On May 11, 1997, IBM’s Deep Blue chess computer defeated world champion Garry Kasparov in the final game of a six-game rematch held in New York, winning the overall match. IBM’s own history page and the Computer History Museum exhibit both document the event; the museum record notes this was the date on which a reigning World Chess Champion first lost a match to a computer.
Deep Blue did not “think” like a modern neural network. It won through brute-force search, evaluating enormous numbers of board positions per second using custom hardware, after IBM doubled the machine’s performance following Kasparov’s win against an earlier version in 1996. The museum record describes how Kasparov won the opening game, faltered in the second, and made a decisive error in the final game.
The match matters as a cultural and technical turning point. It proved that machines could surpass the best humans at a task long held up as a symbol of intelligence, and it set the public expectation that computers would keep climbing harder peaks. Nineteen years later that baton passed to a very different, learning-based system when DeepMind’s AlphaGo beat Lee Sedol at Go.