Skip to content
TrustList
Funding news

Flapping Airplanes raises a $1bn Series A for data-efficient AI models

Editorial

By TrustList Editorial

The young AI lab says it spent a year building new GPU primitives to cut how much data models need. Index and Lux co-lead, nine months after a $180 million seed.

About Flapping Airplanes raises a $1bn Series A for data-efficient AI models

Flapping Airplanes raises a $1bn Series A for data-efficient AI models

21 October 2026: Flapping Airplanes, an AI research lab founded this year, says it has raised a $1 billion Series A co-led by Index Ventures and Lux Capital. In its announcement the company says it has spent the past year "developing new primitives for GPUs" which it believes will enable new learning algorithms and "dramatically reshape AI's cost structure".

Not yet independently verified. The company's own post states only the amount and the two co-leads; the other investors, the reported $6 billion valuation and the founders' background come from press reporting. We will update this when it can be confirmed, and remove this note.

Who backed it

The company names only its two co-leads. TechCrunch, reporting the round the same day, lists Sequoia, GV and Kleiner Perkins among the other participants and Valor, Atreides, A* and NVIDIA as investors, and cites Bloomberg for a valuation of about $6 billion. That would be four times the $1.5 billion valuation of its $180 million seed round from GV, Sequoia and Index in January.

Less data, not more compute

The lab's bet runs against the industry's habit of scaling models on ever larger data sets. Its founders, Aidan Smith, Asher Spector and Ben Spector, want models that learn from far less data and are cheaper to train and run. "A model that's a million times more data efficient is probably a million times easier to put into the economy," Asher Spector told TechCrunch when the company launched. The technical claim behind it is a framework of GPU "megakernels" that, in the founders' description, overlaps memory-bound inference with compute-bound training so that the chips stay busy.

Smith previously founded a company and worked at Neuralink as a student; Asher Spector was a statistics PhD student at Stanford; Ben Spector, a computer science PhD student at Stanford, founded the student incubator Prod at MIT.

What it means for buyers of AI

Nothing is on sale yet. The company has not said when it will release a model or a product, or where it is based. For companies buying AI today, the relevance is the direction of travel: a lab raising this much on data efficiency is a bet that training costs, and with them model prices, can fall by a different route than bigger clusters. Our artificial intelligence company category lists more than 1,000 firms, most of them building on models rather than training their own.

Sources

Categories & features

TrustList Weekly

The week in software and IT, in one email

The news that matters to buyers, new rankings and our own research. Every Thursday, free, and easy to leave.

We will email you to confirm. Unsubscribe with one click in any issue. Privacy policy