Across the board, the message seems to be that it takes about 18 to 24 months to build any kind of data center, and that the bigger they are, the less likely they are to be completed on schedule.

Not graded on evidence

Outcome: Can’t be checked

An opinion, a recommendation, or too vague to test. Listed, not scored.

Evidence up to 2026-10-28. Checked 2026-10-31.

Why

The statement is hedged and framed as a general pattern. It begins "Across the board, the message seems to be fairly simple", so it reports the impression his own survey of projects gave him. It is not a firm claim about the industry. He does not say whether "build" means the first phase or the whole campus. He gives no size bands and no schedule baseline for "less likely to get completed on schedule." Testing it would mean adding measurement choices he did not make. The rules allow a plain reading only for firm language. As context only: on 2025-11-10, Epoch AI reported that for the large US AI sites it tracks that broke ground in the past three years, the time from starting construction to 1 GW of total facility power ranges from 1 to 3.6 years. The same note says xAI was projecting 12 months for Colossus 2. It also says the first gigawatt-scale sites were expected to come online in early 2026, so these figures include expected milestones, not only completed ones.

Caveats

• The Epoch AI context covers only the largest US AI data center campuses and one milestone, 1 GW of total facility power.
• The Epoch AI figures, dated 2025-11-10, include projected or expected milestones rather than only completed builds.
• No evidence in the packet measures schedule slippage by project size.

Evidence (1 item)

Claims like this aren’t scored, so there is no evidence to show.