Neyshabur and Mehta departed Anthropic in December 2025, less than a year after joining the lab. The company has not yet shipped a product. Specific details about the technical approach and timeline remain undisclosed.
The funding reflects a broader bet that AI systems capable of automating their own research cycles could disrupt the current dominance of well-resourced frontier labs. Mirendil's stated goal is to compress research cycles from months of human work into days of automated iteration, potentially enabling universities and research institutions to build and refine AI models without dedicated machine-learning engineering teams. For attorneys tracking AI governance and competitive dynamics, this signals accelerating consolidation of AI research talent into new ventures and growing investor confidence that self-improving systems will materially advance scientific discovery in biology and materials science within the next two to three years.