Flapping Airplanes, a one-year-old startup aiming to train large models with less data, announced on Wednesday that it closed a $1 billion Series A round co-led by Index and Lux at a reported $6 billion valuation. Existing backers Sequoia, GV and Kleiner Perkins participated in the round; other investors include Valor, Atreides, A*, and Nvidia.
Rapid value increase after launch
This Series A arrives just nine months after the company launched with a $180 million seed round led by GV, Sequoia, and Index. That seed round had valued the company at $1.5 billion, meaning the reported valuation has quadrupled since the seed.
Founders and background
Flapping Airplanes was founded by Aidan Smith and brothers Asher and Ben Spector. Aidan Smith is a former founder and Thiel fellow who worked at Neuralink during college. Asher Spector completed a PhD in statistics at Stanford; Ben Spector was a computer science PhD student at Stanford and earlier founded Prod while at MIT.
Prod is a nearly secretive, student-run accelerator that has been operating for about five years; Sequoia partner David Cahn, a Flapping Airplanes investor, described Prod as “one of the most powerful talent incubators in Silicon Valley” and said the founders together are building what he called the “young person’s AGI lab.”
Technology and ambitions
The founders describe their work as “the world’s most complete megakernel framework,” which they say can perform full training runs, serve inference, run video games, and even do all three simultaneously within a single kernel. (A kernel here refers to a program that runs on a GPU; a megakernel bundles multiple tasks into one program.)
They claim the framework can overlap memory-bound inference with compute-bound training, allowing tasks that wait for data and tasks that perform heavy computation to run together so the chip is seldom idle. Their premise is that AI today has barely scratched the surface of GPU potential, and they argue this approach could enable intelligent systems “without ingesting half the internet.”
The stated goal is to build systems that cost less to train and operate. Asher Spector told TechCrunch shortly after the company launched: “A model that’s a million times more data efficient is probably a million times easier to put into the economy.”
Next steps with the new funding
With this fresh $1 billion, Flapping Airplanes plans to expand its team and continue developing the megakernel framework, attempting to turn its data-efficiency claims into practical, deployable systems. The financing brings together the company’s early backers as it scales toward commercial applications.
