Google Cloud has introduced that Protected Superintelligence (SSI), a synthetic intelligence startup based by Ilya Sutskever, is leveraging its tensor processing items (TPUs) for superior AI analysis.
This collaboration, revealed on April 9 at Google Cloud Subsequent 2025, positions Google Cloud as SSI’s main computing supplier, enabling the startup to speed up its efforts in growing protected superintelligent AI techniques.
SSI, which emerged from stealth in June 2024 following Sutskever’s departure from OpenAI, secured $1 billion from outstanding traders, together with Andreessen Horowitz and Sequoia Capital. The funding will assist the computational calls for of coaching large-scale AI fashions and recruiting prime researchers.
Sutskever, who co-founded OpenAI and beforehand labored at Google Mind, has emphasised SSI’s mission to focus completely on growing protected superintelligent AI. The corporate’s web site underscores this as its sole goal. Except for the partnership with Google Cloud, it stays unclear whether or not SSI has agreements with different computing suppliers as properly.
In addition to SSI, AI music startup Udio is utilizing TPUs to assist prepare its fashions for music technology and serve its quickly rising buyer base.
In the meantime, Google unveiled Ironwood, its seventh-generation TPU, designed particularly for inference. It’s a key a part of Google’s broader AI Hypercomputer structure.
“Ironwood is our strongest, succesful and energy-efficient TPU but. And it’s purpose-built to energy considering, inferential AI fashions at scale,” Google stated. The tech big stated that at this time, we dwell within the “age of inference”, the place AI brokers actively search, interpret, and generate insights as a substitute of simply responding with uncooked knowledge.
The corporate additional stated that Ironwood is constructed to handle the complicated computation and communication calls for of considering fashions, equivalent to giant language fashions and mixture-of-experts techniques. It added that with Ironwood, prospects not have to decide on between compute scale and efficiency.
Ironwood can be accessible to Google Cloud prospects later this yr, the tech big stated. It at present helps superior fashions, together with Gemini 2.5 Professional and AlphaFold.
The TPU helps two configurations, one with 256 chips and one other with 9,216 chips. The total-scale model delivers 42.5 exaflops of compute, over 24 instances the efficiency of the El Capitan supercomputer, which provides 1.7 exaflops per pod. Every Ironwood chip gives 4,614 TFLOPs of peak compute.
The submit Ilya Sutskever’s Startup Makes use of Google’s TPUs to Construct Protected, Superintelligent AI appeared first on Analytics India Journal.
