Alphabet, Nvidia Invest in Sutskever’s AI Startup Superintelligence

Alphabet (GOOGL.O) and Nvidia (NVDA.O) have joined leading venture capital firms in backing Safe Superintelligence (SSI), an AI startup co-founded by former OpenAI chief scientist Ilya Sutskever, according to a source familiar with the matter.

Despite being only months old, SSI has quickly emerged as one of the most valuable startups in the artificial intelligence space. The investment signals renewed interest from major tech players and infrastructure providers in supporting AI startups pushing the boundaries of innovation—efforts that demand immense computing power. Earlier this week, Alphabet’s cloud division also announced a deal to provide SSI with access to its proprietary tensor processing units (TPUs), reinforcing its commitment to the AI frontier.

SSI, reportedly valued at $32 billion in a recent funding round led by Greenoaks, has quickly become one of the most prominent startups in AI model research—largely due to co-founder Ilya Sutskever’s strong reputation for anticipating major breakthroughs in AI. Like many of its peers, the company faces a significant demand for high-performance chips. The specific details of Alphabet’s and Nvidia’s investments in SSI remain undisclosed, and spokespersons for all three companies declined to comment.
Alphabet’s coordinated efforts—through both its corporate arm and cloud division—to partner with leading AI labs like SSI and Anthropic reflect the company’s shifting strategy around AI hardware.

Google, which initially reserved its tensor processing units (TPUs) for internal use, is now broadening access to these chips. The recent deal to supply SSI with a substantial volume of TPUs to support advanced AI research highlights this transition. Darren Mowry, managing director overseeing Google’s startup partnerships, confirmed the move in an interview with reporter this week.
“With these foundational model builders, the gravity is increasing dramatically over to us,” he said.
While AI developers have traditionally favored Nvidia’s graphics processing units (GPUs), which dominate more than 80% of the AI chip market, SSI has primarily relied on Google’s tensor processing units (TPUs) for its research and development, according to two sources.

Google Cloud offers both Nvidia GPUs and its proprietary TPUs, which are designed for specific AI workloads and offer greater efficiency than general-purpose GPUs. These chips have powered the development of large-scale AI models, including those from Apple and Anthropic—the latter being a major OpenAI rival backed by significant investments from Google and Amazon.
Google and Nvidia now face growing competition from Amazon, which is developing its own AI processors, Trainium and Inferentia. As early as 2023, Amazon announced that Anthropic would begin building its models using these in-house chips. In December, the company revealed that Anthropic would be the first to leverage a massive new supercomputer powered by hundreds of thousands of Amazon-designed processors.

Despite this partnership, Anthropic continues to rely heavily on Google’s TPUs for its AI development and has not scaled back its spending on Google’s chips, according to two sources.

It has become increasingly common for major cloud providers to make substantial investments in AI startups that not only develop foundational models but also rely heavily on their cloud infrastructure. Amazon and Google, for example, have both invested in Anthropic, while Microsoft has made significant investments in OpenAI. Nvidia has also supported OpenAI, along with Elon Musk’s AI venture, xAI.

Reported by Kenrick Cai in Las Vegas, Max Cherney in San Francisco, and Krystal Hu in New York; Editing by Diane Craft.

Reuters/Jide Johnson