AI Absorption Ledger / NEE

NextEra Energy

NEE · Q2 2026 · reported 2026-07-24 · revenue $7.53bn

Assessment

The quarter says less about AI than Q1. The call does not mention Rewire, its tools or the products for the utility industry; the CEO names artificial intelligence once, in a list of practices the pending merger with Dominion Energy would share. The 10-Q and the release repeat the risk that planned productivity increases through AI may not be realized. The operations channel stays described and unsized, and the products channel is not mentioned.

Large-load demand again fills the call and is absent from the 10-Q, and again no covered source names AI as its cause. FPL raised its large-load expectation for 2032 from GW to GW with no customer yet signed, Energy Resources added GW to its backlog with no hyperscaler share given, and Duane Arnold and the federal projects have not started. The facilities channels stay described and unsized, with the ceilings quoted: NEER revenue of and capital expenditures of in the quarter, capital and out of totals.

Other operations and maintenance rose by on the year on growth across the NEER businesses, while FPL’s line fell; no operating line is attributed to AI. No channel carries a size.

Paid for AI1 channel · 1 not sized

other

Capital spending on generation and grid built for data-centre and large-load customers, inside the capital plan

Not sized

AI named beside another cause, and only at the anchor: no data-centre or AI share of capital spending is stated, so no ballpark is built on a judgment share; the ceiling (the quarter’s capital expenditures, FPL’s capital spending and the capital per gigawatt of large load) is in the metrics.

described, no sizedisclosure: described· motive: exploratory· before LLMs: expanded· layer: facilities

Capital expenditures, independent power and other investments and nuclear fuel purchases were in the quarter, the first half less Q1, against a year earlier, a change of ; FPL’s part was approximately (claim c14). The CEO still expects roughly of capital per gigawatt of FPL large load, none yet signed (claim c8). Capital spending is left out of totals.

Evidence: 3 quotes, 8 figures, 2 confounds, 3 from before coverage

The part of FPL’s and Energy Resources’ capital spending that builds generation, storage, transmission and pipelines for data-centre and other large-load customers. The CEO says every gigawatt of large load under FPL’s tariff is expected to be equivalent to roughly a stated amount of capital, and that Energy Resources builds energy infrastructure for hyperscalers, who pay for it; the same capital lines also serve Florida’s population growth, utility and co-op customers, grid hardening, nuclear fuel, the gas business and transmission. The company names AI as a cause of data-centre demand only at the anchor, beside other causes, and no source states a data-centre share of capital, let alone an AI share. Under the joint-cause rule the channel is described and unsized, with no ballpark built on a judgment share; the ceiling is quoted instead (the quarter’s capital expenditures, FPL’s capital spending and FPL’s investment plan through 2032). Capitalized, so it is traced and shown and left out of flow totals; it reaches the income statement as depreciation and, at FPL, as the return regulators allow on rate base. The payees are equipment makers (solar panels, batteries, wind components and transformers secured through the end of the decade), EPC contractors and other suppliers. Funding is mixed: operating cash flow, debt, equity and, at Energy Resources, noncontrolling and tax-equity investors; for the projects selected under the U.S.-Japan framework the CFO says the company puts essentially no capital down.

Why this motive

Carried from Q1: capital committed ahead of large-load contracts that have mostly not started (claims c7 and c8).

Before LLMs: expanded

At the anchor capital spending already went chiefly to expanding FPL’s system and generation and to Energy Resources’ investments in independent power (claim nee-anchor-c5), in fiscal 2024, and the CEO named data centers for artificial intelligence applications among new sources of demand beside other causes (claims nee-anchor-c1 and nee-anchor-c4). Capital spending on generation predates LLMs; no quarter of the part serving data centres can be traced, so the level has no baseline. The size is the whole of an activity that existed before.

