AI Chips

Twitter

Facebook

Copy Link

Key Facts

Key ChipsH100, B200, TPU v5, Trainium

Key PlayersNVIDIA, AMD, Broadcom, Cerebras

Top News

HP Unveils Next-Gen PCs Powered by Nvidia RTX Spark for AI Development and Personal Agents

HP Inc. announced its latest lineup of PCs powered by Nvidia RTX Spark, designed for the next wave of Windows PC experiences. The HP OmniBook Ultra 16 and OmniBook X 14, launching later this year, will be the world's thinnest laptops featuring the RTX Spark superchip with 1 petaflops AI performance. The devices target developers, creators, and gamers who want to run AI personal agents locally with all-day battery life.

Technology2 sources

HP Unveils Next-Gen PCs Powered by Nvidia RTX Spark for AI Development and Personal Agents

Technology2 sources

E Ink and MediaTek partner to bring AI capabilities to next-generation e-readers

E Ink and MediaTek announced an expanded partnership at Computex 2026 to integrate AI-powered features into future e-readers. The collaboration combines E Ink's color display technology with MediaTek's MT8115 and MT8126 processors, promising on-device language translation, document summaries, and improved color ePaper displays. While the technology offers practical benefits for productivity users, it raises questions about whether traditional e-reader fans want these features.

Technology4 sources

E Ink and MediaTek partner to bring AI capabilities to next-generation e-readers

Technology4 sources

Qualcomm unveils Dragonwing IQ10 to simplify building lifelike robots with 700 TOPS AI power

Qualcomm introduced the Dragonwing IQ10 Robotics Reference Design at Computex, offering developers an off-the-shelf platform with 700 TOPS of AI performance and 18 Oryon CPUs. The full-stack robotics platform aims to reduce integration complexity and accelerate the development of industrial and humanoid robots, with early access starting June 2026.

Technology3 sources

Qualcomm unveils Dragonwing IQ10 to simplify building lifelike robots with 700 TOPS AI power

Technology3 sources

Nvidia squeezes trillion-parameter AI supercomputer into desktop workstations for Windows

Nvidia unveiled the DGX Station for Windows at GTC Taipei, bringing data center-class AI infrastructure to individual desks. Built on the GB300 Grace Blackwell Ultra chip with up to 748GB of memory, the system can run trillion-parameter AI models locally. HP, Dell, and other partners will launch versions starting Q4 2026, with prices expected to reach six figures for enterprise deployments.

Technology3 sources

Nvidia squeezes trillion-parameter AI supercomputer into desktop workstations for Windows

Technology3 sources

Intel Crescent Island targets AI inference with cheaper memory, air-cooling to challenge Nvidia

Intel unveils Crescent Island, an AI chip designed for inference workloads that uses cost-effective LPDDR5X memory instead of expensive HBM. The air-cooled GPU supports up to 480 GB of memory and ships later this year as Intel pivots from its failed Gaudi training chips to focus on the growing inference market.

Technology4 sources

Intel Crescent Island targets AI inference with cheaper memory, air-cooling to challenge Nvidia

Technology4 sources

Microsoft Surface Laptop Ultra brings Nvidia RTX Spark chip to redefine AI PCs this fall

Microsoft announced the Surface Laptop Ultra at Computex, featuring Nvidia's Arm-based RTX Spark superchip with up to 128GB unified memory and 1 petaflop of AI compute. Designed for developers and creators, this high-performance portable computer marks Microsoft's most powerful Surface yet, capable of running 120-billion-parameter AI models locally without cloud dependency.

Technology9 sources

Microsoft Surface Laptop Ultra brings Nvidia RTX Spark chip to redefine AI PCs this fall

Technology9 sources

US moves to close loophole that let Chinese firms buy advanced AI chips through overseas subsidiaries

The Commerce Department issued new guidance closing a year-old loophole that allowed Chinese firms to purchase Nvidia's most advanced Blackwell and Rubin processors through overseas subsidiaries. Industry sources estimate hundreds of thousands of chips may have reached Chinese entities during the gap, which emerged after the Trump administration rescinded Biden-era AI Diffusion rules in May 2025.

