AI chips

Samsung plans to more than double production of its HBM4 and HBM4E high-bandwidth memory DRAM next year, according to industry sources cited by South Korean outlet Sedaily. The expansion targets booming demand from AI chip makers, as Samsung races rivals like SK hynix to supply next-generation memory for processors from Nvidia and other customers.

Taiwan's dominance in semiconductor manufacturing, led by TSMC, is being described as a 'silicon shield' that deters Chinese aggression while making the island indispensable to the United States in the AI era. As artificial intelligence drives global demand for advanced chips, control over Taiwan's chip industry has become central to both its security and US strategic interests, intensifying debate over cross-strait tensions.

Reporting from Taipei frames Taiwan as the potential deciding factor in the escalating technology rivalry between the United States and China. Taiwan's tech industry produces more than 90 percent of the world's most advanced semiconductors, the chips that power AI servers, giving the island outsized strategic importance as both superpowers compete for dominance in artificial intelligence.

AT&S, the Austrian substrate manufacturer, is featured in a new report on the growing competition to produce advanced IC substrates, the critical layer that connects AI chips to circuit boards. The technology is becoming a bottleneck as demand for AI accelerators surges, and companies and governments are racing to expand capacity and secure supply of these components.

The United States and China have struck a trade agreement reportedly worth $30 billion, covering American purchases of Chinese chips and toys alongside tariff adjustments. The deal also establishes an AI safety hotline between the two governments, a step aimed at reducing risks around advanced artificial intelligence. The accord signals a partial easing of tensions in the ongoing trade conflict between Washington and Beijing.

Data centers, power plants and chip facilities for artificial intelligence are being built across the United States at a pace that could surpass any previous infrastructure boom in the country's history. Analysts point to massive capital commitments from tech companies racing to secure computing power, raising questions about electricity demand, grid strain and whether the investment will pay off.

Wall Street analysts are divided over whether AMD or Broadcom will come out ahead in the next phase of the AI chip market. The debate centers on how each company is positioned to capture growth as demand for artificial intelligence hardware broadens beyond the current generation of accelerators, with investors weighing their differing strategies and customer bases.

Major technology companies are increasingly using guarantees to support debt raised for AI data centres and chips, issuing up to $300 billion in commitments within a year. Analysts note that much of this exposure is not recorded on their balance sheets, raising questions about transparency and hidden financial risk as AI infrastructure spending accelerates.

Huawei's Ascend chips have overtaken Nvidia in China's AI chip market, according to a new report. The shift reflects the impact of US export controls that have restricted Nvidia's access to Chinese customers, pushing domestic buyers toward Huawei's homegrown alternatives. Analysts see it as a milestone for China's semiconductor self-sufficiency push, though questions remain about Ascend's performance and production capacity at scale.

Nvidia chief executive Jensen Huang has dismissed predictions that artificial intelligence could destroy humanity within the decade, saying there is a '0% chance' of such an outcome by 2030. His comments push back against warnings from AI safety advocates, including Anthropic, who have cautioned about catastrophic risks from increasingly powerful systems. The remarks are drawing attention given Nvidia's central role in supplying the chips powering the AI boom.

A Brookings Institution researcher argues that the current AI buildout will amount to the largest infrastructure investment bet in United States history, surpassing previous eras of massive spending on railroads, highways, and telecommunications. The claim underscores the enormous scale of capital flowing into data centers, chips, and energy capacity to support artificial intelligence development.

India's Finance Minister Nirmala Sitharaman has urged increased investment in artificial intelligence and semiconductor manufacturing. She positioned the sectors as key to India's economic and technological ambitions, aligning with the government's push to attract chip fabrication projects and build AI capacity. The remarks come as countries compete for investment in these strategic industries.

Morningstar argues that the popular claim that artificial intelligence spending is driving inflation deserves a more nuanced look. The piece pushes back on the simple narrative that data center buildouts and AI investment are broadly pushing up consumer prices, suggesting the picture is more complicated than headlines suggest. It comes amid ongoing debate over how much AI infrastructure spending is affecting energy demand, chip costs and the wider economy.

Microsoft announced expanded Azure transparency reporting, new Surface devices and its own custom AI chips, strengthening its cloud and hardware lineup. The company is positioning the chips to reduce reliance on outside suppliers and to lower costs for AI workloads running across its data centres. Coverage focuses on how the moves fit Microsoft's broader push into artificial intelligence infrastructure alongside rivals such as Amazon, Google and Nvidia.

Chinese AI startup Moonshot is reportedly seeking access to additional Nvidia Blackwell chips, according to a report circulating this week. The news comes just a week after a Trump administration official flagged an export control breach involving Nvidia's advanced chips. The development highlights continuing tensions over US restrictions on selling cutting-edge semiconductors to Chinese companies, a dispute that has become a central flashpoint in the technology rivalry between Washington and Beijing.

