Einstein's Biggest Blunder Was Actually Right

Everything you have ever seen, touched, built, or loved is made of matter and energy that accounts for roughly 5% of the universe. The other 95%? We have no idea what it is.

Sit with that for a second.

A new NOVA PBS documentary, Decoding the Universe: Cosmos, traces the long, strange arc of cosmology from Copernicus to the 1998 discovery that the universe isn't just expanding — it's accelerating. That acceleration is powered by something called dark energy, a force so mysterious that Johns Hopkins astrophysicist Adam Riess, one of the people who discovered it, describes it plainly: "Dark energy is really the name we give to our ignorance of what's causing the accelerating expansion of the universe."

The name we give to our ignorance. That's not a throwaway line. That's the honest state of modern cosmology.

A Fudge Factor That Became a Law of Nature

The story starts in 1917, when Albert Einstein applied his brand-new theory of general relativity to the entire universe. He already knew what answer he wanted: a static, unchanging cosmos, which was the scientific consensus of his era. But gravity kept ruining his math. If everything in the universe attracts everything else, what stops it all from collapsing inward?

Einstein found a loophole in his own equations. General relativity permitted a term he called the cosmological constant — essentially, the idea that empty space itself could exert a repulsive pressure pushing outward, counterbalancing the inward pull of matter. Two gravitational forces in a permanent stalemate. Problem solved.

Then Edwin Hubble demolished the premise.

Working at Mount Wilson Observatory outside Los Angeles in the 1920s, Hubble first established that the Andromeda Nebula sat far beyond the edges of the Milky Way — meaning our galaxy was not, as many had assumed, the entire universe. He published his key findings in late 1924 and 1925, though it's worth noting the history here is somewhat tangled: Belgian cosmologist and Catholic priest Georges Lemaître was independently developing similar ideas around the same time, and credit in early 20th-century cosmology is rarely as clean as textbook accounts suggest.

What's unambiguous is what the redshift data showed. Almost every galaxy Hubble and his contemporaries measured was moving away from us. The light from distant galaxies was stretched toward the red end of the spectrum — the cosmic equivalent of a siren dopplering away. And crucially, the farther a galaxy was, the faster it was receding. As theoretical physicist and author Clifford Johnson puts it in the documentary, "the conclusion is that the whole universe is expanding. Everything is moving away from everything else."

If the universe was expanding, Einstein's cosmological constant — invented to prevent exactly that — was unnecessary. According to physicist George Gamow's autobiography (a source historians treat with some caution), Einstein called it the biggest blunder of his life. Whether he said those exact words is disputed, but the sentiment tracks: he had inserted a term into his own elegant equations just to preserve a conclusion he preferred over what the math actually allowed.

Pigeons, Fossils, and the Bang

Once astronomers accepted an expanding universe, the obvious question was: what happens if you run the film backward? Everything moving apart now means everything was closer together before. Keep rewinding and you arrive at an almost unimaginable starting density — what Lemaître called the "primeval atom" and what critics mockingly labeled the Big Bang.

The mockery didn't land well by 1965. That year, physicists Arno Penzias and Robert Wilson were working with a horn antenna in Holmdel, New Jersey, trying to do routine radio astronomy. They kept picking up a faint, uniform hiss from every direction in the sky. They suspected equipment malfunction. They cleared out pigeons nesting in the antenna. It was not pigeon droppings.

It was the cosmic microwave background (CMB) — a faint afterglow of radiation from roughly 380,000 years after the Big Bang, when the universe had cooled enough for light to travel freely for the first time. The Big Bang theory had predicted this radiation should exist. Penzias and Wilson had just found it by accident while trying to get rid of a static noise they assumed was a nuisance.

The CMB is now one of the most-studied signals in science. Its tiny temperature variations — mapped in exquisite detail by satellites like WMAP and Planck — encode information about the early universe's structure. By the 1970s, it had convinced most astronomers. The Big Bang wasn't fringe anymore.

