AI Infrastructure news and analysis on Web Pulse: 35713 curated articles tracking the latest AI Infrastructure developments, tools, and research, updated continuously from vetted sources.
Alibaba released the model weights for Qwen3.8-Flash-Next, a multimodal mixture-of-experts model with 125B parameters plus 51B N-gram embeddings and 6B activated per token, as a preview of the Qwen4 architecture. NVIDIA …
Nvidia's dominance in the AI chip market faces mounting pressure from rivals including AMD, Intel, and hyperscalers' custom silicon, though its CUDA software ecosystem and scale provide a strong defense. Joy Yang, Head o…
FreeToken, an LLM engine designed to maximize bandwidth across all hardware, faces a fundamental challenge: bit-exact reproducibility is impossible across different CPU and GPU architectures due to differing floating-poi…
A two-person startup founder reported that using Grok Bot, an AI workforce tool from SpaceXAI, generated approximately $1,000 in on-demand usage within 24 hours, translating to a $72,000-per-month run rate. The founder, …
A user seeking advice on building a local LLM/AI setup at home is considering three hardware options: the Minisforum MS-S1 Max at $4,600, the ASUS Ascent GX10 at $6,100, and a new Mac with 128GB at $7,100. The user wants…
A new benchmark suite, AgentSysBench, reveals that agentic AI workloads differ fundamentally from conventional LLM serving, with non-LLM components dominating latency in 5 of 10 applications and sandbox memory peaking at…
Aiden AI has shipped context-window recovery infrastructure in its firmware across three merged pull requests (#497, #498, #530), implementing context compression, session switching, and saved-result-file recovery to han…
Microsoft Corp. (NASDAQ: MSFT) closed its biggest fiscal year with annual revenue of $331 billion, up 18%, and Azure revenue surpassing $100 billion, up 41%, yet shares trade at $491.71, up just 2.32% year-to-date. The c…
Anthropic is set to lease a data center from NScale in a deal valued at $45 billion, adding to its total compute commitments that have surpassed $150 billion as of mid-2026. The Claude maker has also signed a roughly $19…
MiniMax's revenue nearly quadrupled to $116.6 million in the first half of 2026, up 283.1% from $30.4 million a year earlier, but its adjusted net loss widened 111.2% to $293 million. The Shanghai-based AI lab, founded i…
Applied Materials Inc. reported $9.12 billion in fiscal Q3 2026 revenue, up 25% year-over-year, but projected a $600 million revenue hit from tightened US export controls, with China's share of sales falling to 28% from …
Lovable, an AI app builder valued at $13.3 billion, is rolling out Model Context Protocol (MCP) capabilities to let users expose app functions as agent-callable endpoints, aiming to make web apps natively accessible to A…
MSI's Vector 16 HX AI gaming laptop, featuring the Core Ultra 7 255HX processor and RTX 5070 Ti GPU, is available on Amazon with a discount coupon valid until September 9, 2026, reducing the price to $3,068.07. The lapto…
Leaked images posted on X by user @hsuchingpo show Apple's M5 AI server, believed to be the Private Cloud Compute system, revealing a rack-mountable chassis holding up to 32 boards with Apple silicon and NAND flash. The …
SD Times reports that design documents and specifications often misrepresent the actual state of production infrastructure, creating false certainty for developers. The article emphasizes that repositories, architecture …
Goldman Sachs flow experts report that US large-cap mutual funds are underweight Nvidia by roughly 100 basis points, the largest underweight among major AI stocks, ahead of Nvidia's Q2 FY2027 earnings. The de-risking acc…
Lumentum Holdings (NASDAQ:LITE) stock rose 4% to $918.10 midday Wednesday, extending a second consecutive session of gains without any fresh company news, while the broader chip complex lagged. The iShares Semiconductor …
SK hynix Inc. published a co-packaged-optics roadmap in Nature Electronics on August 20, describing an AI architecture where optical links directly connect processor and memory pools, expanding the potential market for C…
Neo4j published a startup guide for building graph applications on its Aura cloud database, recommending installation of the Neo4j CLI and agent skills first, then using the neo4j-getting-started-skill to scaffold a proj…
Super Micro Computer's stock surged 9.35% to close at $38.46 on August 25 after Cisco Systems announced it will integrate Supermicro's liquid- and air-cooled rack-scale AI servers into its Secure AI Factory platform, add…