Basic
4
788
2027
40%
chance

In "Situational Awareness: The Decade Ahead", Leopold Aschenbrenner predicts that the largest AI training clusters will consume 1GW of electricity in ~2026.

This market resolves YES if a training run of a single AI model consumes 1GW+ of power sustained through most of the training run. This power cost includes overhead to run the data center, such as cooling.

This is one of a series of markets on claims made in Leopold Aschenbrenner's Situational Awareness report(s):

Other markets about Leopold's predictions:

Get Ṁ600 play money
Sort by:

The training run has to complete before January 1st 2027 ?

Before January 1st 2027, we need to be confident that it's happening, that it has been consuming 1GW of power so far, and that it will continue to do so for the majority of the training run.

Ok, but not that it's necessarily finished?

That's right.

More related questions