Posts
The Incentive Engine
The Inflation Reduction Act didn't build the Battery Belt. It triggered a competition between state governments to build it on the federal government's behalf — and then mostly stepped aside. Tens of billions in production tax credits and loan guarantees moved. The governance terms attached to that capital did not move with the same force. This post follows the money and asks what it bought.
What the IRA Actually Did
```The Inflation Reduction Act of 2022 deployed battery manufacturing incentives through three primary mechanisms: production tax credits that pay per unit of output, investment tax credits that subsidize facility construction, and DOE loan guarantees that de-risk the capital required to build gigafactories at speed. The combination was designed to make American battery manufacturing financially viable before the market could price it that way on its own.
It worked as a trigger mechanism. In the two years following IRA passage, announced battery manufacturing investment in the United States exceeded $70 billion. The Battery Belt's physical inventory — documented in Post 1 — did not exist at that scale before 2022. The IRA created the conditions under which foreign battery companies and American automakers concluded that joint venture construction in American right-to-work states was the highest-return deployment of their capital.
That last sentence is the one that matters. The IRA did not direct where the Belt would be built, who would own it, what labor standards would govern it, or what would happen to the assets when the credits expired. It created financial conditions. Private capital — predominantly foreign-partnered — made the location, ownership, and governance decisions within those conditions. The federal government purchased assembly capacity. It did not purchase sovereignty over it.
```$10/kWh modules
10% electrode materials
Phases to zero 2033
OBBBA curtailment
(e.g., $9.63B BlueOval SK pre-restructure; $2.5B Ultium)
The Phasedown Architecture
```The IRA's production tax credits were not designed to be permanent. They were designed to be large enough, and long enough, to trigger irreversible private investment — to get facilities built, production lines running, and supply chains rooted deeply enough that the manufacturing base would survive the credits' expiration. The theory is sound. The execution carries structural risk that the Belt's governance architecture does not fully address.
Section 45X phases down beginning in 2030, reaching zero for most battery components in 2033. The critical minerals credit portion is permanent — or as permanent as any tax provision is — but the cell and module credits that drove the bulk of gigafactory economics do not survive the decade. The $35 per kilowatt-hour cell credit that made joint venture economics viable in 2022 will be $8.75 per kilowatt-hour in 2032 and zero in 2033.
The joint venture agreements that govern those same facilities carry no equivalent phasedown. The IP, technology licensing, and production volume decision-making structures embedded in those JVs are not indexed to the credit schedule. Post 3 examines what that asymmetry means in practice.
```"The IRA's credits phase down on a published schedule. The joint venture agreements that govern the same facilities do not. That asymmetry is not an oversight. It is the governance architecture."
The Governance Gap
```Federal industrial policy of this scale typically attaches governance conditions to public capital in proportion to the investment. The IRA's governance architecture is more limited than its capital architecture suggests it should be.
The primary governance mechanisms are: job creation benchmarks attached to state incentive packages (not to federal credits), domestic content thresholds embedded in the credit eligibility rules (phased in over time), FEOC restrictions targeting China-linked supply chains (enforced through audit, not structural ownership change), and DOE loan covenants (project-specific, facility-level).
What the IRA does not impose: mandatory collective bargaining requirements for facility operations (prevailing wage applies to construction, not permanent operations in most configurations), clawback provisions triggered by JV restructuring or foreign partner exit, domestic technology development requirements as a condition of credit eligibility, or any binding commitment on facility operation duration beyond the depreciation schedule of the public loan.
The governance gap is the distance between what public capital was deployed to achieve and what the governance terms actually require. The Belt's post-IRA restructurings — BlueOval SK splitting the KY and TN plants between Ford and SK On, Samsung SDI pausing Indiana construction, multiple DOE loan rescissions — occurred within the governance architecture's tolerance. None triggered a clawback. None terminated a credit stream. The architecture absorbed the restructurings because it was designed to do so.
```
BlueOval SK was the flagship Battery Belt joint venture announcement: Ford and SK On, twin plants in Glendale, Kentucky, a third in Stanton, Tennessee, $9.63 billion in DOE loan guarantees, 5,000+ jobs promised, a 1,500-acre manufacturing campus named BlueOval City. The IRA made the economics viable. The joint venture made the ownership structure hybrid.
In late 2025, as EV demand softened and Ford's EV losses mounted, the JV restructured. Ford took the Kentucky plants. SK On took the Tennessee plant. The DOE loan package — originally sized for a unified JV — required renegotiation. Jobs announced for the unified operation were redistributed across a split structure. No federal governance mechanism triggered a credit clawback or a formal accountability review.
The FSA finding: The restructuring was structurally permitted. Which means the governance architecture was not designed to prevent it. Which means the architecture was designed for something other than manufacturing sovereignty. It was designed to trigger construction. What happens after construction is governed by the JV agreements, not the IRA.
The OBBBA Modification
```The One Big Beautiful Bill Act of 2025 modified the IRA's incentive architecture in ways that cut in two directions. On the demand side, the 30D consumer EV credit was effectively ended for vehicles purchased after September 30, 2025. That removed the market demand signal the Belt's production capacity was built to serve, accelerating the demand softness that was already causing facility pauses and JV restructurings.
On the supply chain side, the OBBBA expanded the FEOC framework into a more stringent Prohibited Foreign Entity structure with Material Assistance Cost Ratio thresholds that tighten annually. This tightens the Chinese supply chain exclusion architecture — which is a genuine governance strengthening, but one that creates compliance costs and supply chain restructuring pressure for facilities whose upstream mineral flows still run through Chinese refining infrastructure.
The net effect is a policy architecture that simultaneously removed demand-side support (30D termination) and tightened supply chain compliance requirements (FEOC/PFE expansion), while leaving the production credit schedule (45X) largely intact for facilities that can meet the sourcing requirements. The facilities that cannot meet those requirements lose both the demand pull and the production credit. The facilities that can are operating in a tighter compliance regime with less market support than the original IRA anticipated.
Whether this is a coherent industrial policy or an incoherent one depends on whether you believe the Chinese supply chain exclusion architecture can be enforced at the mineral refining layer — where China's structural dominance is documented in Post 4 — without destroying the economics of the facilities it is supposed to protect.
```The FSA Finding
```The IRA was the largest federal manufacturing incentive deployment in American history since World War II. It worked as designed: it triggered investment at scale, in locations the private market would not have chosen without the subsidy, in a timeframe the market would not have sustained without the credit schedule.
The governance architecture attached to that capital is proportionate to the goal of triggering construction, not to the goal of sustaining sovereignty. Job creation benchmarks are state-level and largely unenforced at the federal layer. Domestic content thresholds are phased in too slowly to prevent the assembly-without-upstream-control structure the Belt currently has. FEOC restrictions address Chinese supply chain exposure without requiring ownership change in the Korean and Japanese JV structures that actually control the technology. The production credits expire. The joint ventures do not.
The IRA purchased a manufacturing buildout. It did not purchase manufacturing sovereignty. Those are not the same transaction, and the governance gap between them is where the series' central question lives.
Post 3 opens the JV agreements. The ownership architecture is where the gap becomes structural.
```
