Google plans to invest at least €13 billion in three data centers in northern Finland and has agreed a 22-year nuclear-power purchase arrangement with Fortum. Separately, NextEra Energy has secured up to $1.9 billion in U.S. Department of Energy financing for a potential restart of the 615 MW Duane Arnold nuclear plant, linked in the review to a 25-year Google contract.
Aon is reportedly expanding its Data Center Lifecycle Insurance Program with an Operational Property Data Center Lifecycle Program for facilities moving out of construction coverage and for operating data centers. The offering is reported to target $10 billion in insurance capacity, rather than representing capacity already fully placed or available.
imec presented a three-metal-layer NbTiN/a-Si/NbTiN Josephson-junction circuit at the Applied Superconductivity Conference 2026, alongside a three-layer NbTiN interconnect stack. The research demonstration reached junction densities of up to 3.8 million per cm² and 30 nm metal-line widths.
NVIDIA and MediaTek have announced an expanded strategic partnership spanning AI data-center infrastructure, AI PCs and automotive electronics. NVIDIA subscribed to $3.5 billion of MediaTek’s overseas unsecured convertible bonds, while Alphabet also participated in a transaction reported at $3.9 billion in total.
SK Group Chairman Chey Tae-won is scheduled to inspect the construction site of the SK AI Data Center Ulsan on September 10. The facility is being developed with AWS, with total project investment stated at KRW 7 trillion and commissioning planned by the end of 2027.
Ho Chi Minh City has approved the lease of a roughly 4.46-hectare site in Saigon Hi-Tech Park to Starmason for a planned hyperscale data-center project. The development is planned to be delivered in phases, with total stated investment of US$480.26 million and maximum planned equipment capacity of 90 MW.
UK Prime Minister Andy Burnham said on 9 September that the government does not intend to introduce a nationwide pause on data center construction. The position maintains policy support for AI Growth Zones while linking new facilities to local investment and tax-revenue benefits.
Japan has included Tomiya, Miyagi Prefecture, in the first 18 areas selected under its GX Strategic Regions programme. The city’s approved initiative envisages a gigawatt-class data center at the Narita district industrial park over roughly the next 10 years, alongside development of an associated industrial cluster.
A TeleGeography review of Asian cloud infrastructure, published on 20 August 2026 and reported by a Japanese blog, shows China retaining the region’s largest concentration of cloud regions and availability zones as deployments broaden across India, Japan and Southeast Asia. The data are current to the first quarter of 2026.
Sumitomo Corp. has announced plans to pursue practical deployment in Japan of a high-performance power conductor using carbon-fibre-reinforced polymer technology from its French portfolio company, Epsilon Composite. The company claims the technology could reduce transmission losses by up to 30%, positioning it as a potential response to grid constraints associated with AI, data-center and semiconductor demand.
SGC Energy has signed a non-binding memorandum of understanding with Vertiv to jointly evaluate infrastructure for a planned AI data center in Gunsan, South Korea. Vertiv intends to deploy its integrated PowerNexus power infrastructure solution at the site, which SGC Energy expects to enter operation in 2028.
A Japanese market review from S1 identifies the expansion of AI infrastructure and data centers as an emerging source of demand for silicon-carbide (SiC) and gallium-nitride (GaN) power semiconductors. The review points to their potential use in data-center power supplies, UPS systems, voltage conversion and related electrical infrastructure, rather than reporting a specific deployment or supply contract.
fantasista has approved plans to advance an AI data center development on land owned by its consolidated subsidiary, NC MAX WORLD, in Aoba Ward, Sendai. The planned facility is designed for approximately 200 MW of total grid-connected capacity, with the first phase preliminarily budgeted at about ¥400 billion.
NVIDIA has announced partnerships with eight Australian data-center and AI-infrastructure providers to expand land, power and shell capacity for NVIDIA DSX AI factories. The programme targets up to 2 GW of future buildout by 2027, rather than commissioned capacity.
Fukuoka Prefecture has been formally designated a national Green Transformation (GX) strategy region, supporting plans to develop a major data-center cluster by 2036. The initiative covers land and infrastructure preparation across Kitakyushu, Nogata and Kurate, with a planned electricity supply exceeding 1 GW.
BizClik convened representatives from Sitetracker, Accenture and Salesforce to examine the delivery handoffs that can delay AI-ready data center capacity from funded programme to operational infrastructure. Speakers identified project execution and power availability as central constraints on data center expansion.
SSK will hold a seminar on 23 October 2026 examining how generative AI could affect Japan’s electricity balance and the potential of photonic-electronic integration to reduce energy consumption in AI and data-center infrastructure. The event is an educational programme, not a data-center construction, capacity-expansion or investment announcement.
Penguin Solutions has announced a collaboration with cloud operator Lektra to deploy and optimise distributed EdgeScale AI micro data centers. Each site is designed to remain below 20 MW, with Lektra stating that it has deployed eight sites and intends to expand further.
The US Department of Defense is discussing an approximately $5bn loan to AI cloud startup Fluidstack through the Office of Strategic Capital. The proposed funding would target US supply-chain and manufacturing capacity for data-center components rather than the direct construction of new AI data centers.
Microsoft is reportedly targeting an expansion of its global data-center footprint from about 12 GW today to more than 38 GW by 2032, spanning facilities it owns and leases. The plan is intended to relieve capacity constraints for cloud and AI workloads, while excluding compute capacity rented from neocloud providers such as CoreWeave.