Tensor Processing Unit
PulseAugur coverage of Tensor Processing Unit — every cluster mentioning Tensor Processing Unit across labs, papers, and developer communities, ranked by signal.
- 2026-06-09 product_launch Google has reportedly ordered over three million Intel-manufactured TPUs for delivery in 2028. source
1 day(s) with sentiment data
-
Foldable phone inner screens require professional care, not aftermarket protectors
Applying screen protectors to foldable phones requires careful consideration, particularly for the inner display. While the outer screen can be protected like a standard phone, the inner screen is made of delicate flexi…
-
InferenceX externalizes TPU stack, challenging CUDA dominance
SemiAnalysis reports that InferenceX is making significant strides in TPU inference externalization, potentially offering up to 50% better performance per dollar. The company is actively developing its TPU stack, with n…
-
Nvidia Invests $3.5B in MediaTek, Integrating Rival AI Chips into its Ecosystem
Nvidia has invested $3.5 billion in convertible bonds issued by MediaTek, a Taiwanese semiconductor company. This strategic move allows MediaTek to adopt Nvidia's NVLink Fusion platform for its custom AI accelerators, a…
-
Baidu achieves dual primary listing, signaling AI-first valuation shift
Baidu has officially transitioned to a dual primary listing on both the Hong Kong and Nasdaq stock exchanges, marking it as the first company with a comprehensive "full-stack AI" strategy to achieve this status. This st…
-
Z.ai releases GLM-5.3 Flash using 100,000 Chinese chips
Chinese AI vendor Z.ai has released its new multimodal model, GLM-5.3 Flash, which is designed for long-context processing and agentic tasks. Notably, Z.ai utilized 100,000 domestically produced chips for the model's in…
-
Gemma 4 E2B deployment issues highlight TPU serving stack limitations
A developer encountered issues deploying Google's Gemma 4 E2B model with quantized checkpoints on a vLLM serving stack using Google Cloud TPUs. Both the int4 and dequantized QAT variants failed to load due to discrepanc…
-
Inco AI releases DFlash 2 for faster LLM inference
Inco AI has released DFlash 2, an advancement in speculative decoding for large language models. This new version improves output by over 20% per verification pass with minimal latency increase, building on the original…
-
Google Legend Jeff Dean Departs After 27 Years to Launch AI Startup
Jeff Dean, a legendary Google engineer and one of its earliest employees, has announced his departure after 27 years. Dean was instrumental in building much of Google's foundational infrastructure, including systems lik…
-
Google posts first cash-flow negative quarter due to AI data center spending
Google's parent company, Alphabet, reported its first-ever cash flow negative quarter, with a $5.9 billion deficit in Q2 2026. This was driven by a record $44.9 billion in capital expenditures, primarily for AI data cen…
-
JAXBench launches to optimize AI kernels on Google TPUs
A new benchmark suite called JAXBench has been developed to specifically address the optimization of AI kernel performance on Google Cloud TPUs. This suite includes 50 JAX workloads derived from prominent AI models like…
-
Hainan plans eVTOL and drone logistics, boosts commercial space sector
Hainan Province is planning to integrate advanced logistics technologies, including electric vertical takeoff and landing (eVTOL) aircraft and large-payload drones, for cross-sea transportation of goods. The "15th Five-…
-
Google develops 'Rigel' chip to exclusively power Gemini AI models
Google is reportedly developing a proprietary chip designed to significantly outperform its existing Tensor Processing Units (TPUs) in terms of energy efficiency. This new chip, codenamed 'Rigel', is intended to exclusi…
-
Google develops custom chip to boost Gemini AI performance tenfold
Google is reportedly developing a new, proprietary chip designed to significantly enhance the performance and efficiency of its Gemini AI models. This custom silicon aims to outperform existing Tensor Processing Units (…
-
Google develops secret 'Project 7' chip to power Gemini AI models
Google is reportedly developing a proprietary chip designed to significantly outperform its own Tensor Processing Units (TPUs) in terms of energy efficiency. This new chip, codenamed 'Project 7' or '7B', is intended to …
-
Foreign institutions bullish on Chinese stocks; digital yuan sees application surge
Several foreign financial institutions, including Citi and Standard Chartered, have upgraded their ratings and increased their positions in Chinese stocks, expressing optimism for the second half of the year. This posit…
-
Google develops 'Rigel' chip to boost Gemini AI efficiency
Google is reportedly developing a proprietary AI chip, codenamed "Rigel," designed to significantly outperform its existing Tensor Processing Units (TPUs) in terms of energy efficiency for its Gemini models. This move s…
-
Google develops new AI chip; Beijing plans Token factories; Tencent launches AI agent · 1 source tracked
Google is developing a new AI server chip codenamed "Frozen v2" to enhance the performance of its Gemini models, aiming for a significant increase in token processing power per unit of energy compared to its current TPU…
-
China's private fund industry hits record 23.66T yuan in assets, sees strong issuance
China's private fund industry reached a new record in June 2026, with total assets under management hitting 23.66 trillion yuan, marking the ninth consecutive month of growth. Securities-focused private funds exceeded 8…
-
Hangdian Co. plans $400M AI copper foil and fiber project; China Telecom launches TPU cluster
Hangdian Co., Ltd. plans to raise up to 2.88 billion yuan through a private placement of A-shares. The funds will be allocated to a super factory project for the research, development, and manufacturing of optical fiber…
-
Google reportedly taps Intel's EMIB-T for next-gen TPUs, challenging TSMC's packaging dominance
Google is reportedly shifting its next-generation Tensor Processing Unit (TPU) production, codenamed Humufish, from TSMC's CoWoS-L advanced packaging to Intel's EMIB-T technology. This move, if accurate, signifies a sig…