AI Chips

Twitter

Facebook

Copy Link

Key Facts

Key ChipsH100, B200, TPU v5, Trainium

Key PlayersNVIDIA, AMD, Broadcom, Cerebras

Top News

AMD FSR 4.1 AI upscaling expands to older Radeon RX GPUs, bringing ML-powered visuals to 300+ games

AMD confirms FSR 4.1 AI upscaling will arrive on Radeon RX 7000 series GPUs this July, followed by RX 6000 series in early 2027. The machine learning-powered technology, previously exclusive to RX 9000 cards, will support over 300 games and deliver sharper visuals without hardware upgrades. AMD optimized the model for INT8 processing on older architectures.

Technology4 sources

AMD FSR 4.1 AI upscaling expands to older Radeon RX GPUs, bringing ML-powered visuals to 300+ games

Technology4 sources

TSMC raises semiconductor market forecast to $1.5 trillion by 2030 as AI demand surges

TSMC now expects the global semiconductor market to exceed $1.5 trillion by 2030, up from a previous $1 trillion estimate. The world's largest contract chipmaker attributes this 50% upward revision to surging AI silicon demand, with AI and high-performance computing projected to account for 55% of the market. The company is accelerating capacity expansion with nine new wafer fabs planned for 2026.

Business3 sources

TSMC raises semiconductor market forecast to $1.5 trillion by 2030 as AI demand surges

Business3 sources

Intel confirms Googlebook partnership, bringing x86 support to Google's Gemini AI laptops

Intel has officially confirmed its partnership with Google for the newly announced Googlebook lineup, marking a significant shift in the AI laptop landscape. Google VP John Maletis revealed that these premium AI-first devices will ship with processors from Intel, Qualcomm, and MediaTek, suggesting the new operating system will support both x86 and Arm architectures. The first wave launches this fall from major manufacturers including HP, Dell, Acer, Asus, and Lenovo.

Technology6 sources

Intel confirms Googlebook partnership, bringing x86 support to Google's Gemini AI laptops

Technology6 sources

Fractile raises $220 million for AI inference chips as Anthropic eyes UK silicon

London-based Fractile secured $220 million in Series B funding to develop AI inference chips that promise to run frontier models 25 times faster at one-tenth the cost of current GPUs. The round, led by Accel with former Intel CEO Pat Gelsinger joining as an angel investor, comes as Anthropic enters early discussions to become a customer when the chips launch in 2027.

Startups2 sources

Fractile raises $220 million for AI inference chips as Anthropic eyes UK silicon

Startups2 sources

Tower Semiconductor signs $1.3 billion AI chip deals as silicon photonics demand surges

Tower Semiconductor landed $1.3 billion in silicon photonics deals for 2027, driven by surging AI chip demand for data centers. The Israeli contract chipmaker raised its Q2 revenue forecast to $455 million and nearly doubled operating profit to $64.57 million, while securing $290 million in customer prepayments. The company targets $2.8 billion in annual revenue by 2028.

Business3 sources

Tower Semiconductor signs $1.3 billion AI chip deals as silicon photonics demand surges

Business3 sources

China threatens countermeasures as US chip equipment bill targets AI semiconductor production

China is pushing back hard against the proposed MATCH Act, legislation that would restrict Chinese chipmakers from producing AI semiconductors by limiting access to critical chipmaking equipment. Beijing has summoned U.S. diplomats, unveiled countermeasures, and is expected to raise the issue during high-level talks this week between Presidents Trump and Xi Jinping.

Policy2 sources

China threatens countermeasures as US chip equipment bill targets AI semiconductor production

Policy2 sources

Lenovo ThinkPad debuts Ryzen AI 400 chip with rare 9/10 repairability score from iFixit

Lenovo has completed its 2026 ThinkPad lineup with the X13 Gen 7, marking the debut of AMD's Ryzen AI 400 processors. Starting at $1,499 and weighing just 2.05 pounds, the ultralight laptop delivers 50 TOPS AI processing with Thunderbolt 4 connectivity. The device earned a rare 9/10 repairability score from iFixit, with replaceable battery, SSD, and WWAN modules.

