The material with the highest modulus of elasticity would deform the least and minimize the amount of energy lost to the surface.
Since we don't have large surfaces made of pure diamond, and thick sheets of exotic metals are also very expensive, the practical answer is steel plate.
But there is probably a more practical answer that beyond a certain point, it would be difficult to measure the difference.
The question is, where is that difference?
The ITTF has an approval process for testing tables (link 1). They test a complete table, and a 50 x 100 cm section of the top material. One would suspect that the uniformity of the ball rebound across the surface would be important, and we find this in the technical leaflet (link 2):
"2.1.3 The playing surface may be of any material and shall yield a uniform bounce of about 23 cm when a standard ball is dropped on to it from a height of 30 cm."
Concrete is listed as an acceptable material.
Later it is specified that the ITTF limits surfaces to wood and wood derivatives for tournament use, and tests the tabletop at 16 different points and then performs multiple regression analysis on the data.
"Both speed and spin of a ball are affected by the resilience of the playing surface and other properties that together define the bounce. This is measured by dropping an approved ball of average bounce on to the table; from a height of 300 mm, measured between the playing surface and the bottom of the ball, the ball must rebound to a height of 230-260 mm."
Measurement is described using audio and video techniques.
For further investigation, an inexpensive test fixture and measurement apparatus would generate a lot of interesting data in short order.