AI neolab Flapping Airplanes raises $1 billion at $6 billion valuation
Archived — this story has rotated out of today’s deck. It is kept here in full.
The gist
San Francisco AI startup Flapping Airplanes raised $1 billion in a Series A at a $6 billion valuation.
The one-year-old lab has no product on the market, betting efficiency beats raw scale.
Background
Flapping Airplanes is a San Francisco-based AI research startup founded in late 2025, one of a growing group of "AI neolabs" that focus on research rather than racing to launch products and revenue. It works on training AI models with less data and less compute, and on making the chips that run models more efficient. The raise lands as investors scrutinize the huge sums being spent on chips and data centers for AI, and as demand for AI is seen outrunning the supply of compute and data.
How it unfolded
- late 2025Flapping Airplanes is founded in San Francisco as an AI research startup.
- Jan 2026A seed round of $180 million from GV, Sequoia and Index values the company at $1.5 billion.
- Oct 21, 2026The company announces a $1 billion Series A at a $6 billion valuation, led by Index Ventures and Lux Capital, with Sequoia, Alphabet's GV, Kleiner Perkins, Nvidia and others participating.
- Oct 22, 2026Reports describe the company as a one-year-old AI neolab with no commercial product on the market; investors have now put $1.18 billion into it.
Who’s saying what
- Bulls
- Co-founder Asher Spector argues AI demand will outpace chip and data supply, so the winner will be whoever squeezes the most out of the inputs.
- Investors
- Index Ventures' Mark Xu and Shardul Shah say AI demand will keep outrunning compute and data supply, naming robotics, trading and life sciences as markets needing more efficient AI.
- Caution
- Open Magazine notes the megakernel claims are investors' descriptions of the work rather than an independent assessment, and that reproducible performance and comparisons with credible alternatives are still needed.
Still unverified
The $6 billion valuation is described as "reported" by Bloomberg in some accounts. Open Magazine says the megakernel performance claims come from investors rather than independent assessment and still need reproducible results and comparisons with credible alternatives.