AI Chips

Twitter

Facebook

Copy Link

Key Facts

Key ChipsH100, B200, TPU v5, Trainium

Key PlayersNVIDIA, AMD, Broadcom, Cerebras

Top News

Anker debuts Thus AI chip with compute-in-memory design to power local AI in tiny earbuds

Anker unveiled Thus, its first custom AI chip designed to bring local AI processing to compact devices like earbuds. The compute-in-memory architecture processes neural networks directly inside NOR flash memory cells, enabling larger AI models to run on battery-constrained wearable devices. The chip debuts in Soundcore earbuds on May 21.

Technology5 sources

Anker debuts Thus AI chip with compute-in-memory design to power local AI in tiny earbuds

Technology5 sources

Volkswagen deploys AI agents in Chinese cars to compete with local automakers

Volkswagen announced it will equip new cars built for China with AI agents starting in the second half of 2026, as the German automaker scrambles to catch up with Chinese competitors. The AI voice assistants will handle complex tasks like restaurant bookings and parking, using technology from Alibaba, Tencent, and Baidu. The move is part of Volkswagen's strategy to regain market share in the world's largest car market.

Technology6 sources

Volkswagen deploys AI agents in Chinese cars to compete with local automakers

Technology6 sources

AMD prepares Multi-Frame Generation for Radeon GPUs, closing gap with Nvidia's RTX technology

AMD is preparing to launch multi-frame generation support for its Radeon graphics cards, over a year after Nvidia introduced the technology. A recent update to the FidelityFX SDK reveals new frame generation ratio options, suggesting AMD will finally match competitors in AI-driven frame generation capabilities while addressing concerns about latency performance.

Technology3 sources

AMD prepares Multi-Frame Generation for Radeon GPUs, closing gap with Nvidia's RTX technology

Technology3 sources

Victory Giant surges 60% on Hong Kong debut after raising $2.6 billion in largest IPO this year

Victory Giant Technology, a Chinese supplier to Nvidia, saw its shares rocket nearly 60% on its Hong Kong debut after raising $2.6 billion in the city's largest listing in seven months. The printed circuit board maker attracted 37 cornerstone investors and plans to expand production capacity in mainland China as demand for AI hardware intensifies.

Business3 sources

Victory Giant surges 60% on Hong Kong debut after raising $2.6 billion in largest IPO this year

Business3 sources

John Ternus takes over Apple as CEO: Hardware focus may define the company's AI strategy

Apple announced John Ternus will succeed Tim Cook as CEO this September, bringing 25 years of hardware engineering expertise to the role. As the new CEO, Ternus faces immediate pressure to deliver a killer AI product while navigating tariffs, supply chain challenges, and growing competition. His appointment signals a renewed focus on AI-powered devices rather than competing directly in AI model development.

Business27 sources

John Ternus takes over Apple as CEO: Hardware focus may define the company's AI strategy

Business27 sources

Google to invest up to $40 billion in Anthropic as AI infrastructure race intensifies

Google commits to invest up to $40 billion in Anthropic, with $10 billion immediate and $30 billion tied to performance targets. The deal values Anthropic at $350 billion and follows Amazon's $5 billion investment days earlier. Both tech giants are racing to secure computing capacity for Claude AI models amid surging demand.

Business54 sources

Google to invest up to $40 billion in Anthropic as AI infrastructure race intensifies

Business54 sources

AMD partners with GlobalFoundries for Co-Packaged Optics in Instinct MI500 AI accelerators

AMD is preparing to integrate Co-Packaged Optics into its Instinct MI500 AI accelerators, partnering with GlobalFoundries for photonic integrated circuits and ASE for packaging. The 2027 launch aims to address interconnect bandwidth bottlenecks in AI infrastructure as the silicon photonics race with NVIDIA intensifies.

Technology2 sources

AMD partners with GlobalFoundries for Co-Packaged Optics in Instinct MI500 AI accelerators

Technology2 sources

Google splits eighth-gen TPU into training and inference chips to challenge Nvidia dominance

Google unveiled its eighth-generation Tensor Processing Units at Cloud Next, marking a strategic shift by splitting capabilities into two specialized chips. The TPU 8t targets AI model training with 2.8x performance gains, while TPU 8i focuses on inference with 80% better performance per dollar. Both chips ditch x86 for custom Axion ARM CPUs, signaling Google's push for full-stack efficiency in the agentic era.

