IRA phase-out in 2026: how OBBA, FEOC, and the tariff stack reprice US clean energy capex

From 369 billion to 1.7 trillion: why the rescore mattered politically #

The IRA's tax credit architecture was uncapped by design. The Joint Committee on Taxation's August 2022 score of 369 billion US dollars in net energy and climate spending over ten years assumed a deployment trajectory that, in retrospect, was conservative on three fronts: 45X advanced manufacturing uptake, 30D consumer EV credit transferability after the January 2024 IRS dealer transfer rule, and 45Y and 48E technology-neutral electricity credit absorption as utility-scale solar and storage accelerated. The CBO's January 2025 baseline rescore put the 2025 to 2034 cost between 1.4 and 1.7 trillion US dollars, with Goldman Sachs and the Penn Wharton Budget Model bracketing the upper end near 1.97 trillion if 2024 announcement velocity persisted.

That rescore reframed the political economics. A 1.7 trillion dollar program, with roughly 64 percent of announced manufacturing investment landing in Republican congressional districts, became a fiscal target that House Ways and Means leadership could not credibly leave untouched while pursuing the 4.5 trillion dollar 2017 tax cut extension. The OBBA clean energy rollback, scored by JCT in May 2025 at approximately 511 billion US dollars in deficit reduction over ten years, paid for roughly 11 percent of the tax extension. That arithmetic, not the climate debate, explains the structure of the phase-out: surgical cuts to credits where Republican manufacturing constituencies were thin (30D EVs, residential 25C and 25D, technology-neutral electricity) and preservation of credits where they were thick (45X batteries and solar, 45Q carbon capture, 45V hydrogen).

What OBBA actually changed, credit by credit #

The phase-out is uneven and the asymmetry is the investment signal. Section 30D, the 7,500 dollar new clean vehicle credit, terminates for vehicles placed in service after December 31, 2025, eliminating roughly 15 to 18 percent of effective EV pricing power against internal combustion comparables in the 35,000 to 55,000 dollar range. Section 25E used vehicle credit and Section 45W commercial clean vehicle credit follow the same December 31, 2025 cliff. Section 30C alternative fuel refueling property is extended to property placed in service before June 30, 2026, then terminates. Section 25C energy efficient home improvement and Section 25D residential clean energy credits sunset for property placed in service after December 31, 2025, with a narrow extension to June 30, 2026 for projects under binding contract before OBBA enactment.

The technology-neutral electricity credits, Section 45Y production tax credit and Section 48E investment tax credit, remain the largest single fiscal exposure. Original IRA design phased these out beginning the later of 2032 or the year US power-sector emissions fell to 25 percent of 2022 levels. OBBA accelerates the phase-down to 100 percent of credit value for projects beginning construction before December 31, 2025, 75 percent through 2026, 50 percent through 2027, 40 percent through 2028, and zero thereafter. Combined with the FEOC overlay, this collapses the practical investment window to projects that can break ground inside 18 months. Section 45X advanced manufacturing survives the cleanest: batteries at 30 percent of credit value through 2032, solar components through 2032, but wind components phase to zero after 2027, reflecting both the offshore wind political backlash and the absence of a domestic supply chain comparable to solar and battery cell manufacturing.

FEOC: the operational cost the headline number hides #

Foreign Entity of Concern enforcement is the second leg of the phase-out and arguably the more consequential for project finance. The original IRA's FEOC language, sharpened by Treasury and IRS guidance in May 2024 and December 2024, disqualifies any vehicle from 30D or any project from 45X if it contains battery components manufactured by a FEOC (effective January 1, 2024) or critical minerals extracted, processed, or recycled by a FEOC (effective January 1, 2025). FEOC is defined as any entity owned, controlled by, or subject to the jurisdiction of China, Russia, Iran, or North Korea, with a 25 percent ownership, board seat, or voting rights threshold. OBBA tightened the threshold by importing the Section 7701 attribution rules, treating contractual licensing agreements with Chinese parents as effective control, a provision aimed directly at the Ford-CATL Marshall, Michigan licensing structure and similar arrangements at Gotion, Envision AESC, and SVOLT.

The compliance cost is real. Industry filings to Treasury through the December 2024 comment period suggest FEOC documentation, including supplier attestations, beneficial ownership traces, and minerals provenance audits, adds 4 to 7 percent to battery pack cell-level cost, with most of the burden falling on cathode active materials (CAM) and precursor (pCAM) sourcing where Chinese refiners control 70 to 85 percent of global capacity per BNEF and the IEA's 2025 Critical Minerals Outlook. For graphite anode material the share approaches 95 percent, which is why Treasury's December 2024 transition rule for graphite, extended to December 31, 2026, was treated as a tactical lifeline by every domestic battery investor with cells already on US soil. After that date the FEOC restrictions on graphite bind absolutely. Atlas Public Policy and BNEF's April 2025 cancellation tally of 14 billion US dollars in delayed or cancelled solar, wind, and EV manufacturing investment is dominated by projects that could not credibly document a non-FEOC supply chain inside the OBBA window.

The tariff stack: Section 232, Section 301, and reciprocal layers #

OBBA does not operate in isolation. The Trump administration's tariff architecture overlays a second cost shock on the same supply chain. The Section 301 China tariffs, raised by USTR in May 2024 and reaffirmed in February 2025, set duties at 100 percent on Chinese electric vehicles, 50 percent on solar cells and modules (rising to a stacked 50 percent plus existing antidumping and countervailing duties that bring effective rates above 100 percent for many SKUs), 25 percent on lithium-ion batteries for EVs (rising to 25 percent on non-EV batteries from January 1, 2026), and 25 percent on natural graphite and permanent magnets from January 1, 2026. Section 232 reviews opened in early 2025 cover semiconductors, critical minerals, and pharmaceuticals, with the EV-specific Section 232 investigation announced in March 2025 still under review.