“And the tech industry is going to need data centers to support the expected cloud capacity demands that come with artificial intelligence applications.”
CEO, prepared remarks, earnings call, 2024-04-23
“That, of course, includes data centers, technology, AI-driven compute demand, but it is also manufacturing, the re-domestication of important industries in the U.S., and it is also oil and gas and chemicals companies looking to get lower cost energy solutions into their mix.”
Executive, qa, earnings call, 2024-04-23
“NEE's primary capital requirements are for expanding and enhancing FPL's electric system and generation facilities to continue to provide reliable service to meet customer electricity demands and for funding NEER's investments in independent power and other projects.”
Filing, mdna, 10-K periodic report, 2025-02-14

Figures

  • Capital expenditures, the quarter (six months less Q1) · 2026-CQ2
  • Capital expenditures, prior-year quarter (six months less Q1) · 2025-CQ2
  • Capital expenditures for the quarter, year-over-year change · 2026-CQ2
  • Capital expenditures, independent power and other investments and nuclear fuel purchases, six months · 2026-01-01..2026-06-30
  • Capital expenditures, independent power and other investments and nuclear fuel purchases, six months of the prior year · 2025-01-01..2025-06-30
  • FPL capital expenditures in the quarter, approximately · 2026-CQ2
  • Capital expected per gigawatt of large load under FPL’s tariff, roughly · as-of 2026-07-24
  • Base-case goal for new generation to serve large load by 2035, gigawatts · as-of 2026-07-24

What else could explain it

  • other: No Q2 source names AI as a cause of the growth the capital serves; the AI cause rests on the anchor alone (claims nee-anchor-c1 and nee-anchor-c4), beside other causes.
  • line composition: The capital line, in the quarter, also holds FPL’s spending for Florida’s growth, solar and storage, grid work, nuclear fuel, the Duane Arnold recommissioning, the gas business and transmission; no data-centre share is stated.

Quotes

“Initially, we expect every gigawatt of large load under FPL's approved tariff to be equivalent to roughly $2 billion of CapEx and to earn the same return on equity as other FPL investments.”
c8 · CEO, prepared remarks, earnings call, 2026-07-24
“We also have four origination channels feeding into our base case goal of securing 15 GW of new generation to serve large load by 2035.”
c11 · CEO, prepared remarks, earnings call, 2026-07-24
“FPL's capital expenditures were approximately $2.8 billion for the quarter, and we expect FPL's full-year capital investments to be between $12 billion and $13 billion.”
c14 · CFO, prepared remarks, earnings call, 2026-07-24

By quarter

  • Q1 2026described · described, no size · exploratory
  • Q2 2026described · described, no size · exploratory

Cost displaced by AI1 channel · 1 not sized

operations · expensive to verify

Rewire AI tools in plant operations, fleet performance and dispatch (Conduit, Generation Entitlement, Grid Composer)

Not sized

AI named beside another cause, and the money has not started: the quarter gives no tool, saving, rate or count, the only statement is the risk that planned productivity increases through AI may not be realized, and Q1 credited the expected savings to Rewire as a whole. No ballpark is built on a judgment share; the ceiling (other operations and maintenance) is in the metrics.

described, no sizedisclosure: described· motive: exploratory· before LLMs: relabelled

The call does not mention Rewire or its tools. The 10-Q and the release repeat that planned productivity increases through AI may not be realized (claims c2 and c3), and the CEO names artificial intelligence only in a list of practices the Dominion Energy merger would share (claim c1). FPL’s non-fuel O&M is more than better than the industry average per megawatt-hour (claim c19).

Evidence: 5 quotes, 7 figures, 2 confounds, 2 from before coverage

Rewire is described on the Q1 2026 call as a company-wide initiative to reimagine how the company works, paired with an enterprise-wide AI transformation, in partnership with Google Cloud. Its first tools are Conduit, an AI-powered tool to make the renewables workforce more efficient in the field; Generation Entitlement, which identifies abnormal equipment conditions; and Grid Composer, which uses AI to optimize unit commitment, power and fuel dispatch and maintenance scheduling. The work displaced is the company’s own field, engineering and dispatch work, so the counterparty is internal; dispatch and equipment-condition calls are checked by engineers and an error costs reliability, so verification is expensive. The CEO says the tools have the potential to drive significant savings for customers and gives no saving, rate or count. The savings are credited to Rewire as a whole, which pairs reimagining the work with AI, and the CEO credits FPL’s low cost to a relentless focus on technology, so AI is named beside another cause and nothing separates its part. The 10-Qs and releases say only that productivity increases the company plans to achieve through AI may not be realized. The partnership with Google Cloud names no money or term, so there is no vendor-bill channel; Google is on this ledger as GOOGL. The anchor already shows dozens of proprietary artificial intelligence tools used to operate the fleets, and the sources never say the new tools are LLMs, so under the tie-break the activity is read as renamed.

