Infrastructure · Trend

Compute scaling

Training compute keeps climbing roughly four-fold a year, anchoring the cost of frontier work.

Trend strength 9/10

Momentum +1/q

Confidence high

Status stable

Forecast horizon

Gigawatt-scale clusters concentrate capability; capital and power become the real gatekeepers.

Connections

Connections · 3

How this node ties into the rest of the map, and the evidence behind each link.

Larger clusters bind on power and cooling.

Scaling depends on accelerator supply.

More compute underwrites the reasoning paradigm.

Signal sources

Dated facts from primary sources in this direction.

Training compute for frontier AI models grew ~4–5× per year from 2010 to 2024 — a trend Epoch projects to continue toward 2030.

Global data-centre electricity use is projected to roughly double from ~485 TWh (2025) to ~950 TWh by 2030 — about 3% of world electricity.

The price to query a GPT-3.5-level model fell from $20.00 to $0.07 per million tokens between Nov 2022 and Oct 2024 — a 280× drop in ~18 months.

DeepSeek disclosed its open-weight R1 reasoning model's final training run cost just $294,000; R1 was released under open weights on 20 Jan 2025.