New AI chips are roughly ten times more energy efficient than previous generations, but total electricity demand from data centers keeps rising sharply. The paradox driving discussion is that efficiency gains are being overwhelmed by explosive growth in AI workloads, as more powerful models and wider deployment mean operators keep adding servers, pushing overall power consumption higher despite better hardware.

ByteDance, the Chinese parent company of TikTok, gained access to more than 2,000 Nvidia B200 chips through Nscale's data center in Norway, according to a new filing. A Singaporean subsidiary called Spring accounted for 73% of the UK-based neocloud's 2025 revenue. The report is drawing attention because the US restricts exports of advanced Nvidia AI chips to China, raising questions about how US-made hardware reached a Chinese firm indirectly.

Qualcomm has announced two new smartphone chips as it deepens its push into on-device artificial intelligence, positioning itself against rivals in the mobile processor market. Despite the product launch, the company's stock edged lower in after-hours trading, suggesting investors were not immediately impressed or had expected more from the announcement.

Micron is reporting its fiscal first-quarter 2026 earnings, with attention focused on whether a boost from Nvidia's latest chip launch will reach the memory maker later than expected. Analysts and investors are watching Micron's guidance for signs that demand tied to Nvidia's new chips may take longer to translate into sales, a factor that could shape sentiment across the semiconductor sector.

Huawei has announced plans for two advanced AI chips slated for 2027, putting Nvidia stock in focus as investors weigh rising Chinese competition in the AI semiconductor market. Traders are watching how Huawei's roadmap could challenge Nvidia's dominance in AI accelerators and what it means for US export restrictions and Nvidia's China business.

Google has announced Project Suncatcher, an initiative to test its AI chips in space for the first time. The plan points to putting data-processing hardware aboard satellites, a step that could eventually move parts of AI computing off Earth. The announcement is drawing attention across the tech and semiconductor world, with observers weighing the engineering challenges of running chips in orbit and what it could mean for the future of data centers.

Google is preparing to test its AI chips in space, according to a report carried by The Brussels Times. The move suggests the company is exploring how its custom AI hardware performs outside Earth, though details on the mission, timeline and partners have not been disclosed. Observers of the semiconductor and space industries are watching for what this could mean for satellite computing and data processing in orbit.

Korean chip giants Samsung Electronics and SK Hynix are moving into co-packaged optics (CPO) technology, which links AI semiconductors using optical connections to boost data transfer speeds and energy efficiency. As AI computing demand surges, Korean media describe the shift as a new race in which chips compete 'at the speed of light', alongside global players developing similar interconnect solutions.

Huawei has unveiled its Atlas 960E AI computing system with 4,096 chips, not the 15,488 it previously announced. The discrepancy has raised questions about the company's stated specifications for the system, which is part of Huawei's push to build domestic AI computing power amid US export restrictions on advanced semiconductors.

CoreWeave, the GPU-focused cloud provider, is shifting from capitalising on shortages of AI chips to proving it can build a lasting, sustainable cloud business. Commentary centres on whether the company can convert its early advantage in scarce Nvidia GPUs into durable revenue as competition intensifies and chip supply improves. Analysts are watching its contracts, margins and ability to diversify beyond raw computing power.

Jensen Huang has reportedly committed to doubling chip supply, a move that would lock up roughly 37% of the world's high-bandwidth memory (HBM) output for 2026. The pledge underscores Nvidia's aggressive push to secure memory for its AI accelerators amid shortages, leaving rival chipmakers competing for the remaining HBM capacity from suppliers like SK Hynix, Samsung and Micron.

Reports suggest Apple may be considering a return to the server market, a business it exited years ago. The move would put the company in direct competition with established server and data-center players amid growing demand for AI infrastructure. Observers are watching whether Apple's chip ambitions could translate into server hardware, reshuffling the competitive landscape among major technology vendors.

NVIDIA's GTC conference is in the spotlight. The event is the company's flagship gathering focused on AI, chips and computing, where NVIDIA typically unveils new GPU architectures, software tools and partnerships. Attendees and industry watchers are following announcements closely, as the conference often signals the direction of the semiconductor and artificial intelligence markets for the months ahead.

TSMC has reportedly begun pilot production of chips on its A14 node, a 1.4nm-class process, with volume production targeted for 2028. The A14 line would succeed the current 2nm generation and keep the Taiwanese foundry ahead of rivals Samsung and Intel in the race for ever-smaller transistor sizes, with major customers such as Apple expected among early adopters.

Barron's reports that SpaceX is purchasing a significant number of Nvidia AI chips, tying Elon Musk's space company to the booming demand for artificial intelligence computing hardware. The report follows widespread interest in how much AI compute Musk's ventures are securing, as Nvidia's GPUs remain the most sought-after chips in the industry. Investors are watching Nvidia's stock closely as major buyers like SpaceX add to already heavy demand.