Alphabet Inc. and Nvidia Corporation backed a fast-rising artificial intelligence startup, Safe Superintelligence (SSI), which was co-founded by OpenAI’s former chief scientist, Ilya Sutskever, a source familiar with the development disclosed.

The strategic move by the two technology giants came as part of a broader trend of infrastructure providers investing in AI ventures requiring vast computational capacity.

SSI swiftly gained traction and, according to the same source, secured a valuation of $32 billion in a funding round led by venture capital firm Greenoaks. The startup attracted heightened attention in the artificial intelligence space due to Sutskever’s reputation for identifying pivotal advancements in AI technology.

The investment illustrated a renewed focus by Alphabet and Nvidia on startups developing cutting-edge AI solutions. Earlier in the week, Alphabet’s cloud computing division finalised an agreement to provide SSI with access to tensor processing units (TPUs), its proprietary AI chips. Alphabet, already operating its own AI models, continued to expand its hardware strategy through external sales.

Reuters reported that Alphabet originally reserved TPUs for internal purposes. However, the new deal marked a shift in strategy, as SSI became one of the first companies to receive TPUs at scale to support its frontier AI initiatives.

During an interview with Reuters, Google’s managing director overseeing startup partnerships, Darren Mowry, said, “With these foundational model builders, the gravity is increasing dramatically over to us.”

The startup SSI reportedly opted to use TPUs over graphics processing units (GPUs) for its AI research and development, two sources revealed. This choice contrasted with the historical preference among AI developers for Nvidia GPUs, which dominate more than 80 percent of the market.

Google Cloud currently offers both Nvidia GPUs and its in-house TPUs. The TPUs, designed to outperform general-purpose GPUs in specific AI tasks, have been deployed in building large-scale AI models for companies such as Apple Inc. and Anthropic. The latter, a rival to OpenAI, has secured substantial funding from both Google and Amazon.

Anthropic, despite gaining access to Amazon’s in-house Trainium and Inferentia chips, has continued to use TPUs for its AI development, two sources confirmed. The company has also not reduced its investment in Google’s AI chip offerings.

Alphabet and Nvidia both declined to reveal the financial specifics of their investment in SSI. Representatives from SSI also refrained from issuing comments on the matter.

Google and Nvidia now face stiff competition from Amazon, which has advanced its semiconductor initiatives to support partners like Anthropic. In December, Amazon named Anthropic as the inaugural customer for a supercomputer powered by hundreds of thousands of its proprietary chips.

Microsoft, meanwhile, has reinforced its alliance with OpenAI through significant capital injections. Nvidia extended its own backing to OpenAI and also invested in xAI, an artificial intelligence venture founded by Elon Musk.

It became increasingly standard for cloud providers to invest in AI startups that not only develop foundational models but also become heavy users of their infrastructure.