Elon Musk and the Media Chess Match: 10^120 Ways to Prove Chess Cannot Be 'Solved'

People call it a shock; I call it unread data.

On November 25, 2026, on platform X, Elon Musk wrote a short sentence: 'Chess is essentially solved. Like Go. Soon, nothing interesting remains.' Within hours, the official Chess.com account – the world's largest online chess platform with over 200 million users – replied with a cryptic Vietnamese phrase: 'Do kỹ năng thôi' (Just skill). This seemingly playful exchange touched upon one of the most widespread scientific misconceptions about humanity's oldest intellectual game.

As someone who has followed chess since 2026, when I sat in the front row of a regional tournament in Ho Chi Minh City, I have witnessed countless declarations that 'chess is dying.' From Deep Blue in 2026 to AlphaZero in 2026, from engine games to machines defeating world champions, there has always been someone proclaiming that chess would lose its meaning. But this time, when a tech billionaire with over 200 million followers declares that the game will 'soon be uninteresting,' the issue is no longer merely technical. This is an attack on the very cultural value of chess.

So, what does the data actually say? Let me trace the chain of numbers.

Context: A four-year war of words

The dispute between Musk and Chess.com did not happen overnight. It began in June 2026, when Musk, in a series of Twitter statuses, declared that chess was 'dead' at the highest competitive level because computers had surpassed humans by too wide a margin. Chess.com responded with 'Do kỹ năng thôi' – a mocking phrase implying that if Musk lost to a machine, it was because of his own skill, not because the game was dead.

By 2026, Musk deepened his argument in an interview: 'Chess is too simple compared to real life. No fog of war, no tech tree, and the starting positions of both sides are perfectly symmetric.' He compared chess to more complex strategy games like StarCraft – where players must deal with imperfect information and countless uncertain variables.

By late 2026, the conflict reignited. This time, Chess.com did not only use English but also Vietnamese – perhaps a nod to Vietnam's large and passionate chess community. When Musk called the Chess.com account operator an 'intern,' Chess.com replied with witty comebacks that delighted both tech and chess circles.

Chess.com's CEO, Danny Rensch, also spoke up. In an interview with Business Insider, Rensch said: 'We don't view AI as the destroyer of chess. AI is a tool to help humans understand the game more deeply. And we are developing a superhuman chess AI for fans – something that excites them, not frightens them.'

This exchange – seemingly just a tech-world joke – raised serious questions: Is chess truly 'solved'? If so, does that mean the game is about to die? If not, why does an intelligent person like Musk misunderstand it so profoundly?

The number 10^120: The truth behind Chess.com's claim

During the debate, Chess.com cited a number: about 10^120 possible chess games. Musk countered that this number was wrong – that the number of legal positions is only about 10^44, and therefore Chess.com's claim was exaggerated.

Who is right? The answer is: both are correct, but they are talking about two different things.

The number 10^120 was proposed by information-theory pioneer Claude Shannon in 2026 in his famous paper 'Programming a Computer for Playing Chess.' Shannon calculated that the number of possible chess games (i.e., the number of legal move sequences from the starting position to the end) is approximately 10^120. This is the so-called 'game-tree complexity' – a measure of the number of branches in the entire space of possible games.

The number 10^44 that Musk mentions is the number of legal positions on the board – the number of possible board states according to the rules, regardless of move order. This number is far smaller than 10^120 because each position can be reached through many different move sequences.

But the crucial point is: to 'solve' chess – that is, to determine the optimal outcome (win/draw/loss) from every legal position assuming perfect play by both sides – the computer must explore the game-tree space, not just the position space. Even with only 10^44 positions, evaluating each position in a mathematically rigorous way remains impossible with current or near-future technology.

To visualize: the number of atoms in the observable universe is estimated at about 10^80. The number 10^120 exceeds this by an unimaginable margin. If you tried to store the entire chess game tree on hard drives made from the universe's matter, you would need a number of universes greater than the number of sand grains on all of Earth's beaches.

From a computer science perspective, a game is considered 'solved' when there is an algorithm proving the optimal outcome from the initial position. Tic-tac-toe is fully solved – we know with certainty that with perfect play, the result is always a draw. Checkers was also solved in 2026 by a research group at the University of Alberta, Canada, after nearly two decades of computational effort. Chess is not – and likely never will be solved in any foreseeable future.

Why? Because chess's complexity lies not only in its number of positions, but in the branching structure of the game. From the initial position, White has 20 legal moves. After both sides' first move, there are 400 possible positions. After three moves (each side), the number is 8,902. After four moves, 197,281. After ten moves, the number reaches trillions. By move 20, the number of possible positions exceeds 10^30. The chess game tree branches at a dizzying rate, and by depth 40 – that is, a relatively short game at grandmaster level – the number of branches far exceeds the computational capacity of any supercomputer.

Does this mean Musk is wrong in asserting chess can be 'solved' in the near future? Technically, the answer is: almost certainly wrong. But to understand why, we need to look at the history of attempts to solve chess by computer.

