The Biggest AI Companies Stopped Talking About Spending. Now They Talk About Speed to Power.
On their latest earnings calls, Microsoft, Alphabet, and Meta shifted the conversation from how much they are spending to how fast they can get their buildings powered and earning, a concept they call time-to-energy. That shift is the clearest signal yet that access to power, not capital, is the binding constraint on growth. The same logic applies to any building: the fastest path to revenue is increasingly the one that does not wait years for the grid, and owners can borrow the playbook.
By Keith Reynolds | Publisher & Editor, ChargedUp!
A Telling Change in Language
For years, the biggest technology companies competed on how much they would spend to build artificial intelligence infrastructure. In their most recent quarterly earnings calls, Microsoft, Alphabet, and Meta changed the subject. They emphasized how quickly their data center campuses can be energized and turned into revenue-generating computing, a focus industry observers labeled time-to-energy. The pivot from capital spending to speed of energization is subtle in wording and enormous in meaning. It signals that these companies can raise money faster than they can get power, and that power, not cash, now sets the pace of growth.
When the most capital-rich companies on earth start measuring themselves by how fast they can plug in, the constraint has moved. Money is abundant. Power is scarce. And the race is no longer to announce the biggest spending number but to be first to a working, powered building.
Why a Property Owner Should Care
This is not only a hyperscaler story, because the constraint they are naming is the same one every developer faces, scaled up. Connecting a large new building to the grid can take years, as this column examined recently, a wait driven by the same demand surge and connection backlog straining the whole system. A finished building that sits waiting for power earns nothing while its debt keeps accruing. The hyperscalers have simply put a name to a cost every owner bears: the gap between when a building is ready and when its power is.
The lesson they are acting on is one where a commercial owner can borrow directly. If power is the thing that gates revenue, then controlling power is controlling revenue. A building that can generate a share of its own electricity does not wait in the same line. It can open and start earning in the months an equipment installation takes rather than the years a grid connection can require. Speed to power is speed to revenue, and in a market where the grid cannot keep up, that speed has become one of the most valuable features a building can have.
How Hyperscalers Are Solving It, and What Scales Down
The biggest players are pairing their buildings with dedicated power, on-site or contracted, rather than waiting for the grid. Meta's planned 1-gigawatt data center in Alberta is being built alongside its own power arrangements, part of a broad pattern of self-supply. Across the industry, developers are turning to onsite generation and storage to bypass the connection queue entirely, because the wait has grown longer than the buildings take to build.
Most of that architecture scales down. A commercial owner will not build a gigawatt of generation, but the same principle, make and store a share of your own power so you depend less on a slow grid, works at the scale of a warehouse, a shopping center, or an office park. Onsite solar, a battery, and the controls that manage them let a building lean on the grid less, open sooner, and hold steadier costs. The hyperscalers have the resources to prove the model at scale. The model itself is available to any owner.
The Bottom Line
When Microsoft, Alphabet, and Meta stop bragging about spending and start racing on time-to-energy, they are telling the whole market where the real bottleneck is. Power has become the scarce input that decides how fast growth can happen, for a data center and for any building that needs electricity to earn. An owner cannot fix the grid, but an owner can refuse to be entirely dependent on it. The building that controls a share of its own power gets to market faster, holds its value better, and turns the industry's biggest constraint into its own advantage. Speed to power is the new competitive edge, and it is available well below hyperscaler scale.