Policy9 sources

US moves to close loophole that let Chinese firms buy advanced AI chips through overseas subsidiaries

Policy9 sources

Nvidia unveils RTX Spark chip to chase $200B CPU market with AI agent PCs from Microsoft, Dell, and HP

Nvidia launched its RTX Spark chip at Computex in Taiwan, marking a major push into the consumer PC market. The 1-petaflop superchip will power AI agent PCs from Microsoft, Dell, HP, Lenovo, and others this fall. CEO Jensen Huang aims to capture a $200 billion CPU market opportunity by enabling AI agents to replace traditional mouse and keyboard interactions.

Technology24 sources

Nvidia unveils RTX Spark chip to chase $200B CPU market with AI agent PCs from Microsoft, Dell, and HP

Technology24 sources

Acer Veriton RA110 AI Workstation targets Mac mini with AMD Ryzen AI Max+ and local AI power

Acer unveiled the Veriton RA110 AI Mini Workstation powered by AMD Ryzen AI Max+ 395 processors, capable of running 200-billion-parameter AI models locally. The compact desktop PC features up to 128GB RAM, 2TB storage, and 126 TOPS of AI performance, challenging Apple's Mac mini in the professional compact desktop market.

Technology2 sources

Acer Veriton RA110 AI Workstation targets Mac mini with AMD Ryzen AI Max+ and local AI power

Technology2 sources

XCENA raises $135M on a bet that AI's real bottleneck is memory, not compute

XCENA, a chip startup founded by Samsung and SK Hynix veterans, secured $135 million in Series B funding at a $570 million valuation. The company's MX1 chip addresses the AI memory bottleneck by placing compute capabilities directly near DRAM, eliminating costly data round trips between CPUs, GPUs, and memory that slow down AI inference and drive up infrastructure costs.

Technology2 sources

XCENA raises $135M on a bet that AI's real bottleneck is memory, not compute

Technology2 sources

Rivian ditches Apple CarPlay and Android Auto, betting AI infotainment will reshape in-car controls

Electric vehicle maker Rivian is doubling down on its proprietary AI-powered infotainment system, rejecting Apple CarPlay and Android Auto integration. Chief Software Officer Wassym Bensaid says deep AI integration makes screen mirroring obsolete, as customer demand for CarPlay has dropped from over 70% to under 25% in five years. The company's Rivian Assistant aims to control apps through voice interaction rather than traditional buttons or phone mirroring.

Technology4 sources

Rivian ditches Apple CarPlay and Android Auto, betting AI infotainment will reshape in-car controls

Technology4 sources

TSMC targets 30% power savings as AI energy crisis forces chip design rethink

TSMC says surging electricity demands from AI are making energy efficiency the main constraint shaping future chip development. The world's largest contract chipmaker expects to cut power consumption by up to 30% by 2028 while delivering over 20% higher computing performance, as customers from smartphones to AI data centers prioritize performance with lower power usage.

Technology2 sources

TSMC targets 30% power savings as AI energy crisis forces chip design rethink

Technology2 sources

Nvidia pours $6.5 billion into photonics to solve AI's copper bottleneck problem

Nvidia has committed at least $6.5 billion to photonics companies since March, targeting what many see as AI's next critical infrastructure challenge. The chip giant is betting that light-based data transfer will replace copper wiring as the backbone of AI data centers, addressing bandwidth and energy constraints that threaten to slow AI progress.

Technology4 sources

Nvidia pours $6.5 billion into photonics to solve AI's copper bottleneck problem

Technology4 sources

Ambarella and Hanwha forge $800M edge AI partnership spanning video security and robotics

Ambarella has signed a long-term strategic partnership with Hanwha Group valued at over $800 million across more than ten years. The agreement focuses on co-developing edge AI systems-on-chip and software for video security, robotics, industrial automation, and life sciences applications, marking one of Ambarella's largest partnerships to date.

Business2 sources

Ambarella and Hanwha forge $800M edge AI partnership spanning video security and robotics

Business2 sources

Apollo and Blackstone arrange $36 billion debt deal to fund Anthropic's AI chip expansion

Apollo Global Management and Blackstone are orchestrating a $36 billion debt financing package for Anthropic to acquire Google's tensor processing units through a special-purpose vehicle. Broadcom is providing payment guarantees on roughly $31 billion of the senior debt, effectively underwriting demand for its own chips in one of the largest private credit transactions ever assembled around AI infrastructure.

Business3 sources

Apollo and Blackstone arrange $36 billion debt deal to fund Anthropic's AI chip expansion

Business3 sources

Prev 1…353637…92 Next