New record! What impact will a 15-year ultra-long-term power contract for a city-level population of 200k have on the upstream and downstream of AI data centers?

Applied Digital Corporation (apld) announced today that it has signed a 300MW, 15-year long-term contract worth approximately $7.5 billion with a U.S. investment-grade hyperscaler.
The stock price surged nearly 20% in response.
The 300MW scale is already close to the load of a city with a population of 200k, and the 15-year term clearly exceeds traditional data center contracts; moreover, the capacity-locked model is different from GPU-based or hourly-based compute leasing.
This has set a new record for infrastructure-level AI contracts based on “mega-scale + ultra-long cycle + clear compute purpose.”
Behind this is a change in industry attributes.
Current data centers are essentially still IT services, although there have been some contracts lasting up to 5 years, but expansion is still demand-driven, with resources that can be migrated and replaced;
Now, they are gradually becoming infrastructure assets, beginning to lock in supply through long-term contracts similar to power PPA or energy agreements lasting 15 years.
Compute power is no longer a resource that can be purchased at any time; it requires advance planning and pre-allocation of production capacity.
The fundamental reason for this change is resource scarcity.
Like storage, chips, and optical modules, for hyperscalers, if they do not lock in resources early, they may not be able to obtain them in the future.
In the upcoming evolution, power assets will be re-priced and even redefined.
Having power, access capability, and the ability to expand power capacity to 100–300MW or even GW levels, along with network connectivity and development feasibility, will become attributes of new asset valuation.
Ultimately, all of this points to one fundamental shift: AI competition is moving from the model and algorithm layer to the physical resource layer.
Power, transformers, cooling, land, and networks are becoming new constraints.
Whoever controls these entry points will control future compute supply and thus gain longer-term pricing power.
This is the real significance of this 15-year contract.
Disclaimer: I hold the securities mentioned in this article; my views are biased and do not constitute investment advice. DYOR.