Technology24 sources

Google splits eighth-gen TPU into training and inference chips to challenge Nvidia dominance

Technology24 sources

Quantum computers could boost AI by processing large datasets with minimal memory requirements

Researchers have developed a breakthrough method for using quantum computers to boost AI performance on massive datasets. The technique allows quantum machines to process data in smaller batches without requiring impossibly large memory systems. A quantum computer with just 300 logical qubits could outperform a classical computer built using every atom in the observable universe for certain AI tasks.

Science3 sources

Quantum computers could boost AI by processing large datasets with minimal memory requirements

Science3 sources

SK hynix begins mass production of 192GB SOCAMM2 memory for NVIDIA Vera Rubin AI platform

SK hynix has started mass production of its 192GB SOCAMM2 memory module targeting NVIDIA's Vera Rubin AI platform. Built on LPDDR5X technology and 1c-class DRAM process, the module delivers more than double the bandwidth and over 75% improved power efficiency compared to conventional RDIMM modules, addressing critical memory bottlenecks in AI server deployments.

Technology4 sources

SK hynix begins mass production of 192GB SOCAMM2 memory for NVIDIA Vera Rubin AI platform

Technology4 sources

Memory shortage driven by AI demand could keep prices high until 2027 or beyond

The global memory shortage shows no signs of easing, with manufacturers expected to meet only 60% of demand by 2027. AI's appetite for high-bandwidth memory is pushing DRAM and NAND prices to record highs, forcing Samsung to consider its first-ever smartphone loss while consumers face steeper prices for phones, laptops, and other devices.

Business8 sources

Memory shortage driven by AI demand could keep prices high until 2027 or beyond

Business8 sources

Google in talks with Marvell Technology to develop two AI chips, diversifying chip suppliers

Google is negotiating with Marvell Technology to design two new AI chips—a memory processing unit and an inference-optimized TPU—adding a third partner to its custom silicon supply chain alongside Broadcom. The move signals Google's strategy to diversify chip suppliers as inference workloads become the dominant compute cost in AI.

Technology15 sources

Google in talks with Marvell Technology to develop two AI chips, diversifying chip suppliers

Technology15 sources

Cerebras Systems files for IPO after landing $20 billion OpenAI deal and challenging Nvidia

Cerebras Systems, an AI chipmaker building wafer-scale processors, has filed publicly for an IPO months after withdrawing a previous attempt. The company brought in $510 million in revenue in 2025 with a $24.6 billion backlog, including a massive $20 billion deal with OpenAI. But despite 20x revenue growth since 2022, Cerebras remains operationally unprofitable with 86% of revenue coming from just two customers.

Business11 sources

Cerebras Systems files for IPO after landing $20 billion OpenAI deal and challenging Nvidia

Business11 sources

Google negotiates Pentagon deal to deploy Gemini AI in classified military environments

Google is in talks with the U.S. Department of Defense to deploy Gemini AI models in classified settings for the first time. The negotiations include deploying tensor processing units inside secure environments and adding GPU capacity to Google Distributed Cloud, which holds DoD Impact Level 6 authorization. Google pushes for contract language prohibiting domestic mass surveillance and autonomous weapons without human oversight.

Technology3 sources

Google negotiates Pentagon deal to deploy Gemini AI in classified military environments

Technology3 sources

Upscale AI targets $2 billion valuation in third funding round since September launch

AI infrastructure startup Upscale AI is negotiating a new funding round that would value the company at $2 billion, just seven months after emerging from stealth. The Santa Clara-based company aims to raise $180-200 million despite not yet releasing a product, betting on custom chips and full-stack infrastructure solutions to transform how AI data centers operate.

Startups2 sources

Upscale AI targets $2 billion valuation in third funding round since September launch

Startups2 sources

Prev 1…232425…70 Next