Alphabet and Nvidia have become the latest major tech players to invest in Safe Superintelligence (SSI), a cutting-edge AI startup co-founded by Ilya Sutskever, the former chief scientist at OpenAI. Just months after launching, SSI has rapidly gained attention and valuation, reportedly hitting $32 billion in a funding round led by Greenoaks, according to a source familiar with the deal.
This move signals a renewed wave of strategic investment from tech giants and cloud infrastructure providers into startups driving innovation in artificial intelligence — particularly those requiring immense computational power.
Alphabet, which develops its own AI models, also announced earlier this week that its cloud division will provide SSI with access to its proprietary tensor processing units (TPUs) — AI chips designed in-house to efficiently handle complex AI tasks. This marks a shift in Google’s approach, as TPUs were previously reserved primarily for internal use.
The decision to supply SSI with significant quantities of TPUs reflects Google Cloud’s evolving hardware strategy aimed at supporting leading-edge AI research through external partnerships. “We’re seeing increasing momentum among top model builders coming to us,” said Darren Mowry, Managing Director of Google’s startup partnerships, in a recent Reuters interview.
While Nvidia continues to dominate the AI chip market — holding more than 80% share — SSI is reportedly leaning heavily on Google’s TPUs instead of Nvidia’s GPUs for its model training. That said, Google Cloud still offers both options to customers, providing flexibility to AI developers based on their computing needs.
The exact terms of Alphabet’s and Nvidia’s investments were not disclosed, and representatives from all three companies declined to comment.
The growing demand for efficient and scalable AI computing has also intensified competition among cloud service providers. Amazon, for instance, has invested heavily in Anthropic — another AI startup — and is promoting its proprietary chips, Trainium and Inferentia, as alternatives to Nvidia’s GPUs and Google’s TPUs. In late 2023, Amazon announced that Anthropic would be the first to utilize its massive supercomputer powered by these chips.
Despite Amazon’s advances, Anthropic reportedly continues to use Google’s TPUs heavily and has not scaled back its spending on Google Cloud services.
Strategic investments by major cloud platforms into AI startups like SSI and Anthropic — which are not only developing foundational models but also becoming major infrastructure customers — reflect the increasingly symbiotic relationship between cloud services and next-gen AI development.
, a cutting-edge){kind=link}