Google announces TPU 8tx and TPU 8i Max accelerators
Google has expanded its eighth-generation Tensor Processing Unit family with two new chip offerings designed to enhance hardware security and double inference performance.
FTMQ SI, written by our newsroom1 view
Google has expanded its eighth-generation Tensor Processing Unit lineup by introducing the TPU 8tx and TPU 8i Max accelerators, according to MarkTechPost, Pulse 2.0, and Data Center Knowledge. The Tensor Processing Unit is an application-specific integrated circuit developed by Google for neural network machine learning, which Google began using internally in 2015 and made available for third-party use in 2018. The new processors target emerging enterprise requirements for artificial intelligence inference and long-context applications. [1][2][3][8]
The TPU 8i Max combines two TPU 8i chips into a single package, MarkTechPost, Data Center Knowledge, and Wccftech all report. Data Center Knowledge explains that this dual-chip packaging doubles the compute capacity, SRAM, and memory bandwidth compared to the standard TPU 8i, aiming at higher-capacity inference and enterprise workloads with large context windows. The Next Platform notes that the TPU 8i is a variant of the Zebrafish accelerator that Google first revealed in April. [1][3][4][5]
The second model, the TPU 8tx, adds hardware confidential inference capabilities, according to MarkTechPost and Data Center Knowledge. Pasquale Pillitteri reports that the TPU 8tx is the first TPU with data encrypted while in use. Wccftech reports that Google pairs the TPU 8tx with custom AMD EPYC CPU hosts, while The Next Platform notes that the underlying TPU 8t accelerator, codenamed Sunfish, is manufactured for Google by Broadcom. [1][3][4][5][6]
MarkTechPost reports that Bloomberg cited a 2027 availability timeline for the chips. The Next Platform separately reports that Google will roll out the new accelerators next year for Google Cloud and potentially for select direct customers operating Google compute engines within their own data centers. [1][4]
In short
- Google announced the TPU 8tx and TPU 8i Max as part of its eighth-generation TPU lineup.
- The TPU 8i Max combines two TPU 8i chips in one package, doubling compute, SRAM, and memory bandwidth.
- The TPU 8tx adds hardware confidential inference to encrypt data while in use.
Sources
Every paragraph above points to the numbered items it rests on. Read the originals here.
- [1]Google Unveils TPU 8i Max and TPU 8tx: A Dual-Chip TPU and a Confidential TPUMarkTechPost, 8d ago (the report this story comes from)
- [2]Google Unveils TPU 8tx And TPU 8i Max With 2X Compute And Enhanced Security For AI AgentsPulse 2.0, 8d ago
- [3]Google Adds TPU 8tx, 8i Max for Enterprise AIData Center Knowledge, 8d ago
- [4]Google Rejiggers TPU Accelerators For Broader AI Inference WorkThe Next Platform, 8d ago
- [5]Google Pairs Its New TPU 8tx With Custom AMD EPYC CPU Hosts, While The TPU 8i Max Combines Two TPU 8i Chips In A Single Package To Offer Phenomenal Compute And Memory DensityWccftech, 8d ago
- [6]Google unveils TPU 8tx and 8i Max, first TPU with data encrypted in usePasquale Pillitteri, 8d ago
- [7]Google adds new TPU 8tx and TPU 8i Max for AI workloadsTestingCatalog AI News, 8d ago
Background
- [8]Tensor Processing Unit on Wikipedia
- [9]Google on Wikipedia
Our newsroom writes these reports with the help of software, from the 9 sources listed and nothing else, and checks them against those sources. Facts can still be wrong or move on; the originals are the record. Spotted a mistake? Write to daniel@monsterkong.com.
Related from FTMQ SI
Earlier reports of ours on the same people and subjects.
- Google Cloud expands custom silicon family with TPU 8tx and TPU 8i MaxChips & Infrastructure, 8d ago
- Oracle expands cloud hardware with Google TPUs and OpenAI inference chipsModels & Products, 21d ago
- OpenAI and Oracle expand partnership to integrate software and chipsIndustry Labs, 21d ago
- Google Expands TPU Inference Ecosystem to Challenge Nvidia Software DominanceIndustry Labs, 12d ago
- Google DeepMind Unveils SynthID Bio to Watermark Synthetic ProteinsResearch Papers, 43d ago
More in Models & Products
- Microsoft unveils Copilot Workspaces and ROI tools at Ignite 20264h ago
- Meta deploys LLM test system to catch software bugs in production1d ago
- Alibaba releases Qwen4-Max model with native vision MoE architecture1d ago
- NiCE launches next generation Cognigy platform for agentic contact center automation2d ago
- AI scanning tool uncovers dozens of Linux kernel bugs2d ago
- Cmpnd releases dspy-wsp for weight space programming in DSPy2d ago
Get the day in one email
Reports like this one, the top news of the last 24 hours, every morning. Free, one email a day; readers can comment under every report.
By signing up you agree to our terms and privacy policy. Unsubscribe any time.

Comments
Loading
Join the conversation
Comments are open to readers of our daily email: the top news of the last 24 hours, every morning, free. Sign up and the comment box opens.
Already on the list? Enter the same address and we will send a sign-in link.
By signing up you agree to our terms and privacy policy. Unsubscribe any time.