What If the Universe Is a Simulation? The Science and Philosophy Behind the Idea

The simulation hypothesis — the idea that what we experience as reality is a computation running on some advanced civilization's hardware — sounds like a Hollywood plot. But the argument has been made seriously by philosopher Nick Bostrom, endorsed with caveats by figures like Elon Musk, and pushed back on hard by physicists who say the computational math simply does not add up.

Thought experiment. This article explores a hypothetical scenario for educational and speculative purposes. The events described are not predictions and should not be interpreted as warnings, advocacy, or instructions.

Here is an honest, source-by-source look at the simulation hypothesis: where the argument comes from, what evidence, if any, supports it, why some physicists consider it borderline unscientific, and whether the question can ever be answered experimentally.

Key takeaways

• The simulation argument was formalized by philosopher Nick Bostrom in a 2003 paper in The Philosophical Quarterly — it is a probabilistic trilemma, not a physics claim.
• Bostrom's structure requires at least one of three propositions to be true: humanity goes extinct before reaching simulation-capable technology, advanced civilizations choose not to run such simulations, or we are almost certainly inside one.
• A 2017 Science Advances paper calculated that simulating the quantum behavior of just a few hundred particles would require a memory device built from more atoms than exist in the observable universe.
• The hypothesis is currently unfalsifiable by the standard tools of physics, which places it closer to philosophy than to an empirical scientific claim.

What the simulation hypothesis actually claims

Most people who have never read Bostrom's paper assume the "simulation hypothesis" means something like The Matrix: humans plugged into pods, unaware their bodies are batteries. That is not the argument. Bostrom's 2003 paper, "Are You Living in a Computer Simulation?", makes a narrower, more mathematical claim about probability and population statistics, not about the mechanics of how a simulation would work.

The argument starts from a simple observation. If any sufficiently advanced civilization ever builds the computing power to run "ancestor simulations" — detailed re-creations of their evolutionary history populated by conscious minds — and if they run many such simulations rather than just one, then the number of simulated conscious observers vastly exceeds the number of original, biological ones. A civilization with enough computing power could plausibly run millions of such simulations, each populated by billions of simulated people.

From there the logic is almost accounting, not physics. If simulated minds vastly outnumber "real" minds, then a randomly selected conscious observer, statistically, is far more likely to be one of the simulated majority than one of the biological minority. Bostrom is explicit that his paper does not prove we live in a simulation; it proves a trilemma, meaning at least one of three specific propositions must hold (Bostrom, 2003, The Philosophical Quarterly).

A 2017 calculation in *Science Advances* found that simulating the quantum states of even a few hundred particles would require a storage device containing more atoms than exist in the observable universe — a hard computational ceiling that has nothing to do with philosophy and everything to do with physics ([Ringel & Kovrizhin, 2017](https://www.science.org/doi/10.1126/sciadv.1701758)).

The three propositions, precisely

Bostrom's trilemma states that at least one of the following is true:

1. Almost all civilizations at our level of technological development go extinct before reaching the "posthuman" stage capable of running high-fidelity ancestor simulations.
2. Any civilization that does reach that stage is, almost universally, uninterested in running large numbers of such simulations — perhaps for ethical, resource, or cultural reasons.
3. We are almost certainly living inside a simulation run by one of these posthuman civilizations.

None of the three options is obviously false, which is exactly what makes the argument durable. Reject option three and you are implicitly betting on either mass extinction of technological civilizations or a near-universal choice not to simulate ancestors, both of which are themselves strong, unproven claims about how the future unfolds (The Simulation Argument FAQ, Bostrom).

How a simulated universe would actually have to work

Understanding why physicists disagree about this hypothesis requires understanding what "simulating a universe" would actually demand computationally. Think of the difference between a video game rendering a distant mountain and rendering the blade of grass right in front of your character's feet. Games use a trick called level-of-detail rendering: distant or unobserved objects get drawn in low resolution, and only what the player is actively looking at gets full detail. This saves enormous processing power, because most of the game world, at any given moment, does not need to be fully computed.

Some physicists have noticed that quantum mechanics behaves in a strangely similar way. A particle's exact position and momentum are not fully "rendered" until it is measured, a phenomenon called wavefunction collapse. Before observation, a particle exists in a superposition of possible states, described only by probabilities. Critics of the simulation analogy are quick to point out this parallel is suggestive rather than causal. Quantum indeterminacy has a well-developed mathematical description in standard quantum mechanics that requires no simulation to explain, and treating it as evidence of "lazy rendering" mistakes an analogy for a mechanism.

