Skip to content

Invest1 publisher3 min readPublished

NextEra's 15 GW data-center hub goal depends on a few multi-gigawatt deals

NextEra Energy could meet its 15 GW data-center hub goal for 2035 with eight of the roughly 30 sites it is discussing, if each takes 2 GW. Hyperscalers supplied about 30% of recent additions to its 35.1 GW backlog, so a handful of hub contracts will set how far that share rises.

The Investor · Invest desk

Illustration accompanying NextEra's 15 GW data-center hub goal depends on a few multi-gigawatt deals

What happened

  • NextEra Energy reported a backlog of 35.1 GW of renewable energy and storage projects as of the second quarter of 2026.
  • The second quarter added 3.6 GW to that backlog, the company's second-largest quarterly addition on record.
  • About 30% of recent backlog additions came directly from hyperscale clients, a group that includes partnerships with Google and Meta.
  • NextEra Energy Resources' data-center hub initiative targets at least 15 GW of dedicated new generation by 2035, and more than 30 GW in an upside case.
  • Management said it is in discussions on about 30 potential hub locations.

Compiled by The InvestorSomething wrong?How this is made

Why it matters

  • decision Anyone valuing NextEra on AI demand is mostly valuing how many hub talks become contracts, so each quarter's hyperscaler additions matter more than the headline backlog total.
  • capability Because clients now ask for multi-gigawatt blocks, one or two hub signings can move the program by several gigawatts, and a short list of deals decides whether NextEra reaches its base case.
  • exposure NextEra's plan to serve round-the-clock data-center load leans on batteries, so battery prices and delivery schedules become a cost risk inside the hub program.

Of the capacity figures NextEra published, only the 35.1 GW is backlog [1]. The hub numbers are targets, and the site list is a set of talks [4][5]. The regulated utility, Florida Power & Light, has more than 20 GW of large-load projects under consideration [9]. That is about 57% of the size of the entire backlog [6].

Hyperscalers are a minority of what is being added [6]. The source does not define "recent". If the 30% share applies to the second quarter's 3.6 GW [2], hyperscalers supplied roughly 1.1 GW and other buyers about 2.5 GW [1].

Against that run rate the hub program is modest. Spread over the roughly nine years from 2026 to 2035, the 15 GW base case comes to about 1.7 GW a year [2]. The 30 GW upside works out to about 3.3 GW a year [3]. NextEra added 3.6 GW across its whole business in the second quarter alone [2].

Request sizes pull the other way. According to CEO John Ketchum, client requests have moved from traditional project sizes to multi-gigawatt capacities, and clients want them delivered faster [7]. That puts the program's result in a small number of contracts. If a hub averages 2 GW, the base case needs eight signings, a little over a quarter of the roughly 30 sites in talks [4].

The generation plan shows where the capital is going. Through 2032 NextEra projects up to 41.5 GW of solar and 43 GW of battery storage, against 4 to 8 GW of new natural gas [8]. Even at the top of its range, gas is about 9% of 92.5 GW [5], so the company is not planning a gas fleet to carry data-center load. Nuclear appears only as openness to recontracting and expansion [10].

Whether power, and not chips, limits the AI build-out is a question about a whole industry. This is one developer's pipeline, and it supports a narrower point. Buyers want bigger blocks of power, sooner [7], and they are supplying a sizeable minority of NextEra's new backlog [6].

The hub talks could convert near the base case, leaving AI a minority of NextEra's growth. Multi-gigawatt deals could close fast enough to make the upside the working plan. Or the talks could stall while the backlog keeps filling from other buyers. I think the 15 GW base case is the figure to hold management to, because the upside needs the site pipeline both to grow and to convert [5]. The view is wrong if the hyperscaler share of quarterly additions climbs well above 30% [6]. That would mean data centers are becoming the main source of the backlog.

What to watch

  • Whether the hub site list grows from about 30 to the planned 40 by the end of 2026, and how many of those sites move into backlog.
  • Whether any of Florida Power & Light's more than 20 GW of large-load projects move from consideration into contracted capacity.
  • Whether NextEra's stated openness to nuclear recontracting and expansion turns into a named deal.
Loading claim ledger
Loading source directory links
Loading share composer
Loading topic controls
Loading related stories