Paris advocates assert that, despite President Donald Trump’s decision to pull out of the agreement, we face an inevitable transition to a multitude of new low-carbon energy technologies that will consign hydrocarbons—oil, natural gas, and coal—to history’s dust heap. All we need to birth the new energy future are the right “incentives,” which Paris provides: The only way to meet aspirational carbon emissions reductions is to replace hydrocarbons.

This so-called multitude of options actually distills to just three things: biofuels, windmills, and solar panels. Some environmentalists include nuclear fission as well. There’s no new physics here: The first fission reaction was in 1939, the first photovoltaic cell was created in 1954, windmills date to the Middle Ages, and making alcohol (the dominant biofuel) predates recorded history.

But, say Paris supporters, all we need are moonshot government programs to see energy tech emulate the “Moore’s Law” progress that happened in computing (the doubling in computer power every two years). Yet to their inconvenience, the physics of energy won’t cooperate.

Fueling humanity is not like putting a few people on the moon. It’s like putting everybody on Earth on the moon—permanently. No amount of money obviates the costs of fighting gravity. But pundits in awe of Silicon Valley’s prowess analogize today’s green energy companies to the disruption of landline phone businesses after the advent of cell phones or the disruption of taxis by Uber.

Unfortunately, the physics of energy production scales in exactly the opposite direction of the physics of information production. If aircraft, for example, followed Moore’s Law, a loaded Boeing 777 could reach Japan from New York by burning a housefly’s weight in fuel. Engineers and mathematicians can trick bytes to go faster and slip into ever smaller spaces, but only in science fiction are similar tricks possible with physical objects or humans.

Better technology has brought solar and wind costs down dramatically in the past decade. And, fueled by over $100 billion in subsidies, while wind and solar have increased 10-fold over a decade, they still supply, respectively, just 0.3% and 2% of America’s energy.

Remember biofuels? Ten years ago Silicon Valley legend Vinod Khosla famously predicted that with biofuels there is “no doubt that 100% of our gasoline use can be displaced in the next 25 years.” Now, despite 40% of America’s corn crop used distilled into alcohol, that supplies barely 5% of U.S. transportation. No one expects America or the world to get anywhere close to Khosla’s forecast.

But there is a stubborn idea that more money—subsidies, tax incentives, grants, or enforced standards—will make solar and wind 10 times cheaper yet again. While all technologies get better over time, they also approach physics’ limits eventually. Aircraft turbine manufacturers brag about single-digit percentage gains in fuel efficiency precisely because those machines are near thermodynamic limits in converting heat to thrust.

Solar cells and wind turbines are now in the same boat. There are no game-changing advances left in the core technologies; both are now on the curve of diminishing returns. (Biofuels crossed that Rubicon centuries ago.) Nor are there big gains in economies of scale for the underlying components—concrete, steel, fiberglass, silicon, wires, and glass are all already in mass production.

If we want a different energy revolution, we’ll need new discoveries in the physical sciences. That can only emerge from basic research, not from more subsidies for yesterday’s technologies.

Meanwhile, the world’s nearly 8 billion people and $80 trillion economy depend on hydrocarbons to supply over 80% of global energy; oil fuels 98% of transportation. Meaningful changes to this status quo won’t come from the Paris agreement with or without the U.S., unless there’s a revolution in physics. And the latter will require very different priorities. Perhaps it’s time to chase real science.

Mark Mills is a senior fellow at the Manhattan Institute.