A second analogy that shows up constantly in this debate involves pixels. Digital images are built from a finite grid of pixels; zoom in far enough and you eventually hit the smallest indivisible unit, a single pixel that cannot be subdivided further. Quantum mechanics similarly shows that certain physical quantities, including energy levels in an atom and the spin of a particle, come in discrete, indivisible units rather than smooth continuous values, a property called quantization. Advocates of the simulation hypothesis argue this "graininess" looks like evidence of an underlying digital substrate, the same way a photograph looks smooth from a distance but reveals a grid up close.

Why the discreteness argument does not settle anything

Quantization was fully explained by quantum theory decades before anyone proposed the simulation hypothesis, through the work of Max Planck, Niels Bohr, and others in the early twentieth century. Discrete energy levels emerge naturally from the mathematics of wave equations confined to a boundary, the same reason a guitar string can only vibrate at specific harmonic frequencies. Nothing about quantization requires a computer underneath it. The pixel analogy is intuitive, but intuition is not evidence.

The computational cost problem, in real numbers

This is where the simulation hypothesis runs into its most serious scientific obstacle, and it comes from an unexpected corner of theoretical physics: condensed matter research, not cosmology.

In 2017, physicists Zohar Ringel of Hebrew University and Dmitry Kovrizhin of Oxford, while working on an unrelated problem involving quantized gravitational responses in exotic materials, discovered a mathematical proof with a striking side effect. They showed that simulating certain quantum many-body systems, the kind involving genuinely entangled particles rather than simple independent ones, causes computational complexity to grow exponentially with the number of particles simulated. Specifically, the "sign problem" in quantum Monte Carlo simulations makes accurate simulation of quantum systems with topological properties, such as the fractional quantum Hall effect, effectively unsolvable by any physically realizable classical computer, no matter how large (Ringel & Kovrizhin, Science Advances, 2017).

The scale of the problem is genuinely hard to fit in your head. Their calculations imply that a classical simulation of the quantum state of just a few hundred entangled electrons, tracked with full fidelity, would require a memory storage device built from more atoms than exist in the entire observable universe. Extend that logic to a universe containing roughly 10^80 atoms, and the resource requirement to simulate reality at the quantum level, atom by atom, moment by moment, does not just exceed available computing power. It exceeds any conceivable computing power built from ordinary matter, because you would need more matter to build the computer than the universe you are trying to simulate contains.

Physicists have not proven a simulation is impossible in every conceivable form. They have shown that a full, atom-for-atom, quantum-accurate simulation of a universe our size is impossible using any computer built from ordinary matter within that same universe. A "simulator" running our reality would need to sit in a much larger universe with fundamentally different physical limits.

Could level-of-detail rendering solve the resource problem?

This is the most common rebuttal, and it deserves a fair hearing. If a simulation only needs to render what conscious observers are actively measuring, in principle it could skip the astronomical cost of maintaining every particle in full quantum detail everywhere, all the time. The problem is that this requires the simulation to already know, in advance, which particles will eventually be observed and by whom, which reintroduces exactly the kind of infinite regress and hidden-variable requirements that quantum mechanics was specifically designed to avoid. It is a patch that solves one problem by creating a harder one.

Why some physicists take the idea seriously anyway

Despite the computational objections, several respected physicists have engaged with simulation-adjacent ideas without dismissing them outright, largely because of independent developments in theoretical physics that were not originally about simulations at all.

The holographic principle. Proposed by Gerard 't Hooft and developed extensively by Leonard Susskind, the holographic principle suggests that all the information contained within a three-dimensional volume of space can, in principle, be fully encoded on a two-dimensional boundary surrounding that volume, similar to how a hologram stores a 3D image on a flat surface (Susskind, "The World as a Hologram"). This idea grew out of black hole thermodynamics, where physicists found that a black hole's entropy, meaning the amount of information it can contain, scales with the surface area of its event horizon rather than its volume, a deeply strange result now called the Bekenstein bound, formulated by Jacob Bekenstein in 1981. The Bekenstein bound sets a hard mathematical ceiling on how much information any bounded region of space can contain, given its size and energy (Scholarpedia: Bekenstein bound).

Neither the holographic principle nor the Bekenstein bound was proposed to support the simulation hypothesis. They emerged from mainstream general relativity and quantum field theory research. But they show that reality genuinely does behave, in a mathematically rigorous sense, as though information about a volume can be compressed onto a lower-dimensional surface, which is uncomfortably close to how a rendering engine compresses a 3D scene into a stream of data.

