NextDC (ASX: NXT), an Australian data center operator, said on August 27 that contracted capacity had tripled to 740.1 megawatts in its strongest contracting performance to date, while it returned to an A$82.1 million statutory profit for the year ended June 30, 2026. Contracted utilization rose 202% and capital expenditure reached A$3.4 billion as the company moved to meet demand linked to the AI infrastructure boom. The profit recovery was partly driven by a A$128.8 million fair-value gain after several customer contracts were treated as operating leases and the related assets as investment property. Underlying performance also improved, with net revenue rising 16% to A$405 million, underlying EBITDA increasing 15% to A$248.8 million and billing utilization climbing 58% to 175.0 MW. Chief Executive Officer and Managing Director Craig Scroggie called FY26 the largest contracting year in NEXTDC’s history and highlighted a 565.1 MW Forward Order Book, more than three times current billing utilization. NextDC said the commitments are binding contracts rather than options, reservations or pipeline opportunities. Capex totaled A$3,397 million, A$397 million above the top guide, which had already been raised three times. Since August 2025, the company has secured A$9.75 billion in new capital, increasing pro forma liquidity from A$5.5 billion to A$8.7 billion. KL1 in Kuala Lumpur opened to a foundation customer in May 2026, and construction has begun on TK1 Tokyo. The expansion comes as Gartner forecasts global data center electricity consumption will rise 26%, from 447 terawatt-hours in 2025 to 565 terawatt-hours in 2026, with AI-optimized servers accounting for 31% of consumption. McKinsey estimates global data center energy requirements could rise from about 82 GW in 2025 to around 220 GW by 2030, with AI-related demand increasing from 44 GW to 155 GW, or about 70% of total demand. In Australia, grid connection reviews, long waits for electricity access, equipment shortages and public opposition are emerging as constraints on converting booked AI demand into operating capacity.