Alphabet and Nvidia have joined a group of high-profile investors to back Safe Superintelligence (SSI), the AI startup that was co-founded by former OpenAI chief scientist Ilya Sutskever. According to a Reuters report, the startup has rapidly ascended to one of the highest valuations in the AI sector, securing a valuation of approximately $32 billion in a recent funding round (of an additional $2 billion) led by venture capital firm Greenoaks. At that time, Greenoaks put in $500 million in the AI startup, while the infusion of capital also included Andreessen Horowitz, DST Global, and Lightspeed Venture Partners.
Recent months have seen more and more tech companies invest heavily in AI and AI firms to grab a slice of the AI pie, and these startups, like SSI, often demand considerable computing power to develop and train large-scale AI models. Alphabet’s participation in the funding coincides with an agreement through its Google Cloud arm to provide SSI access to its proprietary tensor processing units (TPUs). For now, the specific terms of Alphabet’s and Nvidia’s financial contributions to SSI remain undisclosed.
Speaking of SSI, the firm launched last June, has around 20 employees (as of March 2025) and is based in Palo Alto and Tel Aviv. It was founded by Sutskever (alongside Daniel Gross and Daniel Levy) after his departure from OpenAI in 2023 following internal turmoil, including the controversial ouster and later reinstatement of CEO Sam Altman. At that time, the firm declared its mission to develop a superintelligence, and is currently focused on building advanced AI models that are safe and outperform human intelligence, which is a nod to its co-founder Sutskever’s earlier work at OpenAI.
Since then, the startup has quickly gained industry attention due to its need for AI-specialized hardware. Unlike many of its peers, SSI is reportedly relying heavily on Google’s TPUs rather than Nvidia’s industry-dominant GPUs. According to reports, sources familiar with the company’s technical stack told Reuters that SSI’s AI research and development is primarily powered by TPUs, which are designed to handle specific AI workloads more efficiently than general-purpose graphics processors. While Nvidia still dominates with over 80% of the AI chip market, Google is carving out space with its TPUs, which are efficient in training large-scale AI models.
This marks a shift for Google as well – the tech titan’s cloud unit has historically cautious about offering its TPUs to external customers. Darren Mowry, Google’s head of startup partnerships, said this reflects growing momentum, and that “gTPWith these foundational model builders, the gravity is increasing dramatically over to us.”
While Alphabet and Nvidia are boosting their stake in startups like SSI and Anthropic, firms like Amazon are not falling behind in the AI race either. Amazon Web Services has developed its own AI chipsets—Trainium and Inferentia—and has publicly announced partnerships with AI companies, including Anthropic. As of late 2023, Anthropic was the first customer for AWS’s supercomputer infrastructure powered by hundreds of thousands of Amazon-designed chips.
Content originally published on The Tech Media – Global technology news, latest gadget news and breaking tech news.