Core: A chain of data evidence about unconquerable complexity

Chess's fate differs from tic-tac-toe or checkers. In 2026, after 18 years of continuous computation, the University of Alberta team proved that checkers – with about 5×10^20 legal positions – always ends in a draw when both sides play perfectly. This was an astonishing mathematical achievement. But checkers has only 10^20 positions; chess has 10^44. The gap between 10^20 and 10^44 is a gap that cannot be bridged by ordinary hardware improvements.

To solve checkers, researchers used a method called 'computer-assisted proof,' in which they divided the state space into smaller regions and used retrograde analysis to determine the value of each position. For chess, this method applies only to endgames with very few pieces. Currently, endgame tablebases have solved every position with 7 pieces or fewer. In 2026, the Lomonosov project completed the 7-piece tablebase, containing complete results for all positions with at most 7 pieces on the board – about 500 trillion positions. That is a massive number, but only a tiny fraction of 10^44.

From my 37 years of experience covering major tournaments, I have realized there is an important difference between a computer playing chess well and a computer 'solving' chess. A computer playing chess well does not mean it knows the optimal answer to every position. It only means it uses search algorithms and heuristic evaluations to find a good move within a limited time. Deep Blue defeated Garry Kasparov in 2026 by evaluating about 200 million positions per second – but that number is still a drop in the ocean of 10^44 positions. AlphaZero, launched in 2026, uses neural networks and self-play to learn position evaluation – but it still has not 'solved' chess in the mathematical sense.

Another way to visualize chess's complexity: consider the number of games that end after 40 moves. At grandmaster level, a game typically lasts 40–60 moves. Assuming an average of about 30 legal moves per side per turn (the actual figure is around 30–35), the number of possible games in 40 moves is roughly 30^80 – approximately 10^118. This number approximates Shannon's 10^120. And because the search space grows exponentially, adding one more move multiplies the possibilities by about 30.

The number 10^120 is not arbitrary. It has been used in game theory calculations since the 1950s and remains the figure used by computer scientists to describe chess complexity today. Chess.com is not wrong to cite it – they are referring to game-tree complexity, not the number of legal positions. But Musk is also not entirely wrong in citing 10^44 – he is talking about the number of reachable board states. This seemingly minor distinction determines the entire answer to the question 'can chess be solved?'

The critical point is: even if there are only 10^44 positions, determining the value of each of those positions is a computational problem whose complexity far exceeds the capacity of any conceivable computer in the physical universe. Gödel's incompleteness theorem and computational complexity theory tell us that some problems cannot be solved by any algorithm – and chess, while not theoretically unsolvable, belongs to a class of problems where the required computation time grows exponentially with the size of the state space.

To illustrate: suppose we have a supercomputer that can analyze one position per nanosecond (i.e., one billion positions per second). To analyze all 10^44 chess positions, we would need 10^35 seconds – about 3×10^27 years. The age of the universe is about 13.8 billion years, or 1.38×10^10 years. That means we would need a time span roughly 2×10^17 times the age of the universe to complete the analysis of the entire chess position space. This is an absurdly large number compared to any timescale in the universe.

However, Musk could argue that intelligent algorithms – like those AlphaZero employs – can 'compress information' and discover patterns that solve problems far faster than brute-force search. This is a theoretically sound argument. But it remains speculative – there is no evidence whether chess can be solved by some 'smart' shortcut. Computational complexity theory tells us that some problems can be solved faster if they have special structure – but whether chess has such structure remains an open question.

Contrarian angle: 'Solving' chess does not mean chess dies

Even if one day a super-AI 'solves' chess – proving that from the starting position, with perfect play, the result is a draw (or a win for White/Black) – would that make chess boring? Absolutely not. And this is the biggest blind spot in Musk's argument.

Look at football. We have accurate mathematical models of the game, but nobody says football is 'solved' simply because we know the rules. Football's appeal lies in its uncertainty, in individual players' skills, in coaches' tactics, in fans' emotions. Chess is the same. Humans do not play chess like computers – they play with intuition, emotion, understanding of opponents, and the ability to read psychology.

Throughout my career covering chess, I have witnessed countless games where human players chose moves that were not computer-optimal but were psychologically correct – putting opponents in uncomfortable positions, forcing them to miscalculate, or simply creating time pressure. Computers can calculate most accurately, but humans cannot – and this very imperfection creates art.

Magnus Carlsen, the 16th World Champion and the highest-rated player in history (2882 in 2026), once said: 'Chess is not a sport where the best player always wins. It is a sport where the player who handles pressure the best usually wins.' Carlsen understands something outsiders often miss: top-level chess is not just a battle of numbers, but a battle of psychology and endurance.

Another proof that 'chess is dead' is fallacious: look at the explosive growth of online chess in recent years. The COVID-19 pandemic in 2026 pushed millions to Chess.com and Lichess. According to Chess.com statistics, the number of games played daily on the platform increased from about 10 million pre-pandemic to over 30 million during the 2026 peak and remained high afterwards. By 2026, Chess.com surpassed 200 million registered users – a figure that even the most optimistic chess growth advocates would not have dared dream of in 2026, when Deep Blue defeated Kasparov and many predicted chess would become meaningless.

