Even as the conversation around data centers and load growth monopolized much of 2025, it was hardly the year’s only energy story.
In his comprehensive, annual deck on decarbonization, the veteran energy analyst Nat Bullard sifts through a year’s worth of data to outline the most significant, pressing and surprising trends in the sector. And for four years straight, Bullard — the co-founder of Halcyon and former chief content officer of BloombergNEF — has appeared on Catalyst with Shayle Kann for a doubleheader conversation to unpack some of them.
This year, Bullard’s research demonstrates that common narratives about the global energy sector don’t always square with reality. Three such trends that took shape in 2025: the resilience of public markets in the face of fierce political rhetoric in the U.S.; the paradox of rising solar costs; and the actual drivers of global electricity demand.
Listen to part two of Nat Bullard’s interview on the Catalyst podcast:
Taken together, and especially moving into 2026, these serve as a reminder that the energy transition is truly global; while the U.S. remains a major player, it isn’t the primary driver behind how decarbonization unfolds
The public market paradox
On the surface, 2025 did not look like a promising year for clean energy in the United States.
Against a backdrop of political antagonism, investment in U.S. energy startups dropped from over $8 billion in 2022 to $2 billion last year. In the same period, developers canceled nearly 2,000 power projects — a cumulative 266 GW of mostly clean new capacity.
However, Bullard’s data shows that at least on the public markets, clean energy’s performance is not necessarily correlated with political sentiment. The S&P Global Clean Energy Transition Index didn’t just eke by in 2025; it surged.
Speaking on Catalyst, Bullard explained that the index rose 40% between January and December 2025. As such, it dramatically outperformed benchmarks like the S&P 500 and Nasdaq 100, which both rose around 20% last year.
“ There’s activities still ongoing in these markets, regardless of what specifically you might hear in political rhetoric,” said Bullard. “There’s a global nature to this.”
Listen to part two of Nat Bullard’s interview on the Catalyst podcast:
Solar cost dissonance
We often hear that solar prices are on a steady decrease. But Bullard’s data shows that the reality isn’t quite so simple.
Yes, hardware costs keep getting lower, but project costs are on the rise. On Catalyst, Bullard acknowledged that solar modules are cheaper than ever, but the total cost to build systems in the U.S. is increasing.
In 2025, he found, commercial solar prices rose 9%, and utility-scale fixed tilt and tracking prices jumped roughly the same amount. (Prices of solar per watt, he notes, are dramatically lower in countries like Australia and Germany.)
Bullard argues that the savings on panels are being erased (and then some) by rising soft costs. “Basically the things that are driving the cost up — engineering, planning and permission, labor, the cost of equipment — thanks to tariffs, all of these things that are outside the remit completely of your friendly local module manufacturer,” he said.
The biggest drivers of global electricity demand
While conventional wisdom suggests that AI and data centers are the primary driver of new electricity demand especially in the U.S., the reality across the globe is much more complicated.
From a global perspective, data centers are actually fifth on the list of anticipated drivers of new electricity demand through 2030, according to data from IEA.
Bullard’s research indicates that industrial electrification is expected to top the list in the next four years, constituting about 30% of growth. Both electrified transport (13%) and household appliances like refrigerators and dishwashers (roughly 12%) currently rank above data centers, which are currently expected to consume about 8% of new demand. Even the addition of more air conditioning in buildings around the world will consume more new power (10%) than data centers.
But Bullard cedes that the forecasts could easily change.
“Consider this very much a moving target,” he said. “I’ll be very interested to see what this print looks like a year from now, two years from now, three years from now. Certainly [data centers] will move up, but, but where they land is a really big question.”
