According to Omdia’s Research Director for Digital Infrastructure Vlad Galabov, who claimed substantial productivity gains across sectors driven by AI will fuel this progress, more than 50 % of global data center power and more than 70 % of profit option. Galabov made a number of other business estimates during Data Center World 2025’s scientist time:
- NVIDIA and hyperscalers ‘ 1 MW-per-rack goals are unlikely to occur for another few years until architectural innovation catches up to energy and cooling needs.
- Over 35 GW of data middle power is anticipated to become self-generated by 2030, making off-grid and behind-the-meter solutions no more recommended for those looking to build fresh data centers because many utilities struggle to provide the needed power.
- By 2030, data center annual capital expenditure ( CAPEX ) investments are projected to be worth$ 1 trillion, up from$ 500 billion at the end of 2024.
- Natural system, such as electricity and cooling, has the highest level of CAPEX spending, increasing annually by 18 %, according to spending trends.
The investment in physical infrastructure increases as compute concentrations and cabinet concentrations rise, according to Galabov. We anticipate a reduction in the site matter when a small number of scaled-up systems are preferred over a scaled-out client strategy. The price per byte/compute pattern is also decreasing.
Energy in a data center explodes
Galabov cited the authority outages that AI has caused for info centers. Less than 150 GW of installed strength was present in data centers around the world when the AI storm first started in late 2023. He predicts nearly 400 GW of accumulated data center power by 2030, though 600 k cabinet styles are only about two years away and 120 kW cabinet designs are just beginning to appear. By the end of the decade, fresh data center power additions expected to be close to 50 GW per year will be common.
However, not everyone will be able to survive the wild west of the DC and Artificial markets. Many new DC school developments and neoclouds may fail to develop a long-term business model because some lack the necessary business acumen to succeed. Galabov warned against focusing on just one company because some are very likely to fail.
More policy of the Data Center World 2025: NVIDIA’s Vision For AI Factories
Artificial drives innovation in wet cooling
Shen Wang, a principal analyst for Omdia, described the cooling effects of the AI storm. He claimed that heat heating reached its maximum in 2022. It may provide up to 80 W per cm2, with a few manufacturers claiming to raise the temperature of air refrigeration.
Beyond that range, single-phase direct-to-chip ( DtC ) cooling, where heat is removed by placing water or other fluid on cold plates that are placed directly on top of computer chips, is required. Single-phase DtC may reach 140 W/cm2.
The best way to great cards right then is single-phase DtC, according to Wang. The most recent racks did have” by 2026, the boundary for single-phase DtC may be exceeded.” When two-phase wet cooling should start to experience a rise in implementation prices. Fluids are heated up in two-phase heating, which causes the device to turn to mist as part of the cooling process, increasing cooling efficiency.
The heaviest implementation of liquid cooling is seen for sophisticated chips in the 600 watt and higher range, according to Wang. ” By 2028, 50 % of cards in that group will use liquid cooling,” according to the report.
