Google grows TPU stash by ~$7bn in last year
As company seeks to increasingly sell TPUs, not just lease them
News - The Compute, Storage & Networking Channel, Google Cloud
The latest news from the AI, data center, telco, chip, and cloud sector
-
-
Google developing new chip to run Gemini AI model more efficiently – report
The hardware, dubbed Frozen v2, is slated for launch in 2028
-
FBI considers deploying AI LLM supercomputers with Nvidia B300 GPUs or Google TPUs
Could deploy multiple systems, from training to inference
-
Google to sell TPUs to a "select group of customers" for their data centers, increases capex forecast
Company sees strong Q1 for FY2026
-
Google unveils eighth-generation TPUs, two dedicated training and inference chips
“Purpose-build architectures” designed to support model training and inference workloads
-
CoreWeave partners with Google Cloud on cross-cloud training and inference
Launches "CoreWeave Interconnect"
-
Broadcom touts growing sales with Anthropic, AI orders
And a new unnamed XPU customer is teased
-
AWS and Google Cloud partner on multicloud networking
Two cloud giants come together to create a "more open cloud environment"
-
Google must double AI capacity every six months to meet demand - report
All-hands meeting coinciding with GA of Ironwood release reveals extent of AI demand
-
Google Cloud instances with Nvidia GB300s now "shipping in production"
A4X Max instance contains 72 Blackwell Ultra GPUs and 36 Nvidia Grace CPUs
-
Google and Anthropic confirm massive 1GW+ cloud deal with up to 1 million Google TPUs
Contract worth "tens of billions of dollars"
-
Pasqal's neutral-atom quantum computer available via Google Cloud
Follows similar launch on Microsoft Azure earlier this year
-
Google Cloud launches instances equipped with 5th Gen AMD Epyc processors
Available via C4D and H4D virtual machines
-
Oracle and Google expand multicloud offering
Add new regions, bring availability to company partners, and plan new service
-
Nebius and YTL to offer Blackwell Ultra instances, others announce availability of B200s and GB200s
And more cloud news from GTC
-
Google launches A4X VMs powered by Nvidia GB200 NVL72
Currently available in preview
-
Google makes custom Axion-based virtual machines generally available
Early adopters include Paramount Streaming and Spotify
-
Google offers sneak peak at its Nvidia Blackwell GB200 NVL racks
Shortly follows Microsoft's similar preview
-
Google launches confidential cloud instances based on Intel Xeon and AMD Epyc processors
Confidential Computing instances now generally available
-
Google Distributed Cloud integrates NetApp data storage and intelligence services
Integration aims to provide AI infrastructure that meets high security standards
-
LG AI Research uses Google Cloud infrastructure to train Exaone 3.0 genAI models
Reducing operating costs of training by 72 percent
-
Gen AI startup Magic partners with Google Cloud; closes $320m funding round
Claims cloud-based supercomputers will have 160 exaflops of performance
-
Google Cloud gives Y Combinator startups dedicated Nvidia GPU cluster
And access to TPUs
-
Apple to use Google's TPUs for AI instead of Nvidia GPUs
Apple Intelligence rolls out beta test this week
-
Alice & Bob quantum processor available on Google Cloud Marketplace
Company pitching cat qubit-based system
-
Tape Ark migrates tape-based sports broadcast videos to Google Cloud
Cloud ingest totals more than 40 petabytes of video
-
Google unveils Axion, its first Arm-based CPU
Also reveals updates to TPU AI chip
-
Stability AI defaulted on payments to AWS, owed Google Cloud and CoreWeave money
Troubled generative AI business owes millions to cloud companies