Flapping Airplanes is a San Francisco research lab working on training and running models with less data and less compute. Its stated goal is to open markets such as robotics and scientific discovery, where models are still too expensive or training data is scarce, by improving the unit economics of AI. On October 21, 2026 the company announced a $1 billion Series A co-led by Index Ventures and Lux Capital, nine months after a $180 million seed. Bloomberg reported a $6 billion valuation, four times the $1.5 billion valuation on that seed. The lab has not released a commercial product. Most of the past year went into new GPU primitives. The company says it now has a megakernel framework that can run entire training runs, serve inference, and play video games at the same time inside a single GPU program, and that the same program can overlap memory-bound inference with compute-bound training. It also says those primitives make unusual architectures and optimizers practical to run, and frames the work as co-evolving models with the software that drives the chips. The lab was founded by brothers Ben and Asher Spector and Aidan Smith. Ben, the CEO, was a Stanford computer science PhD student, started the student accelerator Prod while at MIT, and was lead author on ThunderKittens, a 2024 open-source toolkit for writing fast GPU kernels. Asher holds a statistics doctorate from Stanford, and Smith is a Thiel Fellow who worked at Neuralink during college. Index says a couple dozen people work in person in San Francisco, across kernels, pre-training, post-training, and inference, and that partner Mark Xu is joining the board. Sequoia and GV returned for this round, and Nvidia, Kleiner Perkins, and others are also investors.