Solar Is Poised To Lead Power Supply By 2035, But AI Data Hubs Will Prolong Fossil Fuel Use

Solar is on track to be the single biggest source of electricity within the next decade, a new industry analysis finds, yet booming demand from artificial intelligence data centers and the electrification of heavy industry will keep gas and coal running for years to come. The consultancy argues the move toward sun-derived generation is being driven largely by economics: panel-mounted systems are now so inexpensive that utilities and developers choose them over thermal plants in many markets. At the same time, operators of massive computing facilities are expected to add vast amounts of capacity, forcing a continued role for dispatchable fossil-fuel plants to meet round-the-clock loads.

The report projects that data center growth could prompt roughly 1,000 gigawatts of new large-scale solar capacity, alongside several hundred gigawatts of new gas-fired plants and a smaller tranche of coal. Because gas and coal can generate power continuously, the firm estimates that a little more than one-half of the extra electricity serving those computing sites through 2050 will still come from fossil sources. In short, decisions by cloud operators and data center builders will materially affect which generation technologies remain economically viable in the coming decades.

That outlook is not uncontested. Longer-duration storage, underground heat systems and advanced fission designs are all jockeying for a share of the market, and a handful of recent commercial moves have intensified interest. A major cloud provider has purchased multi-day battery systems for a recent facility, and a series of public listings of firms working on deep thermal and small modular reactors has drawn investor attention. Energy analysts say those technologies can blunt fossil dependence if they scale fast enough, but solar’s cost advantage makes the path uphill.

Falling panel prices are central to the forecast. The consultancy expects module costs to drop by about one-third by 2035, a trajectory that would allow solar to outcompete most coal and gas projects on price alone. By mid-century, the analysis sees panels supplying more than twice the electricity produced by gas plants today. Industry veterans point to two main drivers: active industrial policy and subsidies that boosted manufacturing capacity in Asia, and the rapid gains from producing equipment at massive scale. As one independent analyst put it, learning-by-doing has pushed unit costs down faster than many anticipated.

The glut of daytime solar is already reshaping markets. In parts of southern Europe, a surplus of midday solar has depressed wholesale prices so sharply that standalone arrays struggle to turn a profit. Developers are responding by building hybrid sites that combine panels with batteries, storing excess output to sell when prices rebound in the evening. The battery sector has entered a growth phase: global grid-scale storage deployments last year were on the order of 112 gigawatts, and the consultancy expects that fleet to expand to about three times that size by 2035. Manufacturers from materials recyclers to automakers have launched storage businesses to capture demand from utilities and large customers.

The report did not fully incorporate the effects of the recent Middle East conflict, which came after the research was well underway. The team did run two stylized scenarios to probe import dependence: one where market forces — mainly cost declines — drive the shift, and another where aggressive regulation pushes economies toward net-zero. In the market-driven case, virtually every nation cuts its reliance on foreign fuels, including major exporters. Under the regulatory pathway, countries could reduce import needs to almost nil. Analysts say the policy choices of governments, together with procurement decisions by tech firms, will determine how fast fossil fuels are phased out.