The mathematical universe hypothesis. MIT physicist Max Tegmark has separately argued that the universe is not merely described by mathematics but literally is a mathematical structure, and that any sufficiently complex, self-consistent mathematical structure exists in the same sense that our universe exists. This is a distinct claim from Bostrom's simulation argument. Tegmark is not saying someone is running a program; he is saying mathematical existence and physical existence may be the same thing. It is philosophically adjacent to the simulation hypothesis, since a sufficiently complete mathematical description is, by definition, computable, but it does not require a "simulator" in any conventional sense.

Why most physicists remain skeptical

The skepticism is not casual dismissal. It rests on specific, articulable objections.

The resource argument, again. The Ringel and Kovrizhin result described above is the single strongest scientific objection, because it is a rigorous complexity-theory proof, not a hand-wavy intuition. Simulating quantum entanglement at scale is not merely difficult; the resource requirements grow exponentially, meaning no incremental improvement in computing technology, however dramatic, closes the gap. You would need a fundamentally different, larger physical universe to host the computer.

Consciousness is assumed, not established. Bostrom's probability argument only works if simulated minds are genuinely conscious in the same sense biological minds are. If a simulated brain is not actually experiencing anything, merely producing outputs that look conscious from the outside, then the entire population-statistics logic collapses, because there would be no reason to count simulated "minds" alongside biological ones in the probability calculation. Neuroscience and philosophy of mind have no settled answer to whether consciousness requires biological substrate, a computational process, or something else entirely, which means this is a load-bearing assumption the argument cannot currently support with evidence.

Unfalsifiability. German theoretical physicist Sabine Hossenfelder has argued publicly that the simulation hypothesis, as currently formulated, makes no testable predictions that distinguish it from ordinary physical reality. Every observation consistent with "we are in a simulation" is equally consistent with "the universe simply has these properties intrinsically." A hypothesis that cannot in principle be distinguished from its negation through observation falls outside the domain the scientific method is built to adjudicate; Hossenfelder has characterized it as philosophy wearing the vocabulary of physics rather than a genuine scientific theory (Hossenfelder, public commentary and Existential Physics, 2022).

No unique predictions. For a hypothesis to earn scientific standing, it typically needs to predict something specific that rival hypotheses do not. So far, no proposed feature of physics, from quantum discreteness to the speed of light, uniquely requires a simulation to explain it. Every candidate "hint" has a complete, self-contained explanation within standard physics that predates and does not depend on the simulation idea.

What proposed tests exist, and why none has worked

A handful of physicists have tried to move the debate from pure philosophy toward testable physics. None of the attempts has produced a result that uniquely confirms or rules out a simulation.

Lattice spacing in cosmic rays

If space itself were a discrete computational grid, similar to pixels on a screen, then particles with extremely short wavelengths, meaning ultra-high-energy cosmic rays, should interact with that underlying grid structure in a detectable way, producing a directional bias, or anisotropy, in how they arrive at Earth relative to the orientation of the hypothetical lattice. Physicists including Silas Beane and collaborators proposed a version of this test in 2012, examining whether the highest-energy cosmic rays observed show any grid-like directional preference. No such anisotropy has been detected at any energy level current detectors can reach, though this does not rule out a grid at a much finer resolution than instruments can currently probe.

Information limits and the Bekenstein bound

If a region of space contained more information than the Bekenstein bound allows, meaning more entropy than the mathematics of general relativity and quantum mechanics permit for its size and energy, that would be a serious anomaly worth investigating, simulation or not. No violation of the Bekenstein bound has ever been observed, which is consistent with ordinary physics and provides no particular evidence for or against a computational substrate.

"Glitches," Mandela effects, and déjà vu

Simulation advocates in popular culture sometimes point to widespread false memories, so-called Mandela effects, as evidence of rendering errors in reality. Cognitive scientists have thoroughly documented how human memory works, and false collective memories are well explained by ordinary mechanisms: source confusion, social reinforcement of a plausible but incorrect detail, and the reconstructive, rather than recording-like, nature of human recall. There is no controlled, reproducible evidence linking these phenomena to anything beyond known psychology.

In practice, every proposed experimental test to date has returned a null result or has been criticized as untestable in principle, which is a large part of why the hypothesis has not migrated from philosophy departments into mainstream experimental physics programs.

Bostrom's argument compared to other big-picture physics questions

It helps to see where the simulation hypothesis sits relative to other major questions in physics that sound similarly speculative but occupy very different scientific standing.