Contrary to pessimistic predictions, the advent of powerful engines like Stockfish, Leela Chess Zero, and AlphaZero has made chess more popular than ever. Fans can now watch top grandmasters compete with analytical depth previously reserved for world champions. Platforms like Chess.com have integrated engines into live viewing interfaces, allowing fans to see the computer's evaluation of each move in real time. Rather than killing excitement, AI has made chess more accessible and engaging for mass audiences.

This brings us to one of the most profound observations about the AI-chess relationship: AI does not replace human emotion; it explains why our hearts race when watching a tense game. When Gukesh Dommaraju – the 18-year-old Indian player – earned the right to challenge Carlsen at the 2026 World Chess Championship, the world witnessed not just a battle between two individuals, but between two generations and two philosophies of playing chess. Gukesh represents the generation of players raised with computers, using engines as an indispensable training tool. Carlsen, though also using engines, retains a romantic spirit of pre-AI chess. This confrontation – if it happens – will be one of the most anticipated matches in modern chess history, and it could not exist if chess were 'dead' as Musk claims.

Another blind spot in Musk's argument is his confusion between 'chess is solved' and 'top-level chess is dominated by AI.' It is true that since around 2026, no human player has been able to defeat a chess engine running on powerful hardware. Since Deep Blue beat Kasparov, the computer's superiority over humans in chess has been incontestable. But that does not mean chess – the human game – has been 'solved.' Just as airplanes can fly faster than birds does not mean birds are about to go extinct. Humans still watch birds fly, still study birds, still delight in birds' beauty – because the value of a bird lies not in whether it can fly as fast as possible, but in its very existence.

Even in the AI field, there is an interesting irony in Musk's argument. If chess is 'soon to be solved' by a super-AI, then why is xAI – Musk's own AI company – investing in developing Grok, an AI model capable of playing chess? And why, when AI models competed against each other in a chess tournament in 2026, did xAI's Grok 4 lose to OpenAI's o3? If chess is truly 'no longer interesting,' why do leading AI companies treat it as a standard benchmark for model capability?

Numbers are asceticism: one must abandon indulgence to see the truth. The truth here is: chess has become a benchmark for artificial intelligence. It is not just a game – it is a laboratory where ideas about intelligence are tested. From Claude Shannon in 2026 to AlphaZero in 2026, chess has accompanied AI's development. This does not diminish chess's value to humanity – on the contrary, it enhances chess's place in human culture.

But if chess fans have a reason for concern, it is something different: the commercialization and 'industrialization' of chess. This phenomenon resembles what I have seen in esports: professionalization is turning young players into products of digital training pipelines, and individual playing styles are being sanded down by data-driven coaching. In chess, young players' increasing reliance on engines may create a generation of technically superb but creatively lacking players. This is the real threat to chess – not the idea that AI might 'solve' the game.

Takeaway: Signals that chess is alive and will continue to live

After all this data analysis, one thing emerges clearly: Elon Musk – though an outstanding entrepreneur – deeply misunderstands chess. And this misunderstanding is not a mere scientific error. It reflects a worldview that reduces intelligence to algorithm, values games only by computational difficulty, and dismisses anything unmeasurable by numbers as meaningless.

This is a worldview I have encountered many times in my data journalism career. I remember in 2026, when I wrote an xG (expected goals) analysis of the Japan-Belgium match at the World Cup, an older male colleague dismissed my article: 'Women looking at football data only pick convenient numbers.' He did not believe a woman – a data journalist – could deeply understand the sport he loved. But data has no gender, and truth needs no permission. My article was eventually published and exceeded 200,000 reads – a figure that silenced him.

Chess is the same. It does not need Elon Musk's permission to exist. It does not need a super-AI to 'solve' it to prove its value. Chess's value lies in humanity itself: in children learning logical thinking through each move, in elderly players training their minds through each game, in fans finding joy and connection through this intellectual sport.

Age is the only variable that never lies. Chess has existed for over 1,500 years. It has witnessed the rise and fall of empires and the changes of eras. It survived the age without computers, the age of computers' birth, and the age of computers' dominance. And it will continue to exist long after all of us – including Elon Musk – have turned to dust.

In December 2026, when I watched Gukesh Dommaraju become world champion as the youngest in history at 18, I recalled a quote from 11th World Champion Bobby Fischer. When asked why he loved chess, Fischer replied: 'Chess is a perfect game. It has everything: art, science, sport.' More than half a century later, that statement retains its value. And it will retain its value for centuries to come, regardless of how AI develops.

For chess lovers, the message is simple: do not fear AI. Do not fear declarations that chess will die. Keep playing, keep learning, keep enjoying the beauty of this game. Because every chess game between two humans is unique – never repeated – and that uniqueness is the magic that no algorithm can replicate.

The final question I want to pose here – and the question every tech person like Musk should ask themselves – is: If one day AI 'solves' chess – proving that the optimal result is a draw – would you stop enjoying a beautiful game between Carlsen and Gukesh? If your answer is 'no,' then you have already answered yourself: chess was never merely a problem to solve. It is a symphony of the human mind – and that symphony will continue to resonate regardless of what AI does.

The data has spoken. And the data supports chess.