In the late Cretaceous, approximately 100million years ago and subsequent to the splitting off from Gondwana of conjoined Madagascar and India, the Indian Plate split from Madagascar. It began moving north, at about 20 centimetres (7.9 in) per year, and is believed to have begun colliding with Asia as early as 55million years ago, in the Eocene epoch of the Cenozoic. However, some authors suggest that the collision between India and Eurasia occurred much later, around 35million years ago. If the collision occurred between 55 and 50 Mya, the Indian Plate would have covered a distance of 3,000 to 2,000 kilometres (1,900 to 1,200 mi), moving faster than any other known plate. In 2012, paleomagnetic data from the Greater Himalaya was used to propose two collisions to reconcile the discrepancy between the amount of crustal shortening in the Himalaya (~1300 km) and the amount of convergence between India and Asia (~3600 km). These authors propose a continental fragment of northern Gondwana rifted from India, traveled northward, and initiated the "soft collision" between the Greater Himalaya and Asia at ~50 Ma. This was followed by the "hard collision" between India and Asia occurred at ~25 Ma. Subduction of the resulting ocean basin that formed between the Greater Himalayan fragment and India explains the apparent discrepancy between the crustal shortening estimates in the Himalaya and paleomagnetic data from India and Asia.
In 2007, German geologists suggested that the reason the Indian Plate moved so quickly is that it is only half as thick (100 kilometres or 62 miles) as the other plates which formerly constituted Gondwana. The mantle plume that once broke up Gondwana might also have melted the lower part of the Indian subcontinent, which allowed it to move both faster and further than the other parts. The remains of this plume today form the Marion, Kerguelen, and Réunion hotspots. As India moved north, it is possible that the thickness of the Indian plate degenerated further as it passed over the hotspots and magmatic extrusions associated with the Deccan and Rajmahal Traps. The massive amounts of volcanic gases released during the passage of the Indian Plate over the hotspots have been theorised to have played a role in the Cretaceous–Paleogene extinction event, generally held to be due to a large asteroid impact.
The Indian Plate is currently moving north-east at 5 centimetres (2.0 in) per year, while the Eurasian Plate is moving north at only 2 centimetres (0.79 in) per year. This is causing the Eurasian Plate to deform, and the Indian Plate to compress at a rate of 4 millimetres (0.16 in) per year.