16 articles
Microsoft closed gaming studios and initiated an Xbox spin-off review on the same day it launched an AI-focused Surface device for developers. With stock down 16.7% year-to-date, global investors are pressing the company to exit non-AI segments. Enterprise buyers worldwide now face a new procurement category: purpose-built AI developer hardware.
Micron's market cap crossed $1 trillion in 2026, with shares tripling year-to-date, as HBM shipments began for Nvidia's Vera Rubin platform. The global memory supply chain — controlled by US-based Micron, South Korea's SK Hynix, and Samsung — has structurally decoupled from consumer electronics cycles. An AI memory supercycle is now in confirmation, not anticipation.
OpenAI has filed confidentially for a $1 trillion IPO — the largest AI company ever to seek public market validation. The filing gives international investors their first public pricing benchmark for frontier AI, reshaping how private AI companies from London to Tokyo are valued. A successful debut would accelerate AI IPOs globally in the second half of 2026.
Palantir and CrowdStrike are running NVIDIA Nemotron-powered AI agents continuously in live production systems — a shift from one-off AI queries to autonomous, persistent decision-making. The deployment spans financial services and cybersecurity clients across North America, Europe, and government sectors worldwide. Enterprise AI is no longer in pilot phase; it is generating revenue in regulated industries.
Dell Technologies recorded $34.1 billion in AI server orders in Q4 FY2026 and is guiding to $13 billion in AI server revenue for Q1 FY2027—a year-over-year doubling. The backlog reflects a worldwide surge in enterprise AI infrastructure spending, with no plateau yet visible. Rivals HPE and Supermicro are contesting the same global market, but Dell's backlog scale signals a commanding early lead.
Meta reassigned 7,000 employees into dedicated AI divisions on May 19, 2026, creating four new AI organizations. The company's 2026 capital expenditure guidance of $115B–$135B exceeds the annual GDP of more than 60 countries. Hardware suppliers across the US, Taiwan, and Southeast Asia stand to absorb the bulk of that spend.
ON Semiconductor guided AI data center revenue to double year-over-year in 2026, among the most aggressive growth targets from any major chip supplier this cycle. The company's stock surged 60.1% over three months after its Q1 2026 earnings beat. The guidance signals that AI infrastructure investment — spanning US, European, and Asian hyperscalers — has moved into a sustained multi-year buildout phase.
Bitdeer posted a $159.53M net loss in Q1 2026 while revenue surged 170% to $188.93M, as the company converts power infrastructure across Norway, Ohio, and Texas into AI compute capacity. The loss reverses $105.32M in net income from Q1 2025. It follows a global pattern of crypto miners repositioning cheap industrial power for AI demand.
America's four largest AI hyperscalers are deploying a combined $650B in capital expenditure, with multiple Broadcom custom accelerator clients nearing simultaneous production launch. The buildout spans data centers, power systems, and semiconductors manufactured largely in Taiwan and South Korea. Broadcom and Nvidia are positioned as primary beneficiaries, with earnings revisions expected across chip, power, and networking suppliers over the next two to four quarters.
Corning has gained 137.4% year to date after expanding its long-term supply agreement with NVIDIA, committing to three new manufacturing facilities. The deal secures fiber optic and specialty glass production for AI data center infrastructure worldwide. It signals that AI investment is now flowing deep into physical supply chains across borders.
Three companies — South Korea's SK hynix and Samsung, and US-based Micron — control virtually all high-bandwidth memory supply while global AI demand is forecast to grow 35x by 2028. DRAM contract prices are projected to rise 58–63% in Q2 2026 alone, with Gartner forecasting a 125% full-year increase. A structural manufacturing constraint, not market timing, is driving the shortage.
DRAM prices are forecast to rise 125% across 2026, with a 58–63% surge expected in Q2 alone. The shock will hit AI cloud providers globally as memory — not compute — becomes the binding constraint on infrastructure expansion. South Korean and US chipmakers stand to gain while AI startups face compressed runways.
US hyperscalers have pledged $660B in combined AI capital expenditure for 2026, driving an 18-day winning streak on the PHLX Semiconductor Index. Meta signed a $100B+ five-year chip deal with AMD, signalling supply chain diversification beyond NVIDIA. Bloomberg Intelligence forecasts 57% semiconductor revenue growth globally in 2026 — double the broader market rate.
Google will invest up to $40 billion in Anthropic, the largest hyperscaler commitment to a single AI lab in the world. The deal gives Anthropic structural advantages — compute access, enterprise distribution, API pricing — that independently funded rivals globally cannot easily match. For AI labs from London to Singapore, the question shifts from whether to seek hyperscaler backing to which one and at what terms.
Palantir secured a $300 million U.S. Department of Agriculture contract to modernize farm programs, pushing its stock up 51.4% over the past year. The win extends a government portfolio spanning the FAA and Pentagon, built on a bureaucratic moat — clearances, certifications, embedded agency relationships — that competitors in the U.S. and abroad find difficult to replicate. As governments worldwide accelerate AI adoption, Palantir's singular focus on public-sector data infrastructure gives it a
Orkes closed a Series B in April 2026, with all four investors from its 2024 Series A returning — including backers spanning the US, India, and Saudi Arabia. The company builds workflow orchestration infrastructure for enterprise AI automation. Repeat participation signals strong institutional conviction in AI pipeline tooling as a durable infrastructure layer.