Mirendil signs $100M+ Google Cloud deal for self-improving AI

AI lab Mirendil has signed a multiyear partnership with Google Cloud to source compute capacity for its self-improving AI research, as TechCrunch exclusively learned. The deal is worth upward of $100 million, according to Mirendil's co-founder and CEO Behnam Neyshabur, roughly half of what the startup raised in seed funding at a $1 billion valuation in late June. The agreement gives Mirendil access to both Google's TPUs and Nvidia GPUs, as well as managed training clusters, which the startup will use to work on its self-improving AI, hoping it will eventually take on the work of an entire frontier AI lab. Self-improving AI, also known as recursive self-improvement, refers to AI systems that iteratively improve themselves, a concept major labs like Anthropic (where Mirendil's co-founders hail from) have been working on, with startups like Recursive Superintelligence and Ricursive Intelligence also emerging. Mirendil believes this process will automate a lot of scientific and AI research, helping progress in medicine, biology, and materials science. Neyshabur noted that training self-improving AI requires enormous computing power, and co-founder Harsh Mehta emphasized the importance of matching workloads to the right hardware. Google's SVP and chief technologist Amin Vahdat said AI advancement is about orchestrating entire systems of intelligence and breaking physical constraints of scaling. In return, Google gets a strategic partner building frontier recursive self-improving AI, which it can eventually offer to enterprise customers.