Google Adds Specialized TPU Systems to Challenge Nvidia in AI Chips
Alphabet’s Google announced two new tensor processing unit systems, TPU 8tx and TPU 8i Max, to expand its AI-chip offering. The 8tx targets confidential computing for governments and regulated companies, while the 8i Max is designed for AI workloads requiring long context windows. Google said the systems will be available next year, as it seeks to expand a TPU business that Bloomberg Intelligence forecasts will reach $135 billion by 2028.
Why it matters
The launch gives Google more specialized products to sell alongside its cloud services and increases competitive pressure on Nvidia in AI computing. Bloomberg Intelligence forecasts TPU sales will rise from an estimated $12 billion this year to $135 billion by 2028, while Google’s Amin Vahdat said specialization can deliver more value than a single general-purpose chip.
Key facts
- 1Google introduced two TPU variations, called TPU 8tx and TPU 8i Max. finance.yahoo.com
- 2TPU 8tx is intended for confidential computing and works with chips from Advanced Micro Devices, while TPU 8i Max targets applications requiring long-context processing. finance.yahoo.com
- 3TPU 8i Max combines two TPU chips in one package to provide more computing power and memory. finance.yahoo.com
- 4Google said the two variations will be available next year. finance.yahoo.com
- 5Bloomberg Intelligence forecasts TPU sales, including related software and hardware, will exceed $135 billion by 2028, compared with an estimated $12 billion this year. finance.yahoo.com
- 6Google’s TPUs cost about 30% less than Nvidia’s GPUs while providing competitive throughput, according to Bloomberg Intelligence analyst Mandeep Singh. finance.yahoo.com
Open questions
- Google did not provide pricing for the new systems in the cited materials.
Summary written by AlphAI AI Desk from 20 of 24 sources. Not investment advice. Figures are as stated by the linked sources.