For years the transmission conversation has suffered from a mismatch of scale. The load-growth forecasts arrived in tens of gigawatts, the interconnection queues in thousands of projects, and the transmission plans in increments that always seemed a step behind both. On October 28, 2026, American Electric Power closed some of that gap with a single announcement: a $100 billion five-year capital plan through 2031, of which approximately $37 billion is earmarked for transmission. It is one of the largest transmission commitments attached to the load-growth story, and it is worth taking seriously precisely because it is attached to contracted demand rather than to a forecast.
The announcement came with AEP’s third-quarter 2026 results, and the quarter itself helps explain the plan’s size. AEP reported third-quarter operating earnings of $1.216 billion, or $2.23 per share, up from $963 million, or $1.80 per share, a year earlier, on revenue of $6.832 billion, up from $6.011 billion. GAAP earnings were $1.187 billion, or $2.18 per share. The company now expects full-year 2026 operating earnings at or above the high end of its $6.25 to $6.55 per-share guidance range, and it introduced 2027 guidance of $6.75 to $7.05 per share and 2028 guidance of $7.40 to $7.70.
The load number behind the capital number
AEP signed 15 gigawatts of new load commitments during the third quarter alone. That took its incremental contracted load to 84 GW expected through 2031, up from 69 GW disclosed on the second-quarter call. The company counts a project in that total only when it is backed by a fully executed electric service agreement or a letter of agreement, a discipline that matters: much of the load-growth debate has been clouded by speculative announcements that never become steel in the ground or wire in the air. AEP’s figure, by its own stated rule, excludes those.
Not all 84 GW carries the same weight. About 35 GW is backed by firm take-or-pay agreements, in which customers commit to pay whether or not they ultimately take the full amount of power. Another large block sits in Texas, where AEP says fully executed letters of agreement complying with Senate Bill 6 support 46 GW of contracted ERCOT load. The distinction matters: it separates demand that has posted collateral from demand that has merely expressed interest. AEP’s management expects demand on its system to more than triple by 2031, but the contracted figure — and especially the take-or-pay subset — is the part of that expectation a planner can actually build against.
The plan covers 2027 through 2031 and, at approximately $100.4 billion in the company’s investor presentation, represents roughly a 30 percent increase over the prior $78 billion five-year plan. AEP expects the spending to lift its rate base from $87 billion in 2026 to $162 billion by 2031, a compound annual growth rate above 13 percent, and it has raised its expected annual operating earnings growth rate to 9 to 11 percent for 2027 through 2031, up from 7 to 9 percent. Those are company expectations, not outcomes, and they depend on the load arriving, the projects being approved, and the capital being deployed on schedule. But the direction is unambiguous: AEP is planning as though the load-growth era is a construction program, not a forecasting exercise.
What $37 billion of transmission actually buys
Transmission is the largest single slice of the plan, at approximately $37 billion, ahead of about $36 billion for generation and about $22 billion for distribution, with the remainder in corporate and other investments. AEP is already the nation’s largest owner and operator of transmission, with roughly 40,000 line miles, and it expects its transmission rate base to nearly double, approaching $70 billion by 2031. That is the figure that should anchor the grid conversation, because it translates an abstract need — more transfer capability — into a rate-based asset base that regulators, customers, and investors can all see.
In physical terms, the plan leans heavily on the backbone AEP knows best: its 765-kilovolt extra-high-voltage system, the highest-voltage network in the country and the workhorse for moving large blocks of power over long distances. According to the company’s investor presentation, AEP plans to nearly triple the mileage of its 765-kV system over the next five years, with projects under development or construction across PJM, ERCOT, SPP, and MISO. The named example is the Piketon project in Ohio, a 765-kV line backed by a 20-year offtake agreement with SB Energy, with construction expected to begin in the first quarter of 2027, subject to required filings and approvals.
The Piketon structure is itself a signal about how transmission is getting financed in this era. Rather than waiting for a regional planning process to allocate costs across beneficiaries — the fight this desk’s “Transmission Bottleneck” explainer identified as the most litigated question in the field — the project pairs a specific line with a specific contracted customer willing to underwrite it. That anchor-tenant model does not replace regional planning, and it cannot by itself build a meshed network. But it does what regional planning has repeatedly failed to do: it converts a line from a contested allocation into a financed project with a start date.
The generation and distribution slices matter to the transmission story, too, because they define what the wires must carry and where. The $36 billion generation program includes a $7 billion increase in regulated natural gas generation versus the prior plan, reflecting contracted demand growing faster than the output of the existing fleet and the need for dispatchable capacity alongside new load. AEP has also completed commercial agreements for its Wyoming fuel-cell project, backed by a 20-year power purchase agreement with Google, with construction begun and full service targeted for December 2028. The $22 billion distribution program focuses on modernization, reliability, resilience, and infrastructure replacement.
Why AEP thinks it can execute at this scale
A capital plan is only as credible as its recovery mechanism, and here AEP’s pitch is unusually specific. The company expects roughly 90 percent of its planned utility investment to be recovered through mechanisms designed to reduce regulatory lag — formula rates, riders, trackers, and forward test-year rate cases. Regulatory lag is the quiet killer of utility construction programs: when a company spends billions and waits years for rates to reflect it, the financing cost compounds and the incentive to build erodes. A plan in which nine-tenths of the spending recovers on a shorter fuse is a plan whose economics do not depend on heroic assumptions about future rate cases.
