Despite AI dominating layoff headlines, SignalFire data shows engineers are actually claiming a larger share of new hires, defying displacement fears.

IBM's new sub-nanometer stacked chip architecture supports CPUs, GPUs, mobile chips, and SRAM, with potential to reshape AI hardware design.

Qualcomm confirmed Microsoft will deploy its HBC-based AI accelerators and Meta will use Dragonfly C1000 CPUs, positioning Qualcomm as a serious datacenter AI chip contender.

A new Google Search history update stores images and media from user interactions to train its AI models, with an opt-out option now available.

Chinese AI startup Z.ai (Zhipu AI) plans a dual listing in Shanghai to fund its quest for artificial general intelligence after Anthropic cut off its access.

Two senior Google AI researchers, Jonas Adler and Alexander Pritzel, are leaving for Anthropic, continuing a high-profile talent exodus from Google DeepMind.

a16z led a $33 million seed round in Ornn, a compute marketplace startup building a venue to trade the power and processing capacity that drives AI workloads.

SoftBank Group is pursuing an investment in Tokyo Electric Power to secure electricity supply needed to scale its artificial intelligence infrastructure ambitions.

Qualcomm unveiled its Dragonfly C1000 data center CPU built for agentic AI workloads, with Meta confirmed as its first Big Tech customer starting production in 2028.

Wired reports on meetings with China's leading AI researchers, revealing mutual anxiety over the US-China AI arms race and the risks of unchecked AI development.

Chinese cybersecurity firm 360 says it has built tools capable of matching Anthropic's Mythos AI, a system designed to detect software vulnerabilities.

Anthropic has accused Chinese tech giant Alibaba of using thousands of fraudulent accounts to illicitly extract capabilities from its Claude AI models at scale.

OpenAI and Broadcom debut Jalapeño, OpenAI's first custom AI accelerator chip, purpose-built for inference workloads and developed using OpenAI's own models.