CMC Research has released a specialist study examining water-resource strategies for AI data centers, positioning water alongside power as a material consideration in site selection, cooling design, operational resilience and investment planning. The publication covers the market outlook through 2035.
Representatives from Abu Dhabi and South Korea signed seven memoranda of understanding at the Abu Dhabi Investment Forum Seoul 2026 on 16 September. Abu Dhabi also outlined plans to develop large data centers to support AI, while seeking Korean investment and technology partners for joint projects.
Thailand’s Board of Investment has approved three data-center projects with combined investment of THB91.3 billion, accounting for nearly all of a THB95.8 billion package of six large projects. The largest approval is a THB84.2 billion TikTok data-infrastructure expansion spanning Bangkok, Samut Prakan and Chachoengsao.
Morgan Stanley has maintained Overweight ratings on Murata, Samsung Electro-Mechanics (SEMCO) and Yageo, citing early-stage adoption of multilayer ceramic capacitors (MLCCs) in AI servers and server racks. The bank forecasts global MLCC shipment value will rise from US$14.67 billion in 2025 to US$44.45 billion in 2031.
Micron has demonstrated a 512GB DDR5 RDIMM across multiple server platforms, with claimed data rates of up to 9,200 MT/s. The company expects volume production to begin in the second half of 2027, subject to customer demand, while AMD and Intel validate the module for next-generation server platforms.
Eindhoven-based AI chip startup Euclyd has announced a $231 million funding round involving Samsung and several European investment organizations. The company plans to use the capital to advance chips and systems for AI inference, including rack-level hardware for enterprise customers.
Hokkaido has issued a statement to data-center operators calling for the active use of decarbonisation and energy-saving technologies, including those using domestic and local resources, alongside measures to coexist with host communities. The move accompanies the region’s push to attract infrastructure supporting AI, semiconductor and digital industries, rather than a specific data-center project announcement.
Crusoe has closed a $3.9 billion Series F financing round at an approximate $30.9 billion valuation. The company plans to expand its Spark small modular data-center platform, manufactured in its own facilities and intended for deployment where power is available.
Emerald AI has raised $150 million in a Series A round, according to the available source. The company is positioned as a provider of solutions to optimise electricity demand at AI data centres.
NVIDIA, Google and Emerald AI announced the AI Energy Management Alliance (AEMA) on September 16, 2026, aiming to advance data centers that can reduce or shift electricity demand in response to grid conditions. The alliance launched with 18 partners, including Anthropic, National Grid and AES.
Japan’s government has established an interagency liaison council on data-center siting and begun considering related policy measures. The initiative addresses friction with local communities and the power-supply requirements of rapidly expanding data-center development.
Moody’s estimates that meeting incremental US data-center electricity demand could require about $110 billion of investment in new generating capacity. The assessment highlights a widening gap between the build-out of AI data centers and the pace of generation and grid development.
ASUS JAPAN plans to exhibit its XA NR1I-E12R, a fully liquid-cooled 2U AI server based on NVIDIA HGX Rubin NVL8, at CTC DISCOVER 2026 in Japan. The company says availability of the model is planned from autumn 2026.
Naver has signed a memorandum of understanding with Gurin Energy to explore an equity investment in the Jindo Green Solar project and a long-term power purchase agreement for its data-center operations. The proposed renewable supply would be directed primarily to AI and cloud infrastructure at Gak Sejong and Gak Chuncheon.
NVIDIA and Google are intensifying their competition across Taiwan’s AI infrastructure supply chain, extending beyond accelerators into server platforms, networking, liquid cooling, power systems and software. At SEMICON Taiwan in Taipei, Google AI infrastructure CTO and senior vice president Amin Vahdat made his first visit to Taiwan and delivered an opening address.
Taiwan’s Industrial Technology Research Institute (ITRI) used its Innovation Day in Shalun, Tainan, to present a high-density chip-cooling design for AI data centers and its operating hydrogen energy demonstration platform. The event highlighted technologies aimed at addressing the thermal and power constraints associated with semiconductor, data-center and AI expansion.
AWS’s cloud infrastructure in Bahrain continues to provide impaired access to some data and compute resources following war-related damage, according to the source. In the UAE, AWS availability zone mec1-az2 has been deemed unrecoverable, while mec1-az1 and mec1-az3 remain under repair.
Meta Platforms plans to expand the use of internally developed AI processors in its data centers, targeting lower model energy consumption and operating costs. The company says it intends to deploy more than 1 GW of such chips over a 12-month period, while MTIA 450 (Arke) remains in testing and MTIA 500 (Astrid) is targeted for data-center deployment by the end of 2027.
Moody’s estimates that meeting rising US data-center electricity demand will require about $110 billion of investment in new generating capacity. The assessment highlights a widening mismatch between the pace of AI data-center development and the build-out of generation and transmission infrastructure.
NVIDIA, SpaceX and Google are advancing announced platforms, concepts and planned demonstrations for orbital computing infrastructure. The initiatives remain pre-commercial and should not be treated as operational space-based data centers.