Washington's Nov. 7 Permitting Bill Prices a $615 Billion AI Buildout Before Any of It Is Energized

The substitute amendment is technical by design — clarifying and conforming edits to a bill introduced Sept. 30. What it deliberately leaves untouched is the part that governs power: FERC's expanded siting authority, the obligation on utilities to build grid upgrades, and the rule that large data centers pay their own incremental grid costs. Together those three clauses replace a decade of state-by-state rate-case fighting with a federal rule set, and the conversion of a queued project into a financed one is what changes.

The Senate votes on a cloture motion Nov. 9 and needs 60 votes. The text was locked Nov. 7, two days before the vote, because the four negotiators wanted the language fixed before the floor moved. The real risk is arithmetic: the Senate that votes is still Republican-controlled, while Democrats won both chambers on Nov. 3 and take office on Jan. 3, 2027. The bill the Senate would pass is the House vehicle, H.R. 2347, amended with the negotiated text — so the floor can finish the job, but the next Congress owns the result. ACORE said as much four days after the midterms: passing the bill now sets up the newly elected Congress to move beyond permitting reform. Treating Nov. 9 as a binary catalyst misses the point: the text is a negotiating baseline, not a vote.

Money is not the constraint. Amazon, Alphabet, Microsoft and Meta spent $615.4 billion on property and equipment in the twelve months to Sept. 30, 2026. In the September quarter alone they spent $201.9 billion — a run rate that annualises above $800 billion. Amazon's quarterly spend alone went from $44.2 billion in March to $77.3 billion in September. These dollars are committed, and they are already producing the specific hardware the grid has to absorb: turbines, transformers, switchgear, substations.

And yet the AI-power complex has de-rated through 2026. GE Vernova trades 35% below its 52-week high of $1,195. Constellation Energy's market value fell from $110.6 billion at the end of 2025 to $88.6 billion. A market that believed $615 billion of committed spending would not mark the suppliers down by a third. It is discounting something else: that most of the pipeline will never be energised, because it sits behind permits, interconnection studies and cost allocation rather than behind factories. The gap between committed spend and de-rated suppliers is the size of the permitting option.

Three numbers make the case. Nearly 50 GW of clean energy came online in 2025. In that same period, 85% of the projects sitting in interconnection queues were clean energy — down from 95% in earlier years — and 75% of them were cancelled outright because wait times and interconnection fees made them uneconomic. Meanwhile GE Vernova shipped 5 GW of gas equipment in the September quarter while converting 13 GW of slot reservations into firm orders, so firm backlog grew from 53 GW to 64 GW even as slot reservations fell from 63 GW to 58 GW. The industry is selling future capacity faster than the grid can absorb it, and the queue behind the book keeps lengthening.

This is the part that is easy to miss. A turbine slot is a contract; a permit is an option. GE Vernova's 122 GW of backlog and slot reservations is a contractual claim on capacity that exists only if the projects behind it clear review, interconnection and cost allocation. The permitting bill barely changes 2027 revenue. What it changes is the probability weight on the 2029 to 2031 tranche, which is precisely the tranche the market is refusing to pay for.

The 20 MW threshold is the entire political economy of the bill. A covered data center has to cover the full incremental cost of the generation, storage, transmission and distribution built for it, has to keep covering that cost if it exits early, and cannot pass it to anyone else. Every data center regardless of size pays published rates on existing grid assets under the \"and pricing\" model, with the revenue credited back to other customers' bills. States gain additional levers: they may charge above cost, allocate service through application windows, and require large loads to bring their own generation or accept limited service.

That is why the same text could be endorsed on the same day by a clean-energy supply chain that wants the grid built and a climate group that wants ratepayers protected. The bill does not resolve the conflict between data centers and households. It prices the conflict and assigns the invoice to the party creating the load. For an investor, that is the part to trade: the beneficiaries are identical whichever party ends up writing the final rule, because the economics sit in the cost-allocation mechanics, not in the preamble.

The ratepayer-protection floor covers incremental cost for covered loads that interconnect after enactment. Data centers already operating fall outside its scope, and the incremental cost is a one-time estimate fixed at the moment the load goes live. That combination creates a deadline effect that requires no bill at all — only a bill that is likely to pass. Columbia University's Center on Global Energy Policy frames the risk plainly: federal pressure only works if developers are willing to go over the heads of state regulators and utility partners, and only if the federal backstop is real enough to justify that commercial cost.

Two features of this chain decide who gets paid when. Equipment is sold long before construction begins — GE Vernova's $42.2 billion electrification backlog is booked years before a substation is energised — so the equipment layer monetises a permitting win first. Construction is paid on delivery: Quanta Services converts a permit into revenue only as crews energise the line, which is a permit-to-revenue lag measured in years. Owning the contractors is a bet that permitting wins arrive before the equipment does.

The 2025 reconciliation law already withdrew the federal tax credit from new wind and solar. To qualify for the clean electricity production and investment credits, those facilities must now begin construction by July 4, 2026 or be placed in service by Dec. 31, 2027. New clean generation that misses those dates gets no production or investment credit, and the $2 billion of Inflation Reduction Act transmission loans was rescinded outright. For developers, permitting is no longer one of several levers. It is the last federal supply-side lever that still works.

Read that table as a market that has already paid for the equipment shortage and refuses to pay for energisation. GE Vernova trades on 20 times trailing earnings with a $197 billion backlog after falling a third from its high. Quanta, the contractor who gets paid only when the line goes live, trades on 2.3 times sales. A framework that raises the conversion rate on the queue should lift the middle of the chain first — the contractors and the hardware makers with grid exposure — because a permit is worthless until someone builds to it, and the builder is the one currently valued like an afterthought.