The reciprocal tariff regime, announced April 2, 2025 and partially paused for 90 days, layered an additional 10 percent universal tariff plus country-specific rates that hit Vietnam at 46 percent, Malaysia at 24 percent, Thailand at 36 percent, and Indonesia at 32 percent before negotiated reductions. This matters for solar because Vietnam, Malaysia, Thailand, and Cambodia became the primary transshipment routes for Chinese-origin modules after the 2022 to 2024 Department of Commerce circumvention determinations. Stacked, the effective landed cost of an imported utility-scale module in Q1 2026 sits 35 to 60 cents per watt above unsubsidized 2024 levels, against a domestic First Solar or Hanwha Qcells module priced near 30 to 35 cents per watt with 45X absorbed. The tariff stack, in other words, has accomplished what 45X alone could not: domestic content is now the cost-minimizing choice for utility-scale procurement, not just the credit-maximizing one.

Project repricing: where capex still clears, where it does not #

The investment math now sorts cleanly into three buckets. Bucket one, capex still clears: utility-scale solar (with 45Y or 48E captured by 2025 notice to proceed and domestic modules under 45X), grid-scale battery storage (45X protected through 2032, FEOC navigable for non-graphite components), nuclear uprates and small modular reactors (45U and 48E preserved with technology-neutral language), carbon capture on existing industrial facilities (45Q at 85 dollars per tonne for storage), and green hydrogen at 3 dollars per kilogram for projects that secure 45V additionality, hourly matching, and deliverability. Energy and Environmental Economics' April 2026 modeling for SEIA shows utility-scale solar levelized cost of energy at 38 to 46 dollars per megawatt-hour with credits and 58 to 72 dollars per megawatt-hour without, against a combined cycle gas plant LCOE of 65 to 85 dollars per megawatt-hour at 4 dollar per MMBtu Henry Hub and prevailing capacity payments. The credit math still wins by enough margin to absorb a 60 percent phase-down through 2028.

Bucket two, capex repriced but viable: domestic battery cell and pack manufacturing (45X intact, FEOC compliance adds 4 to 7 percent to unit cost, 30D loss compresses near-term offtake), heat pumps and residential efficiency retrofits (state-level incentives in California, New York, Massachusetts, Colorado, and Illinois partially backstop demand), and offshore wind already under construction (Vineyard Wind, Revolution Wind, Sunrise Wind grandfathered, no greenfield sanctioning likely through 2028). Bucket three, capex stalls: new EV models priced at the credit-eligible threshold lose the demand wedge January 1, 2026; onshore wind nacelle and blade manufacturing loses 45X after 2027; standalone green hydrogen without industrial offtake; residential rooftop solar absent state credits or net metering. Atlas's tracker, dividing 470 billion US dollars into 268 billion active, 87 billion paused, and 38 billion cancelled, maps almost exactly onto these three buckets.

Investor and operator playbook: four scenarios through 2028 #

The first scenario, base case at roughly 55 percent probability, is OBBA as enacted holds through 2028 with no further rollback, FEOC enforcement tightens incrementally via Treasury guidance, and tariff stack stabilizes. In this world, US clean energy investment runs at 180 to 220 billion US dollars per year through 2028, down from the 280 billion dollar 2024 peak, concentrated in utility-scale solar and storage in ERCOT, MISO, and PJM, plus battery cell expansion in Georgia, Tennessee, Kentucky, Ohio, and Michigan. The second scenario, deeper rollback at 20 percent probability, sees a 2027 reconciliation bill that eliminates 45X for solar in addition to wind, and tightens 45Y and 48E to a 100 percent phase-down by 2027. This compresses utility-scale renewables to gas-plus-storage hybrids and accelerates capacity retirement deferrals.

The third scenario, partial restoration at 15 percent probability, follows a 2028 election outcome that returns Democratic House control. A targeted clean energy package restores 30D in modified form, reinstates 25C and 25D for income-qualified households, and slows the 45Y and 48E phase-down. The fourth scenario, FEOC enforcement breakdown at 10 percent probability, reflects either a US-China commercial agreement that softens the FEOC perimeter or an enforcement failure that allows compliant-on-paper projects with substantive Chinese content to clear credits. This scenario is positive for near-term capex velocity but corrosive for the political durability of the credits that survive. The actionable conclusion across all four: project sponsors who can document non-FEOC supply chains, lock in 2025 notice to proceed, and structure offtake with creditworthy industrial counterparties retain optionality across every plausible policy outcome. Sponsors who modeled flat 30 percent credits and hoped to refinance into 2030 capacity markets are now solving for a problem that no longer has the parameters they assumed.

Sources #

• US Treasury, IRA implementation page
• IRS Notice 2024-49, FEOC and Section 30D
• Congressional Budget Office, IRA budgetary effects rescore, January 2025
• Joint Committee on Taxation, OBBA revenue estimates
• One Big Beautiful Act, Public Law text via Congress.gov
• Atlas Public Policy, Clean Economy Tracker
• BloombergNEF, US clean energy cancellations and pause tracker, April 2025
• USTR, Section 301 China tariff actions, 2024 and 2025 modifications
• US Department of Commerce, solar circumvention determinations
• Department of Energy, Loan Programs Office portfolio
• Energy Information Administration, Annual Energy Outlook 2025
• Federal Energy Regulatory Commission, generator interconnection queues
• International Energy Agency, Critical Minerals Outlook 2025
• SEIA, US Solar Market Insight reports
• Reuters, IRA phase-out reporting, 2024 to 2026
• Financial Times, OBBA reconciliation coverage
• Penn Wharton Budget Model, IRA cost estimates