The affordability argument runs in parallel. Fully executed take-or-pay agreements in AEP’s vertically integrated states are now expected to generate up to $21 billion in fixed-cost offsets for existing residential customers over the life of the contracts, up from a previous estimate of $16 billion. Chairman, president, and chief executive officer Bill Fehrman framed the logic directly: new large-load customers should bear the incremental system costs they create, and spreading fixed costs across a broader customer base can benefit existing customers. AEP points to a second proposed rate reduction in Indiana as an early example. Whether that logic holds at scale is the central political question of the load-growth era, and the $21 billion figure is an estimate over contract lifetimes, not cash in hand.
There is also a Texas-sized caveat that cuts in the company’s favor and against any simple reading of the plan. Management has characterized Texas as a source of upside beyond the current plan, because much of the infrastructure needed for its contracted load there is not yet included in the $100 billion figure. The timing and scope of that upside depend on ERCOT’s Batch Zero review and broader transmission-planning processes, which are still working through how the state’s share of the load wave gets served. The $100 billion plan may thus understate the eventual build if Texas load materializes at the contracted scale — or absorb the disappointment if it does not.
What still constrains it
This is where the transmission bottleneck reasserts itself, because a capital plan is not a construction schedule. The first constraint is permitting. As this desk has documented, a major interstate line typically requires approvals from every state it crosses, each on its own timeline, and a single state’s denial can kill a project. AEP’s footprint spans 11 states and four regional operators, which multiplies both the opportunity and the veto points. The company’s 765-kV expansion will test whether the anchor-tenant model and contracted demand can move lines through state processes faster than regional planning alone has managed — and the plan’s five-year window leaves little room for the seven-to-twelve-year timelines large interstate lines have historically consumed.
The second constraint is physical, and it has nothing to do with politics. Large power transformers — the custom-built units at substations — have carried lead times stretching past two years, with a thin domestic manufacturing base and global competition for slots, and the specialized lineworker workforce is aging faster than apprenticeship pipelines replace it. AEP says its generation investments are supported by secured turbine slots and construction partnerships. The transmission supply chain — transformers, conductors, breakers, crews — must still scale at a pace the industry has not previously demonstrated, and no capital plan can conjure a transformer that has not been manufactured.
The third constraint is cost allocation, which the anchor-tenant model sidesteps for individual projects but does not resolve for the system. Not every line AEP needs will have an SB Energy or a Google willing to underwrite it. The backbone lines whose benefits are diffuse — lower congestion across many states, shared insurance against regional shortfalls — still require someone to agree to pay, and the beneficiary-pays fights that have stranded regional plans before will not disappear because one company’s plan got bigger. FERC’s long-term planning rule requires planners to identify beneficial lines and states to agree in advance on cost-allocation methods, but identifying a line is not building it, and the authority gap this desk has described remains.
The fourth constraint is the load itself. Contracted load is not delivered load. Customers sign agreements, then delay projects, then resize them, then sometimes walk away — and the take-or-pay structure mitigates the financial damage without eliminating the planning damage, because a transmission line built for a customer who never arrives is still a line whose costs someone must carry. AEP’s discipline of counting only executed agreements is a genuine safeguard, but 84 GW is an enormous figure — roughly two and a half times the company’s existing generating capacity of about 34,000 MW — and the history of load forecasting counsels humility about any number that large. The plan’s credibility will be tested not by the next earnings call but by the pace at which contracted gigawatts become metered gigawatts.
What to watch
Three markers will tell whether this plan is a turning point or a press release with a long tail. The first is Piketon: whether construction actually begins in the first quarter of 2027 as expected, and whether its contracted-customer model becomes a template or remains a one-off. The second is Texas: what ERCOT’s Batch Zero process and the state’s transmission planning produce, and whether the upside management describes enters a future capital plan or stalls in process. The third is the simplest and the hardest — the quarterly cadence of contracted load converting into energized load, and of transmission rate base actually approaching the $70 billion the company projects.
There is a broader context worth naming. October 2026 has already shown what happens when the grid’s margins thin: PJM operated under a federal emergency order during a warm October week, with demand response dispatched and maintenance recalled, because shoulder-season load collided with the maintenance season. That episode was a reminder that adequacy is no longer only a summer-peak question, and that transmission deliverability — getting power to the constrained pocket at the stressed hour — matters as much as total megawatts. AEP’s plan is, in part, a response to that lesson: build the wires so the next tight week has somewhere to draw from.
None of this makes AEP’s plan a sure thing, and nothing in it should be read as a recommendation about the company’s shares or any other security. Capital plans are statements of intent backed by contracts and subject to regulators, supply chains, and customers who have not yet plugged anything in. What the plan does establish is a new floor for seriousness in the transmission debate. For a decade, the grid’s problem was described as a shortage of ambition: everyone agreed more transmission was needed, and no one committed money at the scale the need implied. AEP has now committed $37 billion to transmission alone, inside a $100 billion program, tied to 84 GW of contracted load. The bottleneck has not been solved. But for the first time in the load-growth era, the transmission buildout has a number big enough to argue with — and the argument, from here on, will be about execution rather than intent. That is progress, of a kind the grid has not seen in a long time.