Share
Linkedin
Twitter
Facebook
Whatsapp
Copy Link
Cerebras Systems raised $5.5 billion in its IPO, valuing the AI chipmaker at roughly $60 billion. But the company's success story nearly ended in 2019 when it was burning $8 million monthly trying to solve a technical problem that had stumped the semiconductor industry for decades. The breakthrough came in July 2019 when the team finally cracked the packaging challenge for its dinner plate-sized chip.
YouTuber Hardware Haven converted a $100 Nvidia V100 data center GPU with SMX interface into a PCIe card using a custom adapter and 3D-printed cooling. The modded Tesla V100 delivered 130 tokens per second in AI inference tests, outperforming both the RTX 3060 and RX 7800 XT while proving more efficient than modern midrange offerings at just 0.55 tokens per second per watt.
Bangkok-based technology firm SiamAI has formally rejected allegations that it circumvented U.S. export restrictions to ship advanced AI servers to China. U.S. prosecutors claim at least $2.5 billion in American AI technology was illegally shipped, with more than $500 million moved between April and mid-May 2025. The case highlights growing tensions over semiconductor controls as Thailand emerges as a critical digital infrastructure hub.
US prosecutors identified Bangkok-based OBON Corp as the intermediary in a massive smuggling operation that routed restricted Nvidia AI GPUs to China. The $2.5 billion scheme allegedly involved Supermicro executives using falsified documents and dummy servers to bypass export controls, with Alibaba reportedly among the end customers receiving the banned hardware.
Global tech giants are making rare financial offers to SK Hynix, proposing to invest in new production lines and fund expensive manufacturing equipment purchases as they scramble to secure memory chips. The chipmaker remains cautious, with insiders warning that available capacity is essentially zero despite the artificial intelligence boom driving demand.
Qualcomm stock jumped nearly 15% as investors rallied behind the chipmaker's strategic pivot into AI. The surge followed reports of an OpenAI partnership for AI-native smartphone chips, accelerated data center chip shipments to a major hyperscaler, and strong second-quarter earnings. With new Snapdragon platforms and a $2.3 billion acquisition to expand data center capabilities, Qualcomm is transforming from a smartphone chip vendor into a diversified AI computing company.
AMD unveiled the Instinct MI350P, a PCIe AI accelerator card with 144GB of HBM3E memory designed for air-cooled enterprise servers. The dual-slot card delivers up to 4,600 TFLOPS performance and outpaces Nvidia's H200 NVL by roughly 40% in FP16 and FP8 theoretical compute. With support for up to eight cards per system, AMD targets enterprises seeking scalable AI infrastructure without major platform overhauls.
A fresh investor frenzy has shifted the AI bull run's focus to Asia, with Seoul becoming the world's hottest stock market. Samsung, TSMC, and SK Hynix are posting record earnings, with chip revenues at Samsung leaping nearly 50 times last quarter. South Korea's KOSPI has doubled in six months as Asian chipmakers cement their pivotal role in the global AI supply chain.
Samsung semiconductor workers have rejected a $340,000 one-time bonus offer, demanding annual profit-sharing bonuses like SK Hynix's $900,000 payouts. With over 50,000 workers threatening an 18-day strike starting May 21, the labor dispute could cost Samsung up to $11.7 billion and disrupt the global AI hardware industry at a time when high-bandwidth memory chips are in critically short supply.
Arm doubled its AGI CPU sales forecast to $2 billion across fiscal 2027-2028, yet analysts reveal this represents just 4% of the server CPU market. The chip designer now confronts a critical challenge: securing enough semiconductor supply chain capacity at TSMC to meet surging AI data center demand while competing against Intel and AMD.
Elon Musk's SpaceX has filed plans for a massive semiconductor manufacturing facility in Grimes County, Texas, with an initial investment of $55 billion that could eventually reach $119 billion. The Terafab project aims to produce AI chips for Tesla vehicles, humanoid robots, and orbital data centers, marking SpaceX's ambitious entry into chip manufacturing despite having zero experience in the field.
Over 600 million Chinese citizens now use generative AI, marking a 142% annual increase. From booking travel to ordering food and managing health, AI in China has rapidly moved from experimental to essential. As Chinese AI models now consume more data weekly than U.S. models, this mass AI adoption could shape how AI tools are deployed globally.
Goldman Sachs upgraded AMD to buy from hold with a price target of $450, while Baird raised its target to $625. The chipmaker's server CPU total addressable market is now expected to exceed 35% CAGR through 2030, reaching $120 billion, driven primarily by agentic AI workloads that are reshaping enterprise computing infrastructure.
Infineon Technologies upgraded its 2026 financial outlook as AI demand for power supply solutions surged, expecting €1.5 billion in AI data center revenue. The German chipmaker posted Q2 revenue of €3.81 billion, up 6% year-over-year, while net profit rose to €301 million. CEO Jochen Hanebeck cited the AI boom as a key driver, though investors remain cautious about margin pressure and weakness in the electric vehicle segment.
DeepSeek, the Chinese AI lab known for cost-effective large language models, is raising its first venture capital funding at a valuation that skyrocketed from $20 billion to $45 billion in just weeks. China's state-backed semiconductor fund is leading the round, signaling Beijing's strategic push for AI self-sufficiency amid US technology restrictions.
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.