# AI%20infrastructure
Latest news and articles about AI%20infrastructure
Total: 19 articles found
Google’s Blowout Quarter Reinforces AI-Driven Demand for Optical Infrastructure — A Boost for China’s Communications Suppliers
Alphabet’s strong quarterly results — faster search growth, 48% cloud revenue expansion and improved cloud margins — confirm that generative AI is moving into scalable, profitable products. That momentum is driving massive investment in optical interconnects and datacentre infrastructure, benefiting specialised communications suppliers and reflected in strong returns for funds concentrated in optical‑equipment makers. The shift raises strategic supply‑chain and policy questions as co‑packaged optics and silicon photonics become central to AI scale‑up.
China’s Cloud Firms Brace for an AI-Driven Price Shock as Competition Moves Up the Stack
China’s major cloud providers are integrating open‑source AI assistants like OpenClaw while confronting rising upstream costs and surging enterprise demand. Expect selective price increases for AI GPU services, a tighter focus on packaged AI applications, and competition shifting from raw compute to full‑stack offerings.
Musk Merges SpaceX and xAI, Betting on an Orbital AI Datacentre and a $1.25tn Giant
Elon Musk has merged xAI into SpaceX in a stock-swap that values the combined entity at roughly $1.25 trillion and lays out a plan to move large-scale AI compute into orbit using Starship, Starlink and lunar resources. The integration aims to create virtually unlimited solar-powered compute capacity but faces steep technical, regulatory and geopolitical hurdles.
Nvidia’s Jensen Huang Confirms Company Will Invest in OpenAI — but Not the $100bn Hype
Nvidia CEO Jensen Huang said the company will participate in the current OpenAI funding round and that the investment could be Nvidia’s largest ever, though it will be far below earlier $100 billion speculation. The pledge strengthens ties between a leading GPU supplier and a top AI-model developer, with strategic benefits and regulatory complexities for both firms and their customers.
AI Data‑centre Build‑outs Drive Chinese Optical‑Module Profits to Historic Peaks — But Growth Is Uneven
China's leading optical‑module makers have issued 2025 profit forecasts that reflect a surge in demand from AI data‑centre deployments. Xinyisheng and Zhongji Xuchuang expect historically large, core‑business‑driven profit gains, while Tianfu Communication posts solid but smaller growth, underscoring an uneven competitive landscape.
Musk Eyes ‘Space Compute’ and Robot Missions to Moon and Mars as He Weighs Corporate Consolidation
Elon Musk is reportedly advancing plans to shift substantial computing into space, consider mergers among his companies, and explore delivering Tesla’s Optimus humanoid robots to the Moon and Mars. The ambition combines SpaceX’s launch capability with Tesla’s robotics and Musk’s AI interests, but faces steep technical, economic and regulatory hurdles.
Alibaba Stakes a Claim in Nuclear Power to Secure an AI Advantage
Alibaba has taken a stake in a large nuclear power project as part of a broader push by Chinese tech firms to secure the baseload electricity needed for large AI deployments. China’s industrial capacity in power equipment and fast delivery gives its companies an edge in the global struggle for compute, reframing energy as a core element of AI competitiveness.
Shanghai Stakes a Claim to 120 ExaFLOPS of Smart Compute and Pitches Space–Ground 'Supernode' Architecture
Shanghai announced that its aggregate smart‑compute capacity has surpassed 120,000 petaFLOPS and outlined an agenda to build a self‑reliant compute ecosystem linking domestic chips, models and cloud services. Officials and industry groups also floated an integrated space–ground architecture that would treat LEO satellites as programmable compute clusters to achieve global, low‑latency coverage for selected applications.
AI’s Hunger for Memory Could Keep Global Chip Shortages Dragging On Until 2027
Synopsys CEO Sassine Ghazi warns that the current memory-chip shortage, driven by heavy demand from AI data centres, is likely to last through 2026 and potentially into 2027. Concentrated production, long lead times for new fabs and booming demand for HBM mean elevated prices and allocation pressures may persist, benefiting memory suppliers but squeezing device makers and other industries.
Samsung Doubles NAND Prices as AI‑Fueled Storage Supercycle Tightens Supply
Samsung has raised NAND flash prices by over 100% in Q1 2026 as AI‑driven demand for high‑performance storage outstrips supply. Analysts say the industry has entered a storage‑chip "supercycle," with tight capacity likely to persist until at least 2027 and meaningful new supply not expected until 2028.
From Shanghai Markets to Launchpads: Jensen Huang’s China Stopover and a Boost for Beijing’s Commercial Space Push
Jensen Huang’s early-2026 visit to Shanghai and public-facing interactions underline ongoing commercial links between Nvidia and China despite geopolitical headwinds. At the same time, Beijing is accelerating support for a domestic commercial space sector — highlighted by policy measures for satellite-data use and Zhongke Yuhang’s completion of IPO counselling — while tightening data and security regulation, creating both opportunities and risks for firms operating in China.
vLLM Founders’ New Startup Raises $150m Seed at an $800m Valuation — A Big Bet on LLM Infrastructure
Inferact, founded by the vLLM core team, raised $150 million in a seed round led by Andreessen Horowitz and Lightspeed at an $800 million valuation. The deal highlights investor enthusiasm for LLM inference and deployment infrastructure, but sets high expectations for rapid commercialisation amid fierce competition and regulatory questions.