Venture Money Backs AI That Does a Job, While the Biggest Value Stays at the Top

Three rounds, three specific jobs

Accel led a $30M Series A funding round for SE3, a company whose software combines computer vision and AI to help machines turn sensor data into spatial intelligence. Accel says the software allows machines to operate in environments where satellite navigation is disrupted, is platform-agnostic, and integrates with hardware without competing with it. SE3 is working on multiple contracts with Ukraine and on contracts with Germany's armed forces, and Accel says SE3 is an AI partner to HENSOLDT, to Auterion, and to ERC System. m:…9390

Bonfire Ventures co-led a $9 million funding round for Axle, with Eniac co-leading and Teamworthy Ventures participating. Axle's software replaces legacy technology used by investment banks for corporate access, which Bonfire describes as connecting public company management teams with institutional investors through events like conferences and roadshows. Axle uses AI agents to build meeting schedules and identify investors, its platform integrates with a bank's existing CRM, and Axle plans to expand into broker voting. m:…5392

A drug-trial bet with a hard number attached

Orakl Oncology raised $34 million in a Series A funding round, bringing its total funding to $50 million, according to Norrsken VC. The company says its predictive engine has demonstrated 90% accuracy in predicting cancer clinical trial outcomes, and says 18 of the world's top 20 pharmaceutical companies are in its commercial pipeline. Norrsken notes that approximately 95% of oncology drugs entering clinical development fail to reach approval and approximately 75% of oncology candidates fail during Phase II clinical trials, which is the failure rate Orakl is trying to predict around. m:…7006

Orakl was founded in 2023 as a spin-out of Gustave Roussy, its team consists of more than 40 people, with 30% holding PhDs, and the company plans to increase its data generation fivefold. Orakl intends to expand its research beyond colorectal and pancreatic cancers into lung and gastric cancers, and plans to establish teams in the US and UK to deploy its predictions across more than 10 drug programs. It is also adding a capability to discover and validate drug co-targets directly from patient biology. m:…7006

The top of the market tells a different story

The three rounds are small against the market they sit in. Bessemer Venture Partners says the inaugural AI 100 list represents a total private value of $3.6 trillion, with an average valuation of $36 billion and an average ARR multiple of 57x. Frontier labs and foundation model builders represent 9% of the AI 100 companies but hold 55% of the list's total value, and the average AI 100 company grew 480% year over year. m:…4133

The surveyed companies expect the value to move toward applications: 87% of surveyed AI 100 companies believe value will accrue to the application layer rather than to model providers, and 79% believe frontier model quality will converge in the next 2-3 years. That is the bet behind all three rounds in this story: if models converge, the money goes to the company that turns a model into a working system for a defence customer, a bank, or a drug trial. Excluding Anthropic and OpenAI, the average AI 100 company has raised more than $3.9 billion in capital, a reminder of how far behind the small rounds still are. m:…4133