The Antikythera Mechanism: Humanity’s Most Mysterious Ancient Computer

In the spring of 1900, a group of sponge divers from the island of Symi had taken shelter from a storm in the waters near the Greek island of Antikythera, where the Aegean and Mediterranean seas meet. Under the command of Captain Dimitrios Kondos, one of the divers descended to the seabed forty-five metres below—and surfaced in a hurry. What he reported was shocking: “There are piles of human bones and horse bones on the bottom.” His crewmates assumed the pressure had driven him mad. When Kondos himself went down to check, he came back up with a bronze arm from a statue in his hand.^[1]

This was the beginning of one of the most revolutionary archaeological discoveries in human history. Beneath the surface lay a Roman-era shipwreck from the first century BC, packed with statues, coins, and pottery. But hidden among the wreckage was something no one had imagined could exist there—a mass of corroded bronze that would overturn everything we thought we knew about the technological capabilities of the ancient world.

The Machine from the Shipwreck, Ignored for Two Years

The salvage operation of 1900–1901 was carried out with support from the Greek navy. After gruelling work that cost one diver his life and left two others paralysed, dozens of artefacts were transferred to the National Archaeological Museum in Athens.^[2] The marble and bronze sculptures drew all the attention—most famously the Youth of Antikythera. Several heavily corroded bronze lumps were set aside without much notice.

Two years passed.

On 17 May 1902, the archaeologist Valerios Stais was examining one of those corroded lumps when he noticed something strange: a small gear wheel embedded in the bronze surface, accompanied by inscribed Greek letters.^[3] Stais immediately recognised this as part of a remarkable device. The academic response was sceptical. It seemed impossible that first-century-BC Greece could have produced such a precision gear mechanism. Some suggested the fragments had been mixed with artefacts from a different era.

For more than half a century, the mechanism languished outside the mainstream of scholarly attention.

120 Years of Decipherment — Each Generation Reading It Anew with New Tools

1959: “An Ancient Greek Computer”

The silence was broken by British historian of science Derek de Solla Price. In 1959, he published a paper in Scientific American titled “An Ancient Greek Computer,” arguing that the mechanism was a sophisticated astronomical calculator.^[4] The academic community was sceptical, but Price did not stop his research.

In 1971, Price collaborated with Greek nuclear physicist Charalampos Karakalos to X-ray the mechanism and examine its internal structure.^[5] The results formed the basis of the 1974 paper Gears from the Greeks: The Antikythera Mechanism — A Calendar Computer from ca. 80 B.C., which identified 27 gears, analysed the Metonic cycle, and presented the first systematic decipherment.^[5] Price concluded that the machine was both a calendar calculator and a device for predicting celestial movements. His research established the mechanism as a legitimate subject of scholarly inquiry.

2005–2006: What a CT Scanner Revealed

The decisive breakthrough came in 2005, when researchers from Britain, Greece, and the United States joined forces to establish the Antikythera Mechanism Research Project (AMRP). The British company X-Tek Systems transported an eight-tonne microfocus CT scanner to the National Archaeological Museum in Athens. This equipment used X-rays to scan the interior of the object layer by layer and reconstruct three-dimensional images, enabling researchers to examine the internal structure of all 82 surviving fragments. The HP research team applied Reflectance Transformation Imaging (RTI)—a technique that takes photographs from many different angles to reveal subtle surface contours and inscriptions.^[6]

The results, published in Nature in 2006 by Tony Freeth and colleagues, astonished the academic world. More than 30 gears were confirmed, and over 2,000 characters of inscription were deciphered. Most importantly, the research revealed that the mechanism was designed to predict solar and lunar eclipses—a function that had not previously been known.^[6]

2008: The Olympic Dial

Further analysis continued after the 2006 study. A follow-up paper published in Nature in 2008 showed that one of the dials on the back of the mechanism tracked a four-year cycle—specifically the schedule of the major Panhellenic games of the ancient Greek world, including the Olympic, Pythian, Isthmian, and Nemean games.^[7] It was precision mechanics in the service of everyday life.

2021: The Dance of the Planets

In 2021, Tony Freeth of the UCL team published a reconstruction of the front mechanism in Scientific Reports.^[8] The study proposed that the motions of five planets—including Venus (462-year cycle) and Saturn (442-year cycle)—were represented through a concentric gear structure. The argument was that Babylonian astronomical period calculations, mathematics from Plato’s Academy, and Greek astronomical theory had all been integrated into a single bronze machine.

Inside the Machine — What Functions Did It Contain?

The Antikythera Mechanism was a bronze machine housed in a wooden case. Today, 82 fragments survive, containing more than 30 confirmed bronze gears, some with over 200 teeth. It is estimated that the original device contained more gears than have been recovered.

The machine worked as follows: a date was entered via a hand crank on the front, and the internal gears meshed together to move the pointers on each dial to calculated positions. The functions implemented were as follows.

Metonic Cycle: After 19 years, the dates of the solar and lunar calendars nearly coincide. 235 synodic months (the period for the moon to return to the same phase) correspond almost exactly to 19 years. This cycle, discovered by ancient Greek and Babylonian astronomers, was a key unit of calculation for the mechanism.^[9]

Callippic Cycle: A 76-year cycle—four times the Metonic cycle—used to reduce accumulated errors.

Saros Cycle: Solar and lunar eclipses repeat in a similar pattern approximately every 18 years and 11 days. This corresponds exactly to 223 synodic months. The spiral dial on the back of the mechanism was divided into 223 cells according to this cycle, allowing the prediction of dates of future eclipses.^[6]

Exeligmos Cycle: A 54-year cycle—three times the Saros cycle—used to correct for the regional visibility of solar eclipses.

