Flapping Airplanes, a year-old artificial intelligence startup, has secured $1 billion in Series A funding, catapulting the company to a reported $6 billion valuation. The massive capital injection comes just nine months after the company emerged from stealth with a $180 million seed round, effectively quadrupling its valuation in less than a year.
The Series A round was co-led by Index Ventures and Lux Capital. A wide roster of prominent Silicon Valley investors also participated, including earlier backers GV, Sequoia Capital, and Kleiner Perkins. Other investors in the round include Valor, Atreides, A*, and the chip-making giant Nvidia.
Founded by Aidan Smith and brothers Asher and Ben Spector, the company has quickly become one of the most watched entities in the race toward artificial general intelligence (AGI). Smith is a former Thiel fellow who previously worked at the brain-computer interface firm Neuralink. Asher Spector was a statistics doctoral student at Stanford, while Ben Spector, who also studied at Stanford and MIT, founded Prod, a highly regarded student-run startup accelerator.
Flapping Airplanes is positioning itself as a leaner alternative to the data-heavy approaches currently dominating the industry. While industry leaders often rely on scraping vast swaths of the internet to train large language models, Flapping Airplanes is focused on data efficiency. The founders have expressed a goal of creating intelligent systems that do not require “ingesting half the internet” to function effectively.
At the core of the company’s strategy is what the founders describe as a “megakernel framework.” In high-performance computing, a kernel is a program designed to run on a Graphics Processing Unit (GPU). The Flapping Airplanes framework reportedly bundles multiple tasks into a single program, allowing a single kernel to perform training runs, serve inference, and even run complex simulations or video games simultaneously.
This technical approach is designed to solve a persistent bottleneck in AI development: GPU idle time. By overlapping memory-bound tasks, such as data retrieval for inference, with compute-bound tasks, such as mathematical calculations for training, the company claims it can maximize the utility of expensive hardware. The ultimate objective is to drive down the cost of both training and running AI models, making them more viable for broad economic integration.
“A model that’s a million times more data efficient is probably a million times easier to put into the economy,” Asher Spector noted during the company’s early stages.
Industry observers have characterized the startup as a “young person’s AGI lab,” noting the founders’ deep ties to elite academic and talent incubation networks. Much of this reputation stems from the founders’ involvement with Prod, which has gained a reputation in Silicon Valley as a powerful talent incubator for young engineers and researchers.
With the fresh billion-dollar investment, Flapping Airplanes plans to significantly expand its team as it seeks to turn its high-efficiency framework into a commercial reality. The company’s rapid ascent reflects the continued appetite among venture capitalists for foundational AI research that promises to bypass the immense data and hardware costs associated with current industry standards.
