Where does what we can see end — and where does pure speculation begin? A journey to the outermost edge of cosmological knowledge: from the observable horizon to questions that physics itself leaves open.
"How big is the universe?" sounds like a simple question — but it isn't, because it can actually mean two different things. The observable universe is the sphere around us from which light has had time to reach us since the Big Bang. Its boundary is not a physical wall but a matter of timing: everything farther away presumably exists too, but its light simply hasn't arrived yet. The universe as a whole, by contrast, could be far larger than this visible slice — or even infinite. Observations cannot show this directly, because we can fundamentally only measure the part from which light reaches us.
This ambiguity is why "the edge of the universe" often causes confusion in popular accounts. There is very likely no edge in the sense of a boundary or wall — neither for the observable universe (which merely marks our observation limit) nor for the universe as a whole (which simply doesn't seem to need a boundary, whether finite or infinite).
Every statement about the universe beyond our observation horizon rests on an assumption that cannot be directly proven: the cosmological principle. It states that the universe looks approximately the same everywhere on large scales (homogeneous) and has no preferred direction (isotropic) — that Earth, in other words, occupies no special vantage point. All measurements within the observable universe support this assumption so far, but they cannot guarantee it: by nature, we only ever see a single slice.
From our Solar System to the edge of the observable lies more than 25 orders of magnitude in distance. Each shell in this diagram marks a cosmic structure of higher order — and, at the very outside, the subtle but crucial differences between three terms that are often used synonymously in cosmology, even though they aren't the same.
Remarkably, the event horizon lies inside the observable universe, not outside it. The observable universe collects light from the entire past and is correspondingly large; the event horizon describes how far into the future we can still influence or see anything at all — and it shrinks in relative terms, because dark energy accelerates the expansion. Galaxies that are still visible today will, over time, disappear beyond this horizon.
The classical Big Bang model alone cannot explain two observations: Why is the cosmic microwave background nearly identical in temperature in every direction, even though distant regions would never have had enough time to reach thermal equilibrium with each other (horizon problem)? And why is the geometry of the universe so close to perfectly flat, when any small deviation should have amplified exponentially over billions of years (flatness problem)? In 1980/81, physicist Alan Guth proposed a solution: a phase of extremely rapid expansion directly after the Big Bang.
Inflation was originally an ad hoc solution to two problems — but it now also predicts independently confirmed details of the cosmic microwave background, such as the precise statistical pattern of its tiny temperature fluctuations. That makes inflation far more than a mere fix: it is one of the best-tested ideas in early cosmology, even though the exact nature of the inflaton field remains unknown.
General relativity permits three fundamental geometries for space, depending on the density parameter Ω: closed and spherical (Ω>1), open and hyperbolic (Ω<1), or exactly flat (Ω=1). In 2018, the ESA's Planck satellite measured tiny temperature fluctuations in the cosmic microwave background — their size distribution depends sensitively on the curvature of space through which the light traveled for 13.8 billion years.
The result: the curvature parameter Ω_K is 0.0007 ± 0.0037 (95% confidence) — statistically indistinguishable from zero. The observable universe is thus flat to within measurement precision. This does not rule out a very weak curvature whose radius would be far larger than the observable universe — exactly what one would expect if inflation (Chapter 03) stretched any original curvature almost, but not quite completely, flat.
All of the following ideas are mathematically motivated and are discussed by serious physicists — but none of them yet makes a testable, falsifiable prediction about our own observable universe. That is precisely their fundamental problem: if another universe is by definition never observable, no observation can ever refute the hypothesis — a core criterion, according to Karl Popper, for what counts as a scientific theory at all. Many physicists therefore treat multiverse models as speculative but worthwhile consequences of established theories — not as confirmed physics.
Physicist George Ellis and others argue that pure multiverse hypotheses cross into metaphysics as long as they provide no testable consequences for our universe. Proponents counter that accepting unobservable consequences of an otherwise well-tested theory (such as inflation) is methodologically nothing unusual — the same is done for other physical extrapolations. The debate remains unresolved to this day.
How the universe ends depends critically on how dark energy behaves over cosmological timescales — a topic covered in more depth on the previous Dark Matter & Energy page. Here are just the four most-discussed future scenarios, ranked by present-day plausibility.
That the expansion is accelerating rather than — as expected from the gravity of all matter — decelerating was a surprise in 1998. Two independent teams measured Type Ia supernovae, whose absolute brightness is known, and found that distant examples were dimmer than a uniformly decelerating universe would allow — meaning they were farther away than a decelerating model predicts. This discovery was awarded the 2011 Nobel Prize in Physics. To this day, it remains the central empirical support for the existence of dark energy, and thus for the Big Freeze scenario as the currently favored future.
Three thoughts that show just how unfamiliar cosmological thinking is compared to everyday intuition.
Not because we move through space — but simply because time passes. The longer ago the Big Bang happened, the more time light from ever more distant regions has had to reach us. Every second that passes, new light literally arrives from farther out, and the sphere of the observable grows — regardless of whether or how fast Earth itself is moving.
For many physicists, this question may not even be meaningfully posed — not because the answer is unknown, but because "before" is a temporal concept, and according to prevailing models, time itself only begins to exist with the Big Bang. Asking what came before would then be a category error similar to asking what lies north of the North Pole. Models such as Hartle and Hawking's "no-boundary proposal" try to capture this mathematically: time "rounds off" instead of beginning abruptly at a point. None of these models has been proven.
While the observable universe has a radius of roughly 10²⁶ meters, the Planck length (1.6 × 10⁻³⁵ m) is considered the smallest physically meaningful distance — below it, our notions of smooth space break down, because quantum effects of gravity should dominate. Between the smallest and largest known scales, then, lie about 61 orders of magnitude — a span the human imagination can only grasp in numbers, no longer in pictures.