Gimlet’s $300M raise at ~$3B on the back of a potential $100M+/year OpenAI customer that doesn’t yet exist is a clean example. If you’re selling into labs or clouds, expect diligence to shift from current ARR to pipeline credibility and technical indispensability.
READ THE SOURCE
MORE FROM THE WIRE
Startups & VentureChina’s AI-Fueled IPO Boom Delivers Six-figure Windfalls to Retail Investors
China’s AI-linked IPOs throwing off six-figure day-one gains for retail investors is a classic late-cycle signal—capital is chasing narrative more than cash flow. If you’re operating in or around this market, treat current valuations and demand as fragile and plan for a sharper repricing window than your board deck assumes.
Startups & VentureChime agrees to acquire longtime banking partner Stride Bank for $590M; Stride will become Chime Bank, a wholly owned subsidiary; CHYM jumps 8.5%+ after hours
Chime paying $590M to buy its partner Stride and become a bank underscores how expensive it is to outsource your regulatory license. Any fintech at scale should be modeling the crossover point where owning a charter beats partner-bank fees and constraints.
Startups & VentureCognition hits $48B valuation, signaling investors believe AI coding is far from a winner-take-all market
A $48B mark for Cognition—above Cursor’s pre-acquisition multiple—says investors see AI coding as a multi-player category with room for differentiated workflows, not just one dominant assistant. If you build dev tools, assume your customers will run multiple AI coding surfaces and compete on integration depth, governance, and team-level adoption, not just raw model quality.
Startups & VentureAntioch, which creates high-fidelity simulations to reduce the need for hardware validation in physical AI training, raised a $32M Series A led by Greylock
Capital is flowing into sim-first approaches for physical AI—Antioch’s $32M bet is that high-fidelity virtual worlds can substitute for billion-dollar real-world data collection. If you’re building robots or physical agents, your cost curve may be decided by your simulation stack as much as your hardware.