Share
Linkedin
Twitter
Facebook
Whatsapp
Copy Link
The biggest US tech firms now plan to spend up to $725 billion this year on AI infrastructure, with Meta, Alphabet, and Microsoft all raising their capital expenditure guidance. Meta increased its spending ceiling to $145 billion, citing higher memory costs, while investor skepticism sent its stock down 6% despite strong revenue growth.
The US Department of Commerce has directed chip equipment suppliers including Lam Research, Applied Materials, and KLA to stop shipping tools to Hua Hong facilities developing 7-nanometer chipmaking capabilities. The move targets China's second-largest chipmaker as it prepares to produce artificial intelligence chips, potentially costing US suppliers billions while escalating tech tensions ahead of the Trump-Xi summit in May.
LG Electronics and Nvidia are in talks over potential AI collaboration spanning robotics, AI data centres, and mobility solutions. The discussions follow a visit by Madison Huang, Nvidia's senior director for physical AI platforms, to LG's Seoul headquarters where CEO Ryu Jae-cheol attended the meeting directly.
Nvidia CEO Jensen Huang reveals the company now holds zero market share in China's AI accelerator market, down from 66% in 2024. Huawei fills the void with its Ascend 950 chips, expecting $12 billion in revenue this year as US export controls drive Chinese tech firms toward domestic alternatives. The shift highlights how regulatory barriers are reshaping the global AI hardware landscape.
Nvidia's expansion into physical AI has intensified its reliance on Asian partners, with Asian suppliers now accounting for 90% of production costs compared to 65% last year. The shift extends beyond semiconductors into robotics and autonomous systems, triggering stock rallies across the region as companies from LG Electronics to Taiwan's Nanya Technology benefit from deepening partnerships with the chip giant.
ACEMAGIC has launched the F5A mini PC, a compact desktop system powered by AMD's Ryzen AI 9 HX 470 processor delivering 86 TOPS of combined AI performance. The palm-sized device features OCuLink support for external GPU connectivity, dual USB4 ports with 40Gbps bandwidth, and Wi-Fi 7, targeting AI-enabled workloads and high-performance computing in a small form factor starting at $667.
AMD's latest research reveals that over four in five organizations are now planning, piloting, or deploying AI PCs, marking a significant shift from experimentation to real-world implementation. The data shows 70% report faster performance with dedicated AI PCs, while 66% see productivity gains as agentic AI drives demand for localized computing power.
Nvidia has secured 760,000 square feet of office space in Bengaluru through a 10-year lease valued at over INR 12 billion, marking India's largest single-tenant office commitment to date. The deal nearly doubles the chipmaker's footprint in India to over 1.13 million square feet across six cities, signaling a long-term bet on India's engineering talent to meet surging global AI demand.
Intel released a new driver for Arc Pro GPUs that allows users to allocate up to 93% of system RAM to integrated GPUs, up from the previous 87% limit. This memory allocation breakthrough enables users to run substantially larger Large Language Models on affordable hardware without hitting memory capacity constraints.
Intel's AI PC Innovation Day in Bengaluru showcased a fundamental shift in computing: moving AI workloads from cloud to device. The event highlighted BharatGPT Mini 2 running entirely offline on Intel Core Ultra processors, emphasizing privacy, speed, and accessibility for diverse markets like India where connectivity and language support remain critical factors.
Qualcomm stock surged following analyst reports that the chipmaker is collaborating with OpenAI and MediaTek to develop processors for an AI-first smartphone. The reported partnership, with mass production expected in 2028, marks a significant development in OpenAI's hardware ambitions and could signal smartphones retaining their central role in the AI era.
Industry analyst Ming-Chi Kuo reports OpenAI is developing an AI smartphone in collaboration with Qualcomm and MediaTek, with Luxshare as manufacturing partner. The device would replace apps with AI agents that handle tasks directly, targeting 300-400 million annual shipments by 2028—volumes that would rival Apple's iPhone.
Japan's NEDO has selected the ZAM memory development project for government subsidies, backing Intel and SoftBank subsidiary SAIMEMORY's effort to create a power-efficient alternative to HBM. The next-generation memory technology promises 40% lower power consumption and aims for mass production by 2029, marking Japan's strategic push to re-enter the semiconductor market.
Meta has inked a major multi-billion dollar agreement with Amazon to deploy tens of millions of AWS Graviton CPU cores across 32 data centers over three years. The deal highlights a critical shift in AI infrastructure as agentic AI workloads drive unprecedented demand for CPUs, not just GPUs, exposing supply constraints across the industry.
Intel reported Q1 revenue of $13.6 billion, beating Wall Street expectations by over $1 billion as demand for CPUs in AI inference workloads surges. CEO Lip-Bu Tan says the shift from training to inference is changing the CPU-to-GPU ratio, with agentic AI potentially flipping it entirely. The company's share price jumped over 20% in after-hours trading, approaching record highs not seen since the dotcom era.
Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Follow topics that matter to you and stay ahead.