Why this motive

Carried from Q1: the quarter names no tool or result, and the 10-Q and release again say planned productivity increases through AI may not be realized (claims c2 and c3).

Before LLMs: relabelled

At the anchor the CEO said the company had dozens of proprietary artificial intelligence tools for analytical, real-time decision-making and used its tools to operate nearly all its renewable, storage and fossil fleets around the clock (claims nee-anchor-c7 and nee-anchor-c8). The work sits in operations and maintenance, in the first quarter of 2024 across every function. No line, price or rate is shown to move because of the Rewire tools, so the activity is read as renamed.

“Today, NextEra Energy captures 560 billion operational data points each day and has dozens of proprietary artificial intelligence tools to drive analytical, real-time decision-making.”
CEO, prepared remarks, earnings call, 2024-04-23
“And we use our tools to operate nearly all our renewable, storage, and fossil generation fleets around the clock from our headquarters in South Florida.”
CEO, prepared remarks, earnings call, 2024-04-23

Figures

  • Other operations and maintenance · 2026-CQ2
  • Other operations and maintenance, prior-year quarter · 2025-CQ2
  • Other operations and maintenance, increase from the prior-year quarter · 2026-CQ2
  • FPL other operations and maintenance · 2026-CQ2
  • FPL other operations and maintenance, prior-year quarter · 2025-CQ2
  • Increase in O&M expense at NEER, year over year · 2026-CQ2
  • FPL non-fuel O&M better than the industry average per megawatt-hour, more than · as-of 2026-07-24

What else could explain it

  • transformation program: Carried from Q1: the savings are expected from Rewire as a whole, which pairs reimagining the work with AI (claim c1). The CEO credits FPL’s low O&M to a highly efficient operating model, scale and sustained capital investment, without naming AI (claim c21).
  • operating leverage: Other operations and maintenance rose by on the year, which the 10-Q puts down to growth across the NEER businesses (claim c18); FPL’s line fell to from . No movement is attributed to AI.

Quotes

“•The productivity increases and competitive advantages NEE and FPL plan to achieve through the use of artificial intelligence (AI) technologies may not be realized and the use of and reliance on AI may present certain risks, both of which could materially adversely affect their business, financial condition, results of operations and prospects.”
c2 · Filing, risk factors, 10-Q periodic report, 2026-07-24
“planned productivity increases and competitive advantages through the use of artificial intelligence technologies may not be realized and the use of and reliance on artificial intelligence may present certain risks”
c3 · Filing, press release, 8-K earnings release, 2026-07-24
“Operating expenses – net for the three months ended June 30, 2026 increased $331 million primarily due to increases of $221 million in O&M expense and $57 million in fuel, purchased power and interchange expense. Operating expenses – net for the six months ended June 30, 2026 increased $672 million primarily due to increases of $432 million in O&M expense and $165 million in fuel, purchased power and interchange expense. The increases for both periods were primarily associated with growth across the NEER businesses.”
c18 · Filing, mdna, 10-Q periodic report, 2026-07-24
“FPL's non-fuel O&M is more than 70% better than the industry average on a dollar per megawatt- hour basis.”
c19 · CEO, prepared remarks, earnings call, 2026-07-24
“This performance is a direct result of a proven long-term strategy centered on sustained, disciplined capital investment, scale, a strong balance sheet and credit rating, a highly efficient operating model, and a diverse generation fleet.”
c21 · CEO, prepared remarks, earnings call, 2026-07-24

By quarter

  • Q1 2026described · described, no size · exploratory
  • Q2 2026described · described, no size · exploratory

Revenue arriving through AI2 channels · 2 not sized

customer cohort

Power and generation sold to hyperscalers and other data-centre loads (Energy Resources contracts, data-centre hubs, FPL large-load tariff)

