Solar and batteries are thriving — even in Trump's America | Zero: The Climate Race

ELI5/TLDR

The headline says Trump has killed American clean energy. The deployment numbers say the opposite: roughly 90% of all new power capacity added to the US grid in 2025 and 2026 was solar, wind, and batteries. The reason is boring and powerful — solar plus storage is simply the cheapest electricity you can build, even after Trump’s tariffs triple the price of panels and the tax credits start expiring. The guest, a developer building one of America’s biggest solar-and-battery plants in Arkansas, thinks the future is solar regardless of who is in the White House. The fight isn’t really about cost. It’s about lobbying money and natural gas.

The Full Story

The narrative and the numbers don’t match

Bloomberg’s climate podcast opens with the story everyone has heard: the Trump administration gutted tax credits for wind and solar, killed EV subsidies, and spent taxpayer money to stop offshore wind farms. All true. But then there is the awkward fact that renewables keep winning anyway.

“The first part is true, the second part is not true… the industry has been under attack with policy changes… However, the industry is thriving.”

That’s Kevin Smith, CEO of Cypress Creek Energy, a man who spent two decades in oil, gas and nuclear before switching sides. His company just closed $3.5 billion in financing for the Steel River Energy Center in Arkansas — 2.5 gigawatts of solar and 2.9 gigawatt-hours of storage when finished in 2029, enough to power a small city. By US standards it is enormous. By Chinese or Indian standards it is ordinary, which tells you how far behind America actually is.

Cheapest wins, even with a handicap

The core argument is one line repeated many ways: solar plus batteries is the most affordable power on the grid, full stop. Smith puts numbers on it. A solar project runs about $60 per megawatt-hour with tax credits, maybe $85–90 without them. A brand-new natural gas combined-cycle plant — top-of-the-line turbines — costs $120–140. So even stripped of subsidies and loaded with tariffs, solar still undercuts gas.

The tariffs are not small. Americans pay roughly 32–38 cents per watt for solar modules; the rest of the world pays 12–13 cents. Think of it like a tax of two-thirds, passed straight through to your electricity bill. A solar-plus-battery project costs about $2,500 per kilowatt in the US; without the trade barriers, Smith reckons it could be under $1,500.

There’s a tax-credit cliff worth understanding. Projects must be “under construction” by July 2026 and operating by end of 2030 to qualify. That deadline is pulling a wave of building forward — a short boom. Bloomberg’s own forecasts show solar peaking in 2026, dipping to around 40 GW a year, then recovering. Smith’s counter: 40 GW a year is still about $100 billion of annual construction, a figure that would have looked miraculous a decade ago.

The real battleground is data centers

The US is finally seeing electricity demand grow — 3 to 5% a year — after decades of flatness, driven mostly by AI data centers. The bearish case for solar is that data centers want to be built fast and need 99.999% uptime, and that gas-plus-batteries delivers that reliability better.

Smith’s rebuttal hinges on a distinction: behind-the-meter versus grid-connected. To hit “five nines” of reliability, he argues, the vast majority of data centers — 80% plus — have to plug into the grid anyway. And once you’re on the grid, the utility just buys the cheapest mix available, which means lots of solar and storage. The off-grid gas-and-battery campuses making headlines are, in his telling, a handful.

“If the data center load goes down… it’ll likely come out of the natural gas side of the model, not so much out of the solar and battery storage.”

This is where US and global forecasts split hard. Bloomberg NEF projects US data-center demand becoming gas-heavy from 2027, with gas meeting 80–90% of that new load by 2035 — the US is the only large economy raising its gas-power load over that decade. Smith thinks the forecasters are wrong because gas plants need that high $120–140 price round the clock to cover their capital costs, even at 3am when grid power drops to $20. He’s betting economics overrules everything.

Why gas still has a grip

Two reasons the US bucks the global trend. One, it actively penalizes renewables with tariffs while the rest of the world removes barriers. Two, America sits on cheap, abundant natural gas — though Smith notes that as the US builds more LNG export terminals, gas stops being “trapped” and becomes a world-priced commodity, which will push its price up and tilt the math back toward solar.