Technology2 sources

Lenovo ThinkPad debuts Ryzen AI 400 chip with rare 9/10 repairability score from iFixit

Technology2 sources

Sipeed K3 RISC-V SBC runs 30B AI LLMs locally at 15 tokens per second for $299

Sipeed has launched its K3 series RISC-V-powered single board computers that can run 30B-parameter AI LLMs locally at up to 15 tokens per second. Built with SpacemiT's Fusion Architecture and featuring up to 32GB LPDDR5 memory plus a 60 TOPS NPU, the K3 offers an accessible open-source alternative to proprietary AI hardware starting at $299.

Technology2 sources

Sipeed K3 RISC-V SBC runs 30B AI LLMs locally at 15 tokens per second for $299

Technology2 sources

Major exchanges launch futures markets for AI computing power as GPU costs become tradable asset

CME Group and Intercontinental Exchange are launching the first futures markets for AI computing power, allowing investors and tech firms to hedge against volatile GPU rental costs. CME partners with Silicon Data while ICE teams with Ornn to create cash-settled contracts based on GPU pricing indices. The move signals that compute has evolved into a major tradable asset class.

Business4 sources

Major exchanges launch futures markets for AI computing power as GPU costs become tradable asset

Business4 sources

Micron samples 256GB DDR5 memory at 9200 MT/s, targeting AI infrastructure with major efficiency gains

Micron has begun sampling its 256GB DDR5-9200 RDIMM server modules built on 1-gamma DRAM technology, offering speeds over 40% faster than current production modules. The advanced memory solution uses 3D stacking and through-silicon vias to deliver higher capacity and bandwidth while reducing power consumption by more than 40% compared to dual 128GB configurations.

Technology3 sources

Micron samples 256GB DDR5 memory at 9200 MT/s, targeting AI infrastructure with major efficiency gains

Technology3 sources

Cohu secures $5M in orders for GaN power device testing as AI data centers drive demand

Cohu Inc. announced multiple orders totaling $5 million for its DiamondX testing platform from a semiconductor manufacturer. The systems will test gallium nitride power devices designed for AI data center power architectures, marking Cohu's expansion into the rapidly growing AI infrastructure market where GaN devices offer efficiency advantages.

Business2 sources

Cohu secures $5M in orders for GaN power device testing as AI data centers drive demand

Business2 sources

NVIDIA and Elon Musk race to put AI data centers in space as Earth-based limits bite

NVIDIA is partnering with five space companies to build orbital AI infrastructure, while Elon Musk's SpaceXAI teams with Anthropic for multi-gigawatt space facilities. The push comes as terrestrial data centers face severe power, water, and land constraints—with some projects facing seven-year permit delays.

Technology3 sources

NVIDIA and Elon Musk race to put AI data centers in space as Earth-based limits bite

Technology3 sources

Intel CEO confirms Intel and Nvidia collaboration on new products is actively progressing

Intel CEO Lip-Bu Tan publicly confirmed that Intel and Nvidia are actively developing new products together, eight months after their historic partnership announcement. The collaboration spans AI infrastructure platforms and consumer devices, with the first joint products potentially arriving by 2026. Intel's stock rose following the announcement.

Technology2 sources

Intel CEO confirms Intel and Nvidia collaboration on new products is actively progressing

Technology2 sources

HrdWyr Secures $13M Series A to Build AI-Native Chips for Physical AI Era

Indian semiconductor startup HrdWyr has closed a $13 million Series A funding round led by Ideaspring Capital to develop AI-native System-on-Chip products. The company is building purpose-built chips for Physical AI applications across white goods, EVs, and data centers, marking a shift from conventional general-purpose architectures to domain-specific edge intelligence.

Startups2 sources

HrdWyr Secures $13M Series A to Build AI-Native Chips for Physical AI Era

Startups2 sources

Applied Materials and TSMC join forces to accelerate AI chip development at $5 billion EPIC Center

Applied Materials and TSMC have announced a partnership to fast-track AI chip development at the EPIC Center in Silicon Valley. The collaboration builds on 30 years of joint work and focuses on materials engineering, equipment innovation, and process integration to deliver energy-efficient chips for data centers and edge devices.

Technology2 sources

Applied Materials and TSMC join forces to accelerate AI chip development at $5 billion EPIC Center

Technology2 sources

Prev 1…262728…79 Next