Not sized

AI named beside another cause, and only at the anchor: no covered source names AI as a cause of data-centre demand and no data-centre amount is stated; FPL large load, Duane Arnold and the federal projects have not started (dates in the note). No ballpark is built on a judgment share; the ceiling (NEER revenue and its increase from new investments) is in the metrics.

described, no sizedisclosure: described· motive: exploratory· before LLMs: expanded· layer: facilities

The company raised its expectation for FPL large load by 2032 from GW to GW, still with GW of interest and GW in advanced discussions, and still expects to announce a first transaction by year-end (claims c5, c6, c7 and c20). Energy Resources added GW to its backlog, with no hyperscaler share given this quarter, and has potential hubs toward the GW base case (claims c10, c11 and c15). The state is described: the raised expectation and the gigawatt counts measure the channel’s volume, not AI’s part, as at Duke Energy.

Evidence: 13 quotes, 12 figures, 3 confounds, 5 from before coverage

Electricity, capacity and new generation the company sells or builds for data-centre load: long-term contracted renewables and storage that Energy Resources sells to hyperscalers directly and to utilities, co-ops and municipalities that serve them; the data-centre hub strategy (gas-fired and other generation built for large loads, including the projects selected under the U.S.-Japan framework, which the U.S. and Japan would own and Energy Resources would develop, build and operate for fees); the recommissioning of the Duane Arnold nuclear plant in a collaboration with Google, due back no later than Q1 2029; and large-load service at FPL under its approved tariff, where no customer had signed by the end of the covered quarters. Some of this money has started (Energy Resources already had generation in operation for technology customers at the anchor), most of it has not (the hubs, Duane Arnold, the federal projects and FPL large load). The company names AI as a cause of data-centre demand only at the anchor, and there beside manufacturing, re-domestication and oil, gas and chemicals; no covered call, release or 10-Q names AI as a cause of the demand, and the 10-Qs do not mention data centres at all. No source gives a data-centre amount, let alone AI’s part of one, so any amount would be a ceiling on the AI part, whose non-AI remainder is the enterprise and general cloud work the same data centres run. The only measures in coverage are the share of backlog additions driven by hyperscalers and gigawatt counts of interest, discussions and expected load; they measure the channel’s volume, not AI’s part, and separate nothing in dollars, so the channel reads described. The counterparty is mixed: hyperscalers (Google is named; MSFT, AMZN, GOOGL and META are on this ledger, where the same money once delivered is data-centre cost, not netted here) and the utilities, co-ops and municipalities that resell the power. The NVIDIA collaboration on flexing data-centre load names no money or term and is context, as is the Google collaboration on Duane Arnold, which the sources give no price or term for. The layer is facilities.

Why this motive

Carried from Q1: AI is still named as a cause only at the anchor, and the hub, federal, Duane Arnold and FPL large-load business has not started (claims c7, c12 and c13).

Before LLMs: expanded

At the anchor Energy Resources already sold contracted generation to technology customers, GW of it in operation, did business with the top five hyperscalers, and the CEO named data centers for artificial intelligence applications among new sources of demand (claims nee-anchor-c1, nee-anchor-c2, nee-anchor-c3 and nee-anchor-c4); the annual report counted expanded use of data centers among the drivers of customer usage (claim nee-anchor-c6). No data-centre revenue was given; operating revenues for the whole company were in the first quarter of 2024, context and not the channel’s level. In coverage new capacity is contracted for hyperscalers, so volume moved and the tag is expanded (methodology, facilities for data centres); the level has no baseline. The size is the whole of an activity that existed before.

