Imiron secures US$1.08M pre-seed for AI safety verification
What's the deal? Tokyo-based ImironDealroom has a profile for this one. Try Dealroom → has raised US$1.08M (roughly $950,000) in a pre-Series A round to build safety verification tools for AI systems used in autonomous vehicles and medical devices. DG Daiwa VenturesDealroom has a profile for this one. Try Dealroom → led the round, with Mitsubishi UFJ CapitalDealroom has a profile for this one. Try Dealroom → and Gokin Capital also participating.
The startup, based in Chiyoda, is developing SpecForge — a platform that converts vague system requirements into mathematically precise specifications. It uses a proprietary language called Lilo alongside large language models to make formal verification accessible to engineers who aren't specialists in mathematical logic.
Why now? As AI moves from screens into physical systems — cars, robots, surgical tools — the stakes of failure rise dramatically. Unlike a chatbot hallucination, an autonomous vehicle hallucination can be fatal.
Traditional specification methods using Word or Excel documents have long introduced ambiguities that cause bugs during implementation. With AI systems behaving statistically rather than deterministically, clear and verifiable specifications have become essential for both safety and regulatory compliance.
What could go wrong? Imiron is tackling a genuinely hard problem, but it faces hurdles. Formal verification remains a niche discipline, and convincing automotive and medical device manufacturers to adopt a new specification workflow is no small task. The company also needs to recruit top-tier engineers globally — a competitive endeavour on a modest funding base.
The founding team is strong on paper. Chief executive Masakazu Adachi previously led autonomous driving R&D at Toyota Central R&D Labs and DensoDealroom has a profile for this one. Try Dealroom →. Chief science officer Katsuaki Hasuo is a professor at Japan's National Institute of Informatics specialising in mathematical reliability. Chief technology officer James Haydon brings formal methods expertise from Imperial College London and Oxford. But translating academic rigour into a scalable commercial product is a different challenge.
The signal: Imiron's round is modest in size but notable for the calibre of its backers — DG Daiwa Ventures and Mitsubishi UFJ Capital are established institutional funds betting that formal verification will become a non-negotiable layer in autonomous and medical AI development. With the company still at the early growth stage, the investment reflects a longer-term thesis: as regulators tighten safety requirements for physical AI systems worldwide, startups that can bridge the gap between academic formal methods and industrial-scale deployment stand to capture outsized value.
Read more: third-news.com
