Pitch deck: Hiverge raises $5 million to optimise algorithms with AI
Summary
Hiverge, a Cambridge-based startup founded by former Google DeepMind scientists, has raised $5 million in seed funding to commercialise an AI-driven platform that automates algorithm design. The team — including ex-DeepMind engineers who worked on AlphaEvolve and a University of Cambridge professor — has built “The Hive,” described as an “algorithm factory” that uses program synthesis to generate verifiable backend algorithms to meet specified functions.
The platform is aimed at improving backend code where performance and correctness matter: examples include shortening AI training times and speeding up supply-chain runtimes. Hiverge has been in stealth since late 2024, is running proofs of concept, and is likely to licence the platform to customers. The seed round was led by Flying Fish Ventures and included investors such as Ahren Innovation Capital and Google chief scientist Jeff Dean. The company will use the funds to accelerate go-to-market, product development and research.
Key Points
- Hiverge raised $5 million in a seed round led by Flying Fish Ventures, with participation from Jeff Dean and others.
- The startup was founded by former Google DeepMind engineers and a Cambridge professor; they previously worked on AlphaEvolve.
- Its product, “The Hive,” is an “algorithm factory” that uses program synthesis to automatically generate algorithms for clearly specified tasks.
- Target use cases focus on backend performance improvements — e.g. reducing AI training time or improving supply-chain runtime — rather than consumer-facing front-end code.
- Hiverge emerged from stealth in 2025, is experimenting with its business model via proofs of concept, and expects to licence its platform to customers.
- Funding will be used to speed up go-to-market efforts, support product development and expand research capabilities.
Author style
Punchy and direct: this is a tight, technically credible bet from ex-DeepMind founders that could matter to enterprises that rely on bespoke algorithms. Worth reading if you track AI tooling, developer productivity or enterprise ML ops.
Why should I read this?
Quick take: if you care about how AI and clever code actually get faster or cheaper in the real world, this is a neat one. Hiverge isn’t selling flashy consumer apps — it’s building a tool that promises verifiable, performance-focused algorithm improvements. Read it if you want a snapshot of where elite ML talent is putting early-stage bets and how that could shave time and cost off production systems.