“And the tech industry is going to need data centers to support the expected cloud capacity demands that come with artificial intelligence applications.”
CEO, prepared remarks, earnings call, 2024-04-23
“I mean, if you look at just our GW in operation, we have 3.5 GW in operation today. We have another 3GW, or close to 3.5 GW in our backlog, with technology, providers.”
CEO, qa, earnings call, 2024-04-23
“We do business with all the top, you know, what I would call the top five hyperscalers in this country.”
CEO, qa, earnings call, 2024-04-23
“That, of course, includes data centers, technology, AI-driven compute demand, but it is also manufacturing, the re-domestication of important industries in the U.S., and it is also oil and gas and chemicals companies looking to get lower cost energy solutions into their mix.”
Executive, qa, earnings call, 2024-04-23
“Customer growth and customer usage are affected by a number of factors outside the control of NEE and FPL, such as mandated energy efficiency measures, demand side management requirements, installation of distributed generation technologies and economic and demographic conditions, such as population changes, job and income growth, housing starts, new business formation, expanded use of data centers, inflation and the overall level of economic activity.”
Filing, risk factors, 10-K periodic report, 2025-02-14

Figures

  • Large load expected at FPL by 2032, gigawatts (raised in May) · as-of 2026-07-24
  • Large load expected at FPL by 2032 before the May update, gigawatts · as-of 2026-07-24
  • Large-load interest at FPL, gigawatts · as-of 2026-07-24
  • Large load at FPL in advanced discussions, gigawatts · as-of 2026-07-24
  • Potential data-centre hubs under discussion · as-of 2026-07-24
  • Renewables and storage added to the backlog in the quarter, gigawatts · 2026-CQ2
  • Base-case goal for new generation to serve large load by 2035, gigawatts · as-of 2026-07-24
  • Gas-fired generation in development in Texas and Pennsylvania under the U.S.-Japan framework, gigawatts, up to · as-of 2026-07-24
  • NEER segment operating revenues · 2026-CQ2
  • NEER segment operating revenues, prior-year quarter · 2025-CQ2
  • NEER segment operating revenues, year-over-year change · 2026-CQ2
  • NEER revenue increase from new investments, year over year · 2026-CQ2

What else could explain it

  • other: No Q2 call, release or 10-Q names AI as a cause of data-centre demand; the 10-Q does not mention data centres. The AI cause rests on the anchor alone (claims nee-anchor-c1 and nee-anchor-c4), where AI is named beside other causes. Nothing separates AI’s part.
  • line composition: Data centres also run enterprise and general cloud work, so any data-centre amount would be a ceiling on the AI part. NEER revenue of and its increase from new investments of hold every other customer too, and NEER revenue also moved on commodity hedges.
  • other: Most of the channel has not started: Duane Arnold returns no later than Q1 2029, FPL large load could begin as soon as 2028 with no customer signed, and the U.S.-Japan projects still await definitive agreements (claims c7, c12 and c13).

Quotes

“We continue to see very strong interest from hyperscalers and other large load customers that value speed to market, reliability, and competitive power pricing.”
c4 · CEO, prepared remarks, earnings call, 2026-07-24
“That is why in May, we updated our expectations from 6 GW to 8 GW of large load by 2032.”
c5 · CEO, prepared remarks, earnings call, 2026-07-24
“We have roughly 21 GW of large load interest at FPL.”
c6 · CEO, prepared remarks, earnings call, 2026-07-24
“Of that, we are in advanced discussions on 12 GW, a portion of which we believe we could begin serving as soon as 2028.”
c7 · CEO, prepared remarks, earnings call, 2026-07-24
“Major area of opportunity is large load demand and our data center hub strategy. Hyperscalers and other large load customers are increasingly focused on speed, certainty, and scalability.”
c9 · CEO, prepared remarks, earnings call, 2026-07-24
“In fact, we now have 30 potential hubs we are discussing with the market.”
c10 · CEO, prepared remarks, earnings call, 2026-07-24
“We also have four origination channels feeding into our base case goal of securing 15 GW of new generation to serve large load by 2035.”
c11 · CEO, prepared remarks, earnings call, 2026-07-24
“The recommissioning of our Duane Arnold nuclear plant is a perfect example of matching electric load with power generation. We remain on track to bring the plant back online no later than Q1 2029.”
c12 · CEO, prepared remarks, earnings call, 2026-07-24
“In March, the president approved the projects, which are drawn from our existing inventory of data center hubs.”
c13 · CEO, prepared remarks, earnings call, 2026-07-24
“For the quarter, Energy Resources added 3.6 GW of renewables and storage projects to its backlog, its second-largest quarter of additions coming on the heels of last quarter's record 4 GW.”
c15 · CEO, prepared remarks, earnings call, 2026-07-24
“When you think about bringing these hubs together and you think about speed to power and you think about long-term solutions as you ramp with your hyperscale customer as they grow into their compute capacity at a site, that speed to power solution, many times we're finding is starting with a renewable solution.”
c16 · CEO, qa, earnings call, 2026-07-24
“I think there are a lot of spots in the state of Florida, and we are working with those communities that are really interested and have good locations that'll be the perfect host for data centers.”
c17 · Executive, qa, earnings call, 2026-07-24
“FPL is advancing negotiations with large load customers and continues to expect to announce at least one large load transaction under FPL's tariff by the end of the year.”
c20 · CEO, prepared remarks, earnings call, 2026-07-24

