←Search results

English text is being prepared.

By relevance · Total 215 results Clear search

Sunday

  1. Story1 reports · 1 facts

    Saif Khan founds the Center for Technology and Statecraft

    On October 17, 2026, Saif Khan, a former official at the U.S. National Security Council and the Department of Commerce, founded a new think tank, the Center for Technology and Statecraft (CTS). An interview published by ChinaTalk the same day showed that in CTS's DUV report, Khan predicted that if exports of the relevant equipment were stopped now, by the mid-2030s China could produce about 100 times fewer AI chips than the United States, and he argued that DUV controls are the most critical lever for the United States to influence the AI balance of power. In the interview he also discussed how robotics will become the next frontier in U.S.-China AI competition, and how AI is changing policy research itself. At present, the center and the above judgments all come from the same report, with no other independent sources or subsequent developments.

    On October 17, 2026, Saif Khan, a former official at the U.S. National Security Council and the Department of Commerce, founded a new think tank, the Center for Technology and Statecraft (CTS). An interview published by ChinaTalk the same day showed that in CTS's DUV report, Khan predicted that if exports of the relevant equipment were stopped now, by the mid-2030s China could produce about 100 times fewer AI chips than the United States, and he argued that DUV controls are the most critical lev

    Matched fieldsSummary

Saturday

Thursday

Wednesday

Wednesday

Sunday

Tuesday

Monday

Thursday

  1. Story1 reports · 1 facts

    JEDEC approves SPHBM4 standard (JESD330-4)

    On June 26, 2026, the international semiconductor standards organization JEDEC officially approved the SPHBM4 standard, numbered JESD330-4, with the full name Standard Package High Bandwidth Memory 4. The goal of this standard is to reduce AI chip memory costs without sacrificing performance: cutting HBM4's 2048 data signal pins to 512, while raising the per-pin transfer rate to about 44 Gbps, using 4:1 serialization to maintain total bandwidth comparable to standard HBM4, and allowing direct mounting on standard organic substrates, bypassing the silicon interposer. The standard is currently in the post-publication stage awaiting adoption; no memory manufacturer has yet launched a product, and no cloud provider has announced plans to incorporate it into its own chip designs.

    On June 26, 2026, the international semiconductor standards organization JEDEC officially approved the SPHBM4 standard, numbered JESD330-4, with the full name Standard Package High Bandwidth Memory 4. The goal of this standard is to reduce AI chip memory costs without sacrificing performance: cutting HBM4's 2048 data signal pins to 512, while raising the per-pin transfer rate to about 44 Gbps, using 4:1 serialization to maintain total bandwidth comparable to standard HBM4, and allowing direct mo

    Matched fieldsSummary

Saturday

Thursday

  1. Center for Strategic and Security Studies at Tsinghua University

    Sun Chenghao: Southeast Asia's Choices in the China-US AI Rivalry and ASEAN Regional Cooperation

    Sun Chenghao, associate research fellow at the Center for Strategic and Security Studies at Tsinghua University, believes that Southeast Asia's choices in the AI field will be shaped by the China-US rivalry over the long term, but that countries can strengthen their capacity for independent choice through their own investments and regional cooperation. The SEA-LION project of AI Singapore used Llama and Gemma from US companies and Alibaba's Qwen from China to develop Southeast Asian language models. ASEAN has issued the 2025-2030 Responsible AI Roadmap and is advancing the building of a regional AI safety network.

    English text is being prepared.

    Sun Chenghao, associate research fellow at the Center for Strategic and Security Studies at Tsinghua University, believes that Southeast Asia's choices in the AI field will be shaped by the China-US rivalry over the long term, but that countries can strengthen their capacity for independent choice through their own investments and regional cooperation. The SEA-LION project of AI Singapore used Llama and Gemma from US companies and Alibaba's Qwen from China to develop Southeast Asian language mod

    Matched fieldsBody

    #academic#English text is being prepared.#English text is being prepared.

Wednesday

  1. Rohan Paul

    AMD Releases Q3 2026 Financial Results, Expects Revenue Growth of Over 80% in 2027

    AMD announced its Q3 2026 financial results and expects revenue growth of more than 80% in 2027, above analysts' expectations of 77%. Data center chips are the main source of this forecast; CEO Lisa Su said they will account for 80% of next year's sales. That business grew 97% in the third quarter to a record $8.55 billion, driving total quarterly revenue up 45% to $13.38 billion. AMD expects fourth-quarter revenue of about $17 billion (plus or minus 3%), and Meta is testing the first batch of Helios systems under a five-year procurement agreement signed in February worth up to $60 billion.

