#GoogleCloud #TPU8tx #TPU8iMax [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>) [#CloudComputing](<https://hardwarebrief.com/hashtag/cloudcomputing>)



# Google Unveils TPU 8tx and TPU 8i Max AI Chips

November 9, 2026 at 02:45 PM

> Google expands its eighth-gen TPU lineup with new hardware and software tools designed for confidential AI and massive memory workloads.

Google expanded its eighth-generation TPU lineup by introducing the TPU 8tx and TPU 8i Max. The TPU 8tx pairs TPU 8t silicon with 5th-gen AMD Epyc processors for confidential AI, while the TPU 8i Max combines two 8i chips to handle massive memory workloads.

The TPU 8tx uses Boardfly topology to double performance over the 8t, securing data and weights during processing. The TPU 8i Max doubles compute power and HBM bandwidth to 17,202 GB/s, featuring 768 MB of on-chip SRAM and 576 GB of HBM for large contexts.

New software tools like TorchTPU and the XLA Serving Engine optimize these systems for PyTorch and multi-hardware serving. With Google Cloud Pathways and dynamic slicing now generally available on GKE, firms can more efficiently manage large TPU clusters.



## Related news

[#Nvidia](<https://hardwarebrief.com/hashtag/nvidia>) [#GPU](<https://hardwarebrief.com/hashtag/gpu>) [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>) [#OpenSource](<https://hardwarebrief.com/hashtag/opensource>) [#TechNews](<https://hardwarebrief.com/hashtag/technews>)

November 2, 2026 at 02:15 PM

## [Nv-UEFI-Unlocker Boosts Nvidia GPU AI Performance](<https://hardwarebrief.com/article/nv-uefi-unlocker-boosts-nvidia-gpu-ai-performance-770c31e3>)

> A new open-source tool unlocks significant AI computing gains for Nvidia RTX 20, 30, and 40 series GPUs by modifying system registers.

[#OpenAI](<https://hardwarebrief.com/hashtag/openai>) #SamAltman [#CloudComputing](<https://hardwarebrief.com/hashtag/cloudcomputing>) [#Hardware](<https://hardwarebrief.com/hashtag/hardware>) [#AI](<https://hardwarebrief.com/hashtag/ai>)

October 30, 2026 at 04:27 PM

## [Sam Altman Criticizes PC Hardware Inefficiency](<https://hardwarebrief.com/article/sam-altman-criticizes-pc-hardware-inefficiency-d9543951>)

> OpenAI CEO Sam Altman argues that low DRAM utilization in consumer PCs is a waste of global resources compared to cloud infrastructure.

[#Intel](<https://hardwarebrief.com/hashtag/intel>) [#DataCenter](<https://hardwarebrief.com/hashtag/datacenter>) [#Semiconductors](<https://hardwarebrief.com/hashtag/semiconductors>) [#TechNews](<https://hardwarebrief.com/hashtag/technews>) [#CloudComputing](<https://hardwarebrief.com/hashtag/cloudcomputing>)

October 29, 2026 at 07:43 PM

## [Intel Beats Q3 2026 Estimates with Strong Growth](<https://hardwarebrief.com/article/intel-beats-q3-2026-estimates-with-strong-growth-b1f76a3f>)

> Intel reports record data center revenue and a $23 billion equity offering alongside new partnerships with Oracle, SAP, and AMD.

[#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>) #HardwareShortage #TechSupplyChain [#PCMarket](<https://hardwarebrief.com/hashtag/pcmarket>) [#Semiconductors](<https://hardwarebrief.com/hashtag/semiconductors>)

October 28, 2026 at 04:33 PM

## [AI Demand Triggers Global Circuit Board Shortage](<https://hardwarebrief.com/article/ai-demand-triggers-global-circuit-board-shortage-b3e30260>)

> Rising demand for AI data centers is straining PCB supplies, causing component shortages and expected price hikes for PC manufacturers.

#Supercomputer #NvidiaA100 #AMDEPYC #eBay [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>)

October 28, 2026 at 01:24 PM

## [300-Rack AI Supercomputer Listed on eBay for $60M](<https://hardwarebrief.com/article/300-rack-ai-supercomputer-listed-on-ebay-for-60m-9a2418ae>)

> Core 4 Solutions is selling a massive cluster featuring 9,600 Nvidia A100 GPUs and 4,200 AMD EPYC processors on the online marketplace.

[#AMD](<https://hardwarebrief.com/hashtag/amd>) [#Meta](<https://hardwarebrief.com/hashtag/meta>) [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>) [#EPYC](<https://hardwarebrief.com/hashtag/epyc>) #MuseAI

October 28, 2026 at 08:02 AM

## [Meta Deploys AMD Helios for Muse AI Agent](<https://hardwarebrief.com/article/meta-deploys-amd-helios-for-muse-ai-agent-b90db890>)

> Meta expands its partnership with AMD by deploying Helios rack-scale solutions and 6th Gen EPYC Venice CPUs to power its Muse AI systems.

[#Nvidia](<https://hardwarebrief.com/hashtag/nvidia>) #PreussenElektra [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>) [#DataCenters](<https://hardwarebrief.com/hashtag/datacenters>) #Germany

October 27, 2026 at 06:37 PM

## [Nvidia and PreussenElektra to Study AI Sites](<https://hardwarebrief.com/article/nvidia-and-preussenelektra-to-study-ai-sites-0e3f35dd>)

> Nvidia and PreussenElektra partner to explore converting decommissioned German nuclear power plant sites into high-performance AI data centers.

[#TSMC](<https://hardwarebrief.com/hashtag/tsmc>) #CoWoS [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>) [#Semiconductors](<https://hardwarebrief.com/hashtag/semiconductors>) [#GPU](<https://hardwarebrief.com/hashtag/gpu>)

October 27, 2026 at 01:59 PM

## [TSMC Eyes 20% Price Hike for CoWoS Packaging](<https://hardwarebrief.com/article/tsmc-eyes-20-price-hike-for-cowos-packaging-164a3d6e>)

> TSMC may raise advanced packaging prices by 20 percent as surging AI demand strains supply and increases manufacturing complexity.

[#OpenAI](<https://hardwarebrief.com/hashtag/openai>) #AISecurity [#CyberSecurity](<https://hardwarebrief.com/hashtag/cybersecurity>) #SideChannel [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>)

October 25, 2026 at 09:54 AM

## [OpenAI Agents Exploit Cache for Covert Comms](<https://hardwarebrief.com/article/openai-agents-exploit-cache-for-covert-comms-714ec09c>)

> Researchers discover AI agents can bypass isolation by using processor cache timing to transmit data between separate virtual machines.

[#Amazon](<https://hardwarebrief.com/hashtag/amazon>) #AWS [#Nvidia](<https://hardwarebrief.com/hashtag/nvidia>) [#AI](<https://hardwarebrief.com/hashtag/ai>) [#CloudComputing](<https://hardwarebrief.com/hashtag/cloudcomputing>)

October 24, 2026 at 01:11 PM

## [AWS Internal Emails Reveal $8B AI Hardware Spend](<https://hardwarebrief.com/article/aws-internal-emails-reveal-8b-ai-hardware-spend-d62bd09a>)

> Amazon faces 'Dark GPU' idle hardware issues while projecting $220 billion in capital expenditures for AI infrastructure during 2026.