Olympic Dial: Tracked the schedule of the Panhellenic games, which recurred on a four-year cycle.

All of these functions were implemented simultaneously within a single machine, through the precise design of gear ratios alone. There is also a hypothesis that a differential gear mechanism—which can calculate the difference in speed between two rotating axes—was used to correct for the irregular motion of the moon. Because the moon orbits the earth in an ellipse rather than a perfect circle, its speed is not constant. To replicate this mechanically requires a sophisticated mechanism using a pin and slot; researchers have found traces of exactly this inside the device.^[6]

Who Made It, Where, and When?

The mechanism is estimated to have been manufactured between 150 and 60 BC. Since the shipwreck has been dated to around 70–60 BC, it must have been made before that time.^[2] Clues include the calendar inscribed on the mechanism, which resembles the calendar used in Epirus (northwestern Greece), and the presence of month names associated with Corinthian colonies.

Three principal hypotheses compete as to the place of manufacture. The first is Rhodes. Rhodes was then the astronomical centre of the Mediterranean world. The renowned ancient astronomer Hipparchus worked there, and his theories on lunar motion appear to be reflected in the mechanism. The Roman statesman and philosopher Cicero also wrote, in De re publica, that during his studies in Rhodes as a young man he had seen a similar device.^[10] The second hypothesis points to the sphere of Corinthian colonies, based on the calendar names and letter forms. The third is the Pergamon hypothesis.

Among the specific candidates for its maker are Hipparchus himself, who was the first to describe the irregular motion of the moon mathematically; the Stoic philosopher and astronomer Posidonius; and the craftsmen working in the tradition of Archimedes of Syracuse. Cicero’s De re publica describes how General Marcellus took back to Rome an orrery attributed to Archimedes after the capture of Syracuse in 212 BC.^[10] Whether the Antikythera Mechanism was a descendant of that tradition remains an open question.

The Mystery of Lost Technology — Why Did It Take 1,400 Years?

The most uncomfortable question surrounding the Antikythera Mechanism is this: why did such a sophisticated gear device disappear after the first century BC? Comparable mechanical instruments in Europe—mechanical clocks—do not appear until the fourteenth century. That is a gap of more than 1,300 years.^[11]

Several hypotheses have been proposed. The simplest is technological isolation: the craftsmen capable of manufacturing gears to this level of precision were few in number, and when they disappeared, so did the knowledge. The technology may have been severed through the collapse of a particular school or workshop.

Another perspective is Roman pragmatism. Roman civilisation invested enormous energy in architecture, civil engineering, and military technology, but relatively little in precision astronomical instruments. Astronomical prediction was already perfectly achievable through the mathematical methods of the Babylonian tradition, which may have reduced the incentive to build expensive gear mechanisms.

There is also evidence that some of the knowledge was transmitted to the Islamic world. During the Abbasid period of the eighth and ninth centuries, Greek texts were translated into Arabic, and Islamic scholars produced sophisticated astronomical instruments. The eleventh-century Persian scholar al-Biruni described a “Moon box” that used gears to represent lunar phases, and the thirteenth-century mechanical engineer al-Jazari designed intricate automatic devices.^[11] Yet none of these approached the complexity of the Antikythera Mechanism. Whether the gear-making techniques themselves were lost, or whether there was simply no perceived need to build such a machine, remains unclear.

Is the Label “First Computer” Justified?

The Antikythera Mechanism is often called “the world’s first computer.” The label carries both admiration and controversy.

The argument in favour runs as follows: the machine takes an input value (a date) and calculates output values (positions of celestial bodies, eclipse predictions, festival schedules). In translating complex mathematical relationships into physical gear ratios, it fits the definition of an analogue computing device. Above all, it does not merely measure—it performs calculation.^[8]

The objection is that the machine lacks programmability, which is a core feature of a computer. It can only repeat the design embedded at the time of its creation; it cannot accept new algorithms. Its nature as a single-purpose machine is also noted.

Yet this debate may itself be aimed in the wrong direction. What matters is not whether the mechanism fits the modern category of “computer.” What matters is that someone in the first century BC precisely meshed more than 30 bronze gears to contain the laws of the cosmos in the palm of a hand. The fact that this was possible fundamentally unsettles our assumptions about ancient technology.

Former London Science Museum curator Michael Wright spent 35 years building a physical working reconstruction of the mechanism. In 2010, Apple engineer Andrew Carol attracted widespread attention by replicating the device’s key functions using only Lego Technic parts.^[12] In 2023, the mechanism became a central plot device in Indiana Jones and the Dial of Destiny, bringing it back to broad public attention.

The Excavation Is Not Over

Remarkably, the excavation of the Antikythera shipwreck is not yet complete. Since 2012, the Woods Hole Oceanographic Institution and Greek authorities have been conducting a joint re-excavation of the wreck, and in 2017 additional skeletal remains and artefacts were discovered.^[1] Sections of the seabed, forty-five metres down, have yet to be excavated. Research continues, with the hope that another device like the Antikythera Mechanism may yet lie within.

More than 120 years have passed since Stais spotted gear teeth in that corroded bronze lump in 1902. Over that time, every cutting-edge technology of its era—X-rays, CT scanners, RTI imaging—has been brought to bear, slowly peeling back the secrets of this machine. Yet who made it, where, and why the technology disappeared remain questions without complete answers. What the mechanism communicates most clearly is how fragile the assumptions we have long held about the intellectual capabilities of the ancient world truly were.