The Racing Universe

By the 1990s, cosmologists had pivoted to a new question: how fast was the expansion slowing? Because of course it was slowing. Gravity was doing its job. The question was whether there was enough matter to eventually halt the expansion and pull everything back into a Big Crunch, or whether the universe would coast outward forever. Answer that, and you'd know the fate of everything.

The tool for answering it was a type of stellar explosion called a Type Ia supernova. These occur when a white dwarf star in a binary system accumulates enough mass to trigger a thermonuclear detonation — and the resulting blast is extraordinarily bright, sometimes outshining an entire galaxy. Crucially, Type Ia supernovae can be calibrated into reliable distance markers. They aren't intrinsically identical — their brightness varies — but astronomers apply a correction called the Phillips relation, which links how quickly a supernova's brightness fades to its true peak luminosity. Corrected this way, they become what astronomers call standard candles: consistent enough to use as cosmic measuring sticks.

Two competing teams raced to collect enough Type Ia supernovae to measure the deceleration. Adam Riess was part of the High-Z Supernova Search Team. The other was the Supernova Cosmology Project. Both teams crunched their data expecting to measure a slowdown. Both got the same result that made no sense.

"We went from saying, you know, this has got to be wrong, to — this looks like what the data says. We have to report that," Riess says in the documentary.

In early 1998, both teams announced it independently: the expansion of the universe is not decelerating. It is accelerating. Something was counteracting gravity on cosmic scales — not just holding things steady, but actively pushing them apart faster and faster.

That something got a name: dark energy.

What 68% of Everything Being Unknown Actually Means

Under the Lambda-CDM model — our current best framework for cosmology — dark energy accounts for roughly 68% of the total energy content of the universe, with dark matter making up about 27%, and ordinary matter (everything with an atom in it: you, planets, stars, all visible gas) at about 5%. These percentages are derived from the Planck satellite's 2018 observations, and they're the best numbers we have. But Lambda-CDM is increasingly under pressure. The "Hubble tension" — a stubborn disagreement between different methods of measuring the universe's expansion rate — suggests the model may need revision.

And then there's DESI. The Dark Energy Spectroscopic Instrument, which has been mapping the cosmos since 2021 and released major results in 2024, produced something unsettling: preliminary evidence that dark energy may not be a constant force. It might change over time. If that holds up, it would mean Einstein's cosmological constant — the actual mathematical form of dark energy in Lambda-CDM — is not the right description after all. We would need something stranger.

I keep returning to this, because I think it asks a question that goes beyond physics. We have built our entire understanding of cosmology — the origin, structure, and fate of the universe — on observations of the 5% of it we can actually detect. Our instruments touch ordinary matter. Our telescopes catch photons from stars and galaxies. Everything we know, we inferred from that thin visible slice. Dark matter and dark energy are not measured directly; they are deduced from how ordinary matter behaves in their presence.

That's not a failure of science — inference is how science works, and the Lambda-CDM model makes genuinely powerful predictions. But it is a profound epistemic situation that we don't talk about enough. The history NOVA traces is one of humans repeatedly discovering that our confident picture of the universe was built on incomplete information: the Earth wasn't the center, the Milky Way wasn't the universe, the universe wasn't static, and the expansion wasn't slowing. Each correction required abandoning something that felt obvious.

Einstein invented a cosmological constant to preserve a universe that didn't exist. Then the universe handed it back to him as a description of a force he never anticipated. The constant is back in the equations — but now it represents something nobody understands, potentially doing something (changing over time, per DESI's hints) that nobody expected.

Which means the question isn't whether Einstein's blunder was actually right. The question is whether we're currently making an equally confident assumption about something that will also, eventually, turn out to be wrong — and whether we'd even know how to look.

---

Based on NOVA PBS's "Decoding the Universe: Cosmos." Watch the full episode on PBS.

— Mei Zhang, Biotech & Genetics Reporter, Buzzrag