A Billion-Dollar Series A and the Back Office Both Point the Same Way

Efficiency as a billion-dollar bet

Flapping Airplanes, a San Francisco-based foundational AI research lab, announced a 1 billion dollar Series A round in October 2026 led by Index Ventures, with Mark Xu joining the board. The company aims to make AI orders of magnitude more efficient by developing new primitives for GPUs, including megakernels that can run entire training runs in a single kernel, alongside models designed to learn more from far less data. m:…1762

Founder Ben Spector emphasizes holding research to strong baselines, a detail that matters because an efficiency claim at this scale will be judged against existing systems rather than against a straw man. The investment thesis is therefore testable in a way many application bets are not: either the primitives run training materially more efficiently, or they do not, and the baseline discipline is the mechanism that keeps the claim honest. m:…1762

Drug programs, not drug companies

Formation Bio announced a 600 million dollar Drug Development Accelerator Fund that invests in individual drug programs rather than standalone biotech companies, with partner programs receiving upfront payments while retaining participation in future milestones and royalties. Funded NewCos receive money to advance drug assets through clinical milestones and value inflection points, targeting assets 2 to 4 years away from a clinical readout. m:…4601

The backdrop the firm cites is a pipeline that discovers more but approves at a flat rate: discovered drug candidates have doubled over the past decade while annual approvals stayed roughly flat, and more than 300 billion dollars in drug sales face patent expiration over the next decade. The fund structure is the argument: if the bottleneck is advancing individual assets rather than founding companies, then capital should be organized around assets. m:…4601

The back office, automated and attacked

Emergence Capital reports Casap raised a 56 million dollar Series B after growing revenue 10 times in one year and now working with more than 100 banks, credit unions, fintechs and processors. The company says its agents resolve 73 percent of dispute cases autonomously in 8 days, against manual resolution that previously took up to 90 days and 45 minutes of human time per case, now reduced to 1 to 2 minutes, with a 96 percent chargeback win rate against a 60 percent industry average. m:…0104

The same automation wave creates its own adversary: Snappt, a California company helping multifamily operators fight fraud, in a market where bad actors increasingly use AI to forge rental application documents, and fraudulent tenants may refuse to pay rent or vacate after gaining access. 93 percent of multifamily operators have experienced rental fraud, and Emergence notes first-party fraud rose from 15 percent of reported fraud two years ago to 36 percent now, calling fraud an AI versus AI problem. The funding pattern is therefore double-sided: money flows to the automation, and then to the detection of automated abuse. m:…3078 m:…0104

Where the money goes next

The follow-on moves are already visible in the same releases: Casap is working with Synchrony to expand agents into more bank workflows after proving dispute resolution, and Formation fund assets are targeted 2 to 4 years from a clinical readout, with funded programs advancing through milestones toward value inflection points. That pattern favors investors who underwrite a second act, not just a first deployment, because the dispute platform expands across workflows and the drug fund holds assets until a readout reprices them. m:…0104 m:…4601