By quarter

  • Q1 2026described · described, no size · exploratory
  • Q2 2026described · described, no size · exploratory

product revenue · expensive to verify

Rewire AI products delivered to the utility industry with Google

Not sized

The quarter’s sources are silent on Rewire products sold to the utility industry; no price, customer or revenue has been given.

inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: relabelled

No Q2 source mentions the Rewire products or their delivery to the utility industry with Google.

Evidence: 0 quotes, 2 from before coverage

On the Q1 2026 call the CEO said that Rewire also serves as the company’s AI product development platform, that partnering with Google it is delivering these products to the utility industry, and that its first Rewire products were brought to market in the quarter. No price, customer, revenue, term or revenue split with Google is given, and the examples named (Conduit, Generation Entitlement, Grid Composer) are described in the company’s own fleet. Under the rule for products with no price the channel is described and unsized, and no ballpark is built. The buyers would be other utilities, so the counterparty is enterprise; the products automate dispatch and equipment-condition work whose errors cost reliability, so verification is expensive. At the anchor the company already used its data, proprietary technology and tools in work for power customers and commercial and industrial customers, and the sources never say the products are LLMs, so under the tie-break the activity is read as renamed. Google is on this ledger as GOOGL. Q2 2026 does not mention the products.

Why this motive

Carried from Q1 (claim c3); silence is not evidence about motive.

Before LLMs: relabelled

At the anchor the company used its data and proprietary technology to help power customers balance supply and demand and used its tools with commercial and industrial customers to choose locations (claim nee-anchor-c9), drawing on dozens of proprietary artificial intelligence tools (claim nee-anchor-c7). No revenue from software or tools was given then or since, so no line exists to move; the activity is read as renamed.

“Today, NextEra Energy captures 560 billion operational data points each day and has dozens of proprietary artificial intelligence tools to drive analytical, real-time decision-making.”
CEO, prepared remarks, earnings call, 2024-04-23
“We are using our data and proprietary technology to help power customers balance supply and demand while keeping customer bills affordable. We also use our tools with commercial and industrial customers to identify the best locations based on their physical preferences and most important variables.”
CEO, prepared remarks, earnings call, 2024-04-23

By quarter

  • Q1 2026described · described, no size · exploratory
  • Q2 2026not mentioned · inscrutable · exploratory

Reported lines, year-over-year growth

Revenue +12.4%

Q2 2026. Growing slower than revenue: total operating expenses – net (+10.0%). A displaced cost shows up as a line that stays under the dashed revenue line. These are the audited lines, as first reported; nothing here is attributed to AI by the filing.

Other operations and maintenanceTotal operating expenses – netRevenue
0%5%10%15%20%25%Q1 2025Q2 2025Q3 2025Q4 2025Q1 2026Q2 2026Other operations and maintenanceRevenueTotal operating expenses – net
Reported values and filings
LineQ1 2025Q2 2025Q3 2025Q4 2025Q1 2026Q2 2026
Operating revenues$6.25bn$6.70bn$7.97bn$6.50bn$6.70bn$7.53bn
Other operations and maintenance$1.17bn$1.22bn$1.41bn$1.60bn$1.42bn$1.46bn
Total operating expenses – net$4.03bn$4.81bn$5.58bn$4.98bn$4.75bn$5.29bn