In last week's post, I traced the AI energy debate back to one metal, and one number: the projects actually funded to build.

The harder question is the one that comes next — how long does everything behind them take to arrive?

The copper the world needs in 2035 has to be found, permitted, financed, and built before then. Most of it hasn't been found yet.

Of the projects that have been found, most are still in early-stage exploration. Of those that have advanced to feasibility, almost none have committed capital. Of those that have committed capital and begun construction, only 32 are active globally.

Pulse Intelligence Partner data across the full active copper asset universe shows the pipeline at every stage:

  • Grassroots and drilling: 1,910
  • Resource definition through scoping: 444
  • Pre-feasibility and feasibility: 112
  • Construction: 32
  • Producing: 253

The funnel narrows sharply and late.

1,910 companies are searching for copper. 32 are funded to build it. Primary copper assets by development stage: grassroots and drilling 1,910; resource definition to scoping 444; pre-feasibility and feasibility 112; construction 32; producing today 253.A discovery made today produces copper in 2044. 18 years on average from first drill to production, across 135 copper assets in the Pulse database. From a completed feasibility study: 4 years. From construction start: 2 years. The supply response can't be rushed.112 projects sit in feasibility. Almost none have committed capital. If every one advanced without delay they produce 2029 to 2032. S&P Global: 10Mt annual copper deficit by 2040. UBS: 400kt supply gap in 2026 alone. BHP: 6 times data-centre demand growth by 2050.This is not a data scarcity problem. It is a data accessibility problem. The data that decides which assets are real already exists — in NI 43-101 reports, JORC filings and technical studies across ASX, TSX and LSE — but assembling it by hand takes months. The next 12 to 18 months decide who is positioned.That analysis used to take months. It doesn't have to. The intelligence layer to make these calls — at the speed and coverage the market now demands — is a platform problem. That is what we are building. Pulse Intelligence: less searching, more strategising.
1 / 5
Copper × AI · The Supply Pipeline — swipe through all five. Source: Pulse platform, June 2026.

The average time from first drill to first production, calculated across 135 copper assets in the Pulse database, is 18 years. From a completed feasibility study: 4 years. From construction start: around 2 years.

The implication is straightforward. Projects making discoveries today do not produce copper until 2044 on average. The 112 projects currently in pre-feasibility and feasibility, assuming all of them advance without delay, begins producing sometime between 2029 and 2032. Some will. Many will not.

S&P Global models a 10-million-tonne annual copper deficit by 2040. UBS forecasts a 400,000-tonne gap in 2026 alone. AI data centre demand is projected to grow sixfold by 2050, according to BHP's long-range modelling.

The supply response required is not impossible. But it requires decisions to be made now, with the best available data, on which assets are real, which resource estimates hold, and which projects have a genuine pathway to production.

That analysis used to take months. It does not have to.

Now imagine being able to comp/benchmark any asset or company on live, validated data, see geology/exploration evolution in 3D in real time, chat to our AI about permitting trends in the region, or deep-dive on the leadership team's experience and track record? All in Pulse Intelligence Partner, and so much more.

Pulse Intelligence Partner

Less searching. More strategising.™

Pulse Intelligence

Less Searching. More Strategising.™

See the platform running on live, validated copper data — comps, 3D geology, permitting trends, and management track records in one place. Book a demo.

Request a demo → Take the Diagnostic