Flapping Airplanes, an artificial intelligence startup founded just nine months ago, announced on Wednesday that it has raised $1 billion in a Series A funding round. The investment values the company at $6 billion, a significant jump from its $1.5 billion seed valuation earlier this year.
The round was co-led by Index Ventures and Lux Capital. Other participants included existing backers Sequoia Capital, GV, and Kleiner Perkins. Additional investors in the latest round are Valor Equity Partners, Atreides, A*, and Nvidia.
Founding Team and Background
The company was established by Aidan Smith and brothers Asher and Ben Spector. Smith is a former founder and Thiel Fellow who worked at Neuralink during his college years. Asher Spector holds a PhD in statistics from Stanford University. Ben Spector is a computer science PhD student at Stanford and previously founded Prod while attending MIT.
Prod is a student-run startup accelerator at MIT that has gained a reputation as a talent incubator in Silicon Valley. Sequoia partner David Cahn, an investor in Flapping Airplanes, described the founders as building a laboratory for artificial general intelligence tailored for younger developers.
Technical Approach to Efficiency
Flapping Airplanes is developing what it calls a megakernel framework. The company describes this technology as a system capable of performing entire training runs, serving inference, and playing video games simultaneously within a single kernel. A kernel is a core program that runs on a graphics processing unit, or GPU.
The founders claim their framework can overlap memory-bound inference tasks with compute-bound training tasks. This allows the chip to process data fetching and mathematical calculations concurrently, reducing idle time. The underlying premise is that current AI systems use far more data than necessary and that GPUs are not being utilized to their full potential.
Asher Spector stated shortly after the company launched that a model requiring a million times less data would be significantly easier to integrate into the broader economy. The goal is to create intelligent systems that are cheaper to train and operate.
The new funding will allow the company to expand its workforce and continue development of its training infrastructure. The startup previously raised $180 million in seed funding from GV, Sequoia, and Index.
Source: TechCrunch