    AMD announced its Q3 2026 financial results and expects revenue growth of more than 80% in 2027, above analysts' expectations of 77%. Data center chips are the main source of this forecast; CEO Lisa Su said they will account for 80% of next year's sales. That business grew 97% in the third quarter to a record $8.55 billion, driving total quarterly revenue up 45% to $13.38 billion. AMD expects fourth-quarter revenue of about $17 billion (plus or minus 3%), and Meta is testing the first batch of H

    Matched fieldsBody

    #ai-news#Industry news#English text is being prepared.

Tuesday

  1. Yicai

    Is AI Financing About to Hit a Minsky Moment?

    Li Wei, a professor at Cheung Kong Graduate School of Business, analyzes AI infrastructure financing risks, noting that Anthropic finalized approximately $42 billion in syndicated loans and about $18 billion in subordinated debt through the "Project Polaris," with Broadcom providing a guarantee and Blackstone contributing half of the funding. The article also uses Meta's Hyperion data center as an example, explaining that it moved $27 billion in debt off its balance sheet through an SPV, with a debt-to-asset ratio of about 90%, higher than Meta's own ratio of about 43%. The author believes there is insufficient evidence to assert that a Minsky moment is imminent, but signs such as rising issuance of junk-rated data center bonds and falling prices of some bonds are worth watching.

    Li Wei, a professor at Cheung Kong Graduate School of Business, analyzes AI infrastructure financing risks, noting that Anthropic finalized approximately $42 billion in syndicated loans and about $18 billion in subordinated debt through the "Project Polaris," with Broadcom providing a guarantee and Blackstone contributing half of the funding. The article also uses Meta's Hyperion data center as an example, explaining that it moved $27 billion in debt off its balance sheet through an SPV, with a

    Matched fieldsBody

    #chinese#Industry news#English text is being prepared.#English text is being prepared.
  2. Semiconductor Industry Observer

    The Jevons Paradox Meets Chip Iteration: Revisiting AI Compute Investment

    Gartner estimates that by 2030, AI data centers will need approximately $8.5 trillion in revenue to cover their investments, but actual AI-related revenue inflows will reach at most only $3.6 trillion, leaving more than $3.5 trillion in investment exposed to asset impairment risk. Based on Ornn leasing data, a16z points out that the actual rental rates for A100, H100, B200, and H200 are currently 9%, 54%, 111%, and 143% above expected levels, respectively. The author argues that demand growth and impairment of the first wave of infrastructure may occur simultaneously, and that the silicon roadmap determines the true pace of chip depreciation.

    Gartner estimates that by 2030, AI data centers will need approximately $8.5 trillion in revenue to cover their investments, but actual AI-related revenue inflows will reach at most only $3.6 trillion, leaving more than $3.5 trillion in investment exposed to asset impairment risk. Based on Ornn leasing data, a16z points out that the actual rental rates for A100, H100, B200, and H200 are currently 9%, 54%, 111%, and 143% above expected levels, respectively. The author argues that demand growth an

    Matched fieldsBody

    #hot-signal#English text is being prepared.#English text is being prepared.#English text is being prepared.
  3. Semiconductor Industry Observer

    The More Agents, the Better Intel Looks: Xeon 6+ and KV Cache Tiering Complete the AI Infrastructure

    At the 2026 Tencent Global Digital Ecosystem Summit, Intel and Tencent Cloud showcased collaboration results centered on Xeon 6+ processors, the CubeSandbox agent sandbox, IAA memory compression, and KV Cache tiered storage. A single machine running CubeSandbox on Xeon 6+ can support the concurrent creation of 350 sandboxes, an improvement of about 46% over the comparison platform; in tests based on Qwen2.5-32B-Instruct, when KV Cache hits backend storage 100% of the time, TTFT drops by 90%, and throughput at concurrency levels of 8/16/32 increases by more than 5x, 6x, and 8x, respectively. Intel also reported third-quarter revenue of $16.5 billion on the same day, with DCAI revenue of $7.3 billion, up 78% year over year.

    At the 2026 Tencent Global Digital Ecosystem Summit, Intel and Tencent Cloud showcased collaboration results centered on Xeon 6+ processors, the CubeSandbox agent sandbox, IAA memory compression, and KV Cache tiered storage. A single machine running CubeSandbox on Xeon 6+ can support the concurrent creation of 350 sandboxes, an improvement of about 46% over the comparison platform; in tests based on Qwen2.5-32B-Instruct, when KV Cache hits backend storage 100% of the time, TTFT drops by 90%, and

    Matched fieldsBody

    #hot-signal#Industry news#Agents#English text is being prepared.