Hypothesis	Testable predictions?	Scientific status	Key proponents
Simulation hypothesis	None unique to the claim	Philosophical argument; unfalsifiable by current methods	Nick Bostrom, discussed by Elon Musk
Many-worlds interpretation	Same predictions as standard quantum mechanics; debated on interpretive grounds	Serious minority interpretation within physics	Hugh Everett III, championed today by Sean Carroll
Eternal inflation / multiverse	Some indirect predictions (bubble collision signatures) sought in cosmic microwave background data	Active cosmological research program with proposed, largely unconfirmed, tests	Andrei Linde, Alan Guth
Holographic principle	Makes precise, mathematically verified predictions within specific model systems (AdS/CFT)	Established, widely used tool in theoretical physics	Gerard 't Hooft, Leonard Susskind
String theory	Predicts extra dimensions and particles, largely beyond current experimental reach	Mathematically rich but experimentally unconfirmed	Edward Witten, and the broader string theory community

Table: How the simulation hypothesis compares to other speculative physics ideas, ranked by testability and scientific standing.

The pattern that emerges is instructive. The simulation hypothesis is the only entry on this list built entirely from probability reasoning about hypothetical future civilizations, rather than from equations that make falsifiable predictions about our own universe. That is precisely why physicists treat it differently from the multiverse or many-worlds, even though all three sound equally exotic to a casual reader.

What it would mean for physics if it were true

Assume, purely as a thought experiment, that proposition three of Bostrom's trilemma is correct and we are inside a simulation. What would actually change about how physics operates day to day? Almost nothing, and that is one of the more underappreciated features of the hypothesis.

The laws of physics we measure would still hold exactly as measured, because "the simulation" is simply whatever produces those consistent laws, whether that substrate is a base-reality universe or a computational one running inside a larger reality. Gravity would still pull objects together at 9.8 meters per second squared (32.2 feet per second squared) at Earth's surface. Light would still travel at exactly 299,792,458 meters per second in a vacuum. Evolution by natural selection would still explain the diversity of life, because natural selection describes a pattern in how traits propagate under selective pressure, a pattern that would exist regardless of the underlying substrate computing it.

This is sometimes called the "meta-consistency" objection to the simulation hypothesis mattering practically: a simulation good enough to be indistinguishable from base reality is, for every operational purpose that touches your life, indistinguishable from base reality. Medicine still works the same way, engineering still works the same way, and every prediction physics makes about the observable universe remains equally reliable either way.

Common confusions and things people get wrong

Confusion 1: Bostrom claims we probably live in a simulation. He does not. His paper is explicit that it establishes a trilemma, and he has stated in interviews that he considers the probability roughly comparable across the three branches, without strongly favoring the simulation branch as the correct resolution. Popular coverage frequently drops the "trilemma" framing and reports only the third, most dramatic branch.

Confusion 2: Elon Musk's "billion to one" comment is a scientific estimate. Musk's widely quoted remark, made at the 2016 Code Conference, that "there's a billion to one chance we're living in base reality," was an extrapolation from the observation that video games have improved rapidly over a few decades, from Pong to photorealistic 3D graphics, and a projection that this trend continues indefinitely (TechCrunch, 2016, "A running tab of what tech people think about whether we're living in a simulation"). It is an argument by analogy from consumer entertainment technology, not a calculation grounded in the physics of computational complexity, and it does not engage with the Ringel-Kovrizhin resource problem at all.

Confusion 3: Quantum weirdness is direct evidence of a simulation. As covered above, quantization and wavefunction collapse have complete, self-contained explanations inside standard quantum theory that were developed decades before the simulation hypothesis existed. Resemblance to a video game mechanic is an analogy, not a causal link, and analogies are not evidence in physics.

Confusion 4: If it can't be disproven, it must be taken seriously as science. Unfalsifiability cuts the opposite direction in the scientific method. A claim that cannot in principle be distinguished from its negation through any observation is, by Karl Popper's classic demarcation criterion, outside the domain science is equipped to evaluate, not a strong candidate awaiting confirmation.

Famous opinions, and how much weight they carry

• Nick Bostrom (philosopher, University of Oxford, originator of the argument): Has described the probability as worth taking seriously without committing to a specific number in his formal writing, though he has discussed something in the range of a meaningful, non-trivial probability in various interviews.
• Elon Musk (entrepreneur): "There's a billion-to-one chance we're in base reality," reasoning from the trajectory of video game realism rather than from physics.
• Neil deGrasse Tyson (astrophysicist, American Museum of Natural History): Has offered roughly even odds in informal public debates, treating it as an open, interesting question rather than a settled physical claim.
• Lisa Randall (theoretical physicist, Harvard University): Has publicly characterized the probability as extremely low and has criticized the argument for resting on unverifiable assumptions about future civilizations.
• Sabine Hossenfelder (theoretical physicist, author of Existential Physics): Has argued the hypothesis is not falsifiable and therefore sits outside what physics as a discipline can currently adjudicate.