Red states quietly love it

The political punchline: 74% of solar installed in early 2026 went into Texas, Florida, Ohio, Indiana, Michigan, Arizona, Mississippi — all Trump states. More than 80% of renewable-manufacturing investment, tens of thousands of jobs, also landed in red states. The panels for the Arkansas project are 100% US-made (First Solar), the batteries from LG’s US plants, the steel from Mississippi County, Arkansas.

So why does oil and gas keep getting its way? Lobbying. Smith estimates the fossil industry outspends renewables roughly 100 to one in Washington. “We’ve got the right message. We don’t have the right lobbying dollars behind it.”

The wind asterisk

Wind is the cautionary tale. Trump’s executive order halting wind permitting was struck down in court, but a new front opened — the Pentagon is sitting on wind-project reviews, and renewable groups are now suing. Asked whether the anti-wind crusade comes for solar next, Smith doesn’t think so: the economics are too obvious, and choking solar would visibly raise everyone’s bills. He calls the administration’s wind hostility “non-economic, non-technical” — which is a polite way of saying it makes no sense.

Key Takeaways

~90% of new US grid capacity in 2025–26 was solar, wind and batteries; solar+storage alone was ~80%. Deployment, not rhetoric, is the real signal.
Levelized cost stack: solar ~$60/MWh with credits, ~$85–90 without; new gas combined-cycle $120–140/MWh. Solar wins even after losing subsidies.
US solar modules cost 32–38 c/W vs 12–13 c/W globally — Trump tariffs roughly triple the price, all passed through to consumer bills.
Tax-credit cliff: must be under construction by July 2026, operating by end-2030. This pulls building forward (a boom) then normalizes to ~40 GW/year — still ~$100B of annual construction.
The behind-the-meter vs grid-connected distinction is the crux of the gas-vs-solar fight: hitting 99.999% uptime pushes ~80%+ of data centers onto the grid, where utilities buy the cheapest mix (solar/storage).
Gas plants must earn $120–140/MWh around the clock to cover capital — even at 3am when grid prices fall to $20–30 — a structural cost disadvantage vs solar in a low-demand-hour grid.
US is the only major economy projected to increase gas-power load 2025–2035 (per BNEF), due to tariffs penalizing renewables + uniquely cheap domestic gas. LNG exports will erode that gas-price advantage.
High solar penetration creates “negative electricity prices” (Europe now sees hundreds of such hours), which slows new solar but pulls in batteries and flexible demand like crypto mining.
Renewables are low-risk/lower-return vs oil & gas’s high-risk/high-reward; returns have risen to double digits post-Ukraine/Iran, attracting pension and utility capital.
Residential rooftop solar is shrinking (~4 GW in 2026, half of 2023’s level) after federal subsidy removal — but utility-scale (millions of panels on farmland) is a different market and still growing.
Politics is shifting underneath the rhetoric: 74% of early-2026 solar and 80%+ of renewable manufacturing jobs are in red states.
Fossil lobbying outspends renewables ~100:1 in DC — the industry’s main structural disadvantage isn’t technology or cost, it’s political spending.

Claude’s Take

This is a clean, useful interview that punctures a lazy narrative, but you should weigh the source. Smith runs a solar developer and is, by his own admission, “very bullish on solar.” Every answer bends toward “the economics force solar to win.” That’s a real argument, and the cost figures support it — but he waves away the gas-heavy data-center forecasts from BNEF, the IEA, and Wood Mackenzie a little too smoothly. His one rebuttal (gas needs round-the-clock high prices) is genuinely good, yet “every serious forecaster disagrees with me but I’m right because solar has beaten projections before” is a faith argument dressed as an economic one.

The strongest, least self-interested point is the political one: solar and its factory jobs are concentrated in Trump-voting states, which makes a sustained federal war on it politically self-defeating in a way the war on wind isn’t. The behind-the-meter vs grid-connected framing is also the single most clarifying idea here — it reframes the whole “can renewables power AI” debate, because if data centers must grid-connect for reliability, the gas-vs-solar question gets decided by the utility’s merit order, where cheap wins.

Score 7. Substantive, numerate, and the host pushes back with real forecast data rather than lobbing softballs — that tension is what makes it worth the time. It loses points for being a single-guest, single-perspective piece on a question where the guest has money riding on the answer. Treat the cost numbers as solid and the “solar inevitably wins” conclusion as a motivated forecast.