Frontier models, infrastructure, and the policy that shapes them. Reported across Seoul, Singapore, Bengaluru, and Hangzhou.
The company's nanostack architecture packs 100 billion transistors onto a fingernail-sized die, but Asia's fab timelines suggest commercialization won't arrive before 2031.
ZTE's AIR Core prototype shows how carriers plan to sell AI traffic as a service, moving beyond bandwidth into computing, identity, and real…
Consulting giants and multinationals are quietly pulling back on unlimited inference access after discovering that low-value tasks can burn …
Jalapeño, co-designed with Broadcom, represents a strategic pivot toward vertical integration in AI infrastructure - and a quiet admission t…
While OpenAI, Anthropic, Google, xAI and Microsoft have begun submitting unreleased models to the Commerce Department's new review unit, the…
Squidbleed, a 29-year-old memory leak in widely deployed caching software, went unnoticed through three decades of enterprise use - until a …
Developers have announced 160 GW of new capacity by 2032, but half of what's due next year hasn't broken ground - and the math may not even …
Los Alamos labs will debut autonomous agent-driven clusters as GPU maker unveils Vera Rubin platform, three software tools, and a pitch to t…
A generative AI benchmarking platform reveals China's lead across robotaxi, trucking, and delivery categories, challenging assumptions about…
The text-to-image AI company is building ultrasonic scanning spas and targeting a 2031 global rollout - a hardware gambit that raises questi…
As Liu Qiangdong tells APEC that autonomous delivery will eliminate frontline jobs, the company's "Nirvana Plan" bets on retraining workers …
John Jumper's departure marks the second high-profile exit from Google's AI unit this week, raising questions about talent retention at the …
The startup has taped out a 3nm chip that sidesteps conventional multiplication - but the real test lies in software maturity and deployment…
A US government order to pull two AI models over security fears may have inadvertently amplified the very brand it sought to constrain - rai…
Amazon's security VP argues that asking people to approve every automated decision creates a false sense of safety - and that accountability…
Coherent's $650 million Texas expansion reflects a wager that copper's days at hyperscale are numbered - and that photonics supply chains mu…
The San Francisco AI lab's latest update lets its design assistant read local codebases and hand off work directly to the coding agent - a w…
The chipmaker has begun limited production of an enhanced node variant promising nine percent performance gains, betting foundry customers w…
The AI company paused plans to impose API-style metered fees on automation toolkit users just hours before the change was set to take effect…
A 952-node DOE cluster sidesteps InfiniBand and Slingshot with revived Intel-era interconnect tech, now delivering 400 Gbps and 91% scaling …
Researchers at UC Davis have kept a brain-computer interface running for three years in daily use, achieving 99 percent accuracy in controll…
From AWS to Google Cloud, the industry's largest operators are deploying custom Arm chips to balance performance, power, and agentic AI dema…
As Elon Musk floats a wafer factory capable of one million units monthly, the world's leading foundry signals confidence in its position and…
While competitors slash token pricing, the Shenzhen giant is betting on a vertical-integration strategy that bundles compute, tooling, and s…
The financial super app's pivot to conversational AI could set a new template for how a billion-plus users interact with digital payments an…