None of these are peer-reviewed physical measurements. They are informed opinions from people with different levels of direct expertise in the relevant complexity theory, and they should be weighted accordingly rather than treated as competing scientific findings.

Frequently asked questions

Q: Could we ever prove we're in a simulation? Probably not through experiment alone. Any simulation sophisticated enough to be worth building would likely be designed so its inhabitants cannot detect the substrate underneath it, which means the argument is currently stuck at the level of probability and inference rather than direct measurement.

Q: Does the Mandela effect prove the simulation hypothesis? No. The Mandela effect, in which large groups of people misremember the same fact in the same incorrect way, is well explained by established cognitive science, including source confusion and the reconstructive nature of memory. It is a documented feature of how brains work, not evidence of rendering glitches in reality.

Q: Does the simulation hypothesis require consciousness to be simulatable? Yes, and this is one of its weakest assumptions. The probability argument only works if simulated minds have genuine subjective experience comparable to biological minds. If simulated brains merely produce conscious-looking behavior without actual experience, the population-counting logic behind Bostrom's trilemma no longer holds.

Q: How much computing power would it take to simulate the entire universe? Far more than appears physically achievable. The 2017 Ringel and Kovrizhin calculation in Science Advances found that even a few hundred entangled quantum particles would require a memory device built from more atoms than exist in the observable universe, implying that a full quantum-accurate simulation of our roughly 10^80-atom universe is not achievable with any computer built from ordinary matter inside that same universe.

Q: Is the simulation hypothesis considered real science by physicists? Most physicists classify it as a philosophical argument rather than an empirical scientific theory, mainly because it currently makes no unique, testable predictions that distinguish "we are simulated" from "the universe simply behaves this way." Sabine Hossenfelder and other physicists have specifically criticized it on falsifiability grounds.

Q: What is the difference between the simulation hypothesis and the multiverse theory? The multiverse, in its various forms including eternal inflation and the many-worlds interpretation of quantum mechanics, emerges from equations within established physical theories and, in some versions, makes indirect predictions that researchers actively search for in cosmological data. The simulation hypothesis is a probability argument about hypothetical future civilizations and does not derive from any equation in current physics.

Q: Did any physicist actually prove the universe isn't a simulation? Not in an absolute sense. The 2017 Science Advances paper proved that a specific, physically realistic class of quantum simulations becomes computationally intractable at scale using classical computers. That is strong evidence against a straightforward "computer built from ordinary matter" simulating our universe in full quantum detail, but it does not rule out every conceivable form the hypothesis could take, including a simulator operating under entirely different physical laws.

Q: Why do quantum mechanics and the simulation hypothesis get mentioned together so often? Because two features of quantum mechanics, the discreteness of certain physical quantities and the fact that particles behave probabilistically until measured, superficially resemble how digital simulations conserve computing resources. Both features have complete, non-computational explanations within standard quantum theory that predate the simulation hypothesis by decades, so the resemblance is an analogy rather than supporting evidence.

Q: What would change about daily life if we found out we were in a simulation? Very little in practical terms. The laws of physics we measure would continue to behave exactly as they do now, because those laws describe whatever process generates our observations, computational or otherwise. Medicine, engineering, and every predictive model built on physics would remain equally reliable.

The takeaway

The simulation hypothesis survives as one of the more durable thought experiments in modern philosophy of science precisely because Bostrom's trilemma is genuinely difficult to escape on logical grounds alone. But "difficult to logically dismiss" is not the same standard as "supported by physical evidence," and on that second, higher bar, the hypothesis currently falls short. The strongest rebuttal is not philosophical but computational: Ringel and Kovrizhin's 2017 proof that quantum many-body simulation costs scale exponentially shows that no computer built from ordinary matter inside a universe our size could simulate that universe in full quantum fidelity. Whether reality is computed on some vastly larger substrate or simply is what it appears to be, the same laws govern your Tuesday morning either way.

---

Read next:

• What If Parallel Universes Are Real? The Physics Behind the Multiverse
• What If Time Travel Were Physically Possible? The Real Physics Behind the Idea
• What Is the Speed of Dark — And Other Things That Move Faster Than Light