NVIDIA and IREN Signal the Next Phase of AI Infrastructure Scale
NVIDIA’s latest partnership with neocloud operator IREN is another sign that the AI infrastructure market is becoming increasingly defined by access to power and deployable data centre capacity. Under the agreement, NVIDIA has secured the right to invest up to $2.1Bn into IREN over five years, while the two companies plan to deploy up to 5GW of NVIDIA DSX-aligned AI infrastructure across IREN’s global pipeline.
The announcement comes alongside IREN’s acquisition of Spanish data centre developer Nostrum Group, adding approximately 490MW of secured, grid-connected capacity in Spain. Together, the two announcements highlight a broader shift in the market: AI infrastructure players are increasingly prioritising control over land, grid access, and development pipelines, rather than relying only on traditional colocation procurement.
More broadly, the announcements reinforce a structural shift underway across the sector: AI infrastructure is becoming increasingly vertically integrated. NVIDIA is no longer simply supplying chips into the market – it is positioning itself deeper within the infrastructure stack, aligning with developers, cloud operators, networking providers, and energy ecosystems to secure long-term deployment capacity. As AI workloads continue driving demand for compute, the competitive advantage is increasingly moving toward organisations capable of controlling the full chain of delivery, from semiconductors and networking through to power infrastructure and operational campuses.…Read more
SoftBank’s Battery Push Shows AI Infrastructure Is Becoming an Energy Strategy
SoftBank’s launch of a new battery business in Japan highlights how AI infrastructure strategies are increasingly moving beyond data centres alone and into the wider power ecosystem. The company plans to take an end-to-end approach, covering the development and manufacturing of battery cells and battery energy storage systems, with its AI data centre in Sakai City, Osaka, acting as the core hub for the initiative.
The scale of the ambition is notable. SoftBank plans to begin manufacturing battery cells and energy storage systems by the fiscal year ending 31 March 2028, before reaching gigawatt-hour-per-year scale mass production around the fiscal year ending 31 March 2029. The batteries will initially support SoftBank’s large-scale AI data centres in Japan, before being offered for grid, industrial, factory, residential, and potentially overseas applications.
More broadly, the announcement reflects a structural shift in AI infrastructure: access to power is no longer just a utility consideration, but a strategic capability. As AI data centres drive higher and less flexible electricity demand, battery storage, energy management systems, and AI-based power forecasting are becoming increasingly important parts of the infrastructure stack. SoftBank’s move shows how major technology and investment groups are beginning to treat energy resilience, storage, and grid support as core enablers of AI deployment…Read more
STACK’s Reported APAC Sale Highlights the Premium on Powered Growth Platforms
STACK Infrastructure is reportedly weighing the sale of assets in its APAC business, with Bloomberg reporting that the Blue Owl Capital-backed developer is targeting a valuation of around $30Bn. If achieved, this would represent one of the largest data centre transactions to date, behind only the proposed $40Bn BlackRock, GIP and MGX takeover of Aligned Data Centres.
The potential valuation is notable when compared with recent APAC transactions, including Blackstone and CPP Investments’ $16Bn acquisition of AirTrunk. STACK’s reported price tag appears to reflect not only existing operational assets, but also the strategic value of its development pipeline across high-growth APAC markets, including Japan, Australia, Malaysia and South Korea.
More broadly, the report reinforces how investor appetite for scaled data centre platforms remains strong, particularly where operators can demonstrate contracted capacity, access to land and power, and credible future growth. As demand from cloud and AI workloads continues to expand, the market is increasingly placing a premium on platforms that can offer not just existing capacity, but a clear path to large-scale, power-backed development…Read more