Cornell Model Suggests Universe Could Crunch in 20 Billion Years, Not Expand Forever

For most of the twentieth century, cosmologists expected the universe to slow down. Gravity, they reasoned, would eventually brake the expansion that began with the Big Bang. Whether the cosmos would coast forever or collapse seemed to hinge on a single question: how much matter did it contain? That picture fractured in 1998. Two independent teams studying distant supernovae discovered something that shouldn't have been there—the expansion was accelerating, not slowing. The universe was speeding up.

The simplest explanation was a cosmological constant, a term in Einstein's equations representing the energy of empty space itself. If this constant was positive and truly unchanging, it would eventually overwhelm everything else. Matter would thin as space grew, but the constant would not. Structures bound by gravity—galaxies, clusters—would hold together, but everything else would drift apart at an ever-faster rate. The default cosmic ending became heat death: a cold, sparse, eternally expanding void. That model, called ΛCDM, passed test after test and became the standard.

But the standard rested on assumption, not certainty. No one had ever directly observed what dark energy actually is. In recent years, hints began to emerge that dark energy might not be constant at all. The Dark Energy Spectroscopic Instrument, or DESI, used more than 14 million galaxies and quasars to measure the universe's expansion history with unprecedented precision. When researchers combined DESI's measurements with data from other cosmic surveys, a flexible model that allowed dark energy to change over time fit the observations slightly better than the old constant model. The preference ranged from 2.8 to 4.2 standard deviations—intriguing, but below the five-sigma threshold physicists normally require before claiming a discovery. The signal was weak, inconsistent across different datasets, and it could fade as measurements improved.

Yet the possibility was enough to prompt a question: if dark energy does evolve, what could that mean for the universe's fate? A team led by Cornell physicist S.-H. Henry Tye proposed one answer. They built a model combining an ultralight axion-like field—a hypothetical particle far lighter than conventional dark-matter axions—with a hidden negative cosmological constant. The trick was the accounting. The positive field contributed about 2.33 to the universe's total energy density, while the negative constant contributed minus 1.61. They canceled to produce the positive 0.72 that observations require today. This cancellation allowed the universe to accelerate now while concealing a negative vacuum underneath.

As cosmic expansion slowed over billions of years, the axion field would gradually lose influence. Its energy would become increasingly matter-like, diluting as space grew. The negative constant, by contrast, would not dilute. Eventually the balance would flip. The positive contribution would fade enough that the negative constant would dominate. Expansion would stop, then reverse. The universe would begin to contract.

The team numerically evolved their model forward from the present. They found that the cosmic scale factor—a measure of the universe's size—would grow to about 1.69 times its current value before reaching a maximum roughly 11 billion years from now. At that moment, the Hubble parameter, which describes the rate of expansion, would fall to zero. After that, it would become negative. Space would shrink. Matter would grow denser. The contraction would end when the universe reached a total age of 33.3 billion years. Since the universe is currently about 13.8 billion years old, this model placed a Big Crunch approximately 20 billion years in the future.

The authors were careful about what they were and were not claiming. These numbers came from one proposed physical model evaluated at one set of benchmark parameters. The model's parameter space was highly degenerate, meaning many different combinations could fit current observations. At mean values rather than best-fit values, the universe would last 40.3 billion years instead of 33.3. A zero cosmological constant also remained consistent with the data, even though negative values produced a better fit in the region the team examined. The 33.3-billion-year lifespan was not a measured countdown. It was the output of a specific calculation, no more and no less.

The real test will come from better observations. DESI has finished its main survey, with a full dark-energy analysis expected after final data processing. The Euclid space telescope is measuring cosmic structure from orbit. The Vera C. Rubin Observatory will build an enormous dataset of weak gravitational lensing and time-domain observations. NASA's Roman Space Telescope will measure supernovae and galaxy clustering with different instruments and selection effects. Agreement among these independent methods—standard rulers, standard candles, gravitational lensing, galaxy clustering, and the cosmic microwave background—will matter far more than any single dramatic number. The Cornell model is valuable precisely because it translates a speculative idea into a concrete, testable history. It shows how a universe can accelerate today and still end in collapse. But showing how something could happen is not the same as showing it will.