China’s AI chip makers out-earn the models as non-GPU accelerators take the lead

Two shifts are accelerating in China’s AI chip market. Nvidia’s share is being eaten by domestic makers, and non-GPU accelerators are rising. IDC data for the first half of 2026 shows China’s data-centre AI accelerator shipments reached 2.008 million units, with non-GPU at 1.016 million, surpassing GPU for the first time. Nvidia still leads at … Read more

Microsoft’s Maia 300 plan signals Big Tech’s quiet break from Nvidia

On the evening of 12 October, Microsoft chairman and CEO Satya Nadella outlined the company’s latest in-house chip plan: a large-scale deployment of Azure Maia 300 in 2027. Microsoft shares rose 1.14 per cent to US$541.15 that day, pushing its market value back above US$4 trillion. The Maia case Nadella’s point is that everyone is … Read more

Silicon Valley’s price war: Claude Haiku 5.5 and free GPT-6 Luna undercut DeepSeek

For two years the “price butcher” of large models was never Claude. DeepSeek drove per-million-token prices to cents and trained Chinese vendors to match. On 8 October the positions flipped. Anthropic launched Claude Haiku 5.5, a lightweight model for everyday tasks, at USD 0.10 in and USD 0.50 out per million tokens for requests under … Read more

DeepSeek nears RMB 80B raise, sets stage for 2027 IPO; Tencent and CATL lead

DeepSeek is close to finalising a new funding round of at least RMB 80 billion, with Tencent and CATL among the largest committed investors, Bloomberg reported on 6 October, in a deal that paves the way for an initial public offering early in 2027. People familiar with the terms said the final total could approach … Read more

Huawei’s Richard Yu: Domestic Teams Are Fighting to Develop EUV Lithography

At a recent media briefing, Richard Yu, chairman of Huawei’s Consumer Business Group, shared the latest on terminal business and the HarmonyOS ecosystem. Yu said Huawei once followed a global path, cooperating with US, European, Japanese, South Korean and Taiwanese firms, but since 2019, to keep the business alive, Huawei joined domestic supply chains to … Read more

US Blocked the Chips, Then Watched China’s Data Centre Capacity Climb to 24 GW

The United States has long banned its advanced AI chips from reaching China, fearing Chinese labs would use Nvidia silicon to build frontier AI. But Washington miscalculated one thing: blocking the chips does not block the compute. US Blocked the Chips, Then Watched China’s Data Centre Capac (Source: Sohu) One analysis of global data centres … Read more

Tsinghua AIR Founding PhD Launches BASAL Intelligence on Action-Native Embodied Model

LeiPhone WhaleX has learned that Jianxiong Li, the first doctoral graduate of Tsinghua’s Institute for AI Industry (AIR), has founded embodied intelligence company BASAL Intelligence, with a round led by FiveYuan Capital and participation from IVC, Linear Capital and WestSummit Capital. AIR associate professor Xianyuan Zhan serves as chief scientist, with a core team of … Read more

Simplexity Robotics: One Robot Tending Two CNC Lathes at 0.5 mm, 100 per cent Success

Could a robot build the parts needed to assemble a robot? That is the hardest real world test for embodied intelligence. A robot’s joints, chucks and precision structs are cut by CNC lathes and mills. If a robot can tend those machines, grabbing, loading, inserting and removing workpieces, robot production feeds itself: more robots, lower … Read more

Faraday Future Ships 265 EAI Robots in September, Cumulative Total Hits 817

In September, Faraday Future (FFAI) and FFR sold and shipped 265 EAI robot units, a single month record. In the third quarter the robot business moved 575 units, and since late February the cumulative total has reached 817. On 14 October US time, FFAI and FFR plan to launch an education ecosystem brain, Skills products … Read more

Mega-Kernel vs Super-Operator: Inside the Trillion-Parameter Inference War

How do you make a 2.8 trillion parameter model run fast? Recently, teams on both sides of the Pacific fixed their eyes on Kimi K3, the world’s largest open source model by parameter count. With 2.8 trillion total parameters and 104 billion activated, K3 raised the ceiling for open models, but ordinary AI servers struggle … Read more