The 2022 Inflation Reduction Act and CHIPS Act together represent the largest US industrial-policy push in seventy years. The IRA's energy and manufacturing provisions are roughly $369 billion in tax credits and direct spending over ten years; the CHIPS Act adds $52 billion in semiconductor manufacturing subsidies. The implementation data so far is mixed in ways that matter for whether industrial policy can succeed in a system not built for it.

What Got Announced vs. What Got Built

By mid-2024, IRA-related manufacturing announcements exceeded $250 billion in committed private capital across battery, solar, and EV manufacturing. Construction starts ran behind announcements; production ramps ran behind construction. Hyundai's Georgia EV plant began limited production in late 2023. Form Energy's West Virginia iron-air battery plant began construction in 2024. The Treasury Department's Section 30D implementation, which determines which vehicles qualify for the consumer tax credit, has been revised multiple times under industry pressure, generally in the direction of looser foreign-content rules.

The Brookings tracking work shows the geographic distribution skewing toward Republican-leaning states — South Carolina, Georgia, Texas, Tennessee — partly by accident of where energy projects make sense and partly by deliberate design to build durable political coalitions for continued funding. Whether the political-economy strategy succeeds, in the sense that Republican legislators defend IRA spending in their own districts, will determine whether the program survives an unfriendly administration.

The CHIPS Implementation

The CHIPS Act's biggest awards — TSMC Arizona ($6.6B), Intel Ohio ($8.5B), Samsung Texas ($6.4B), Micron New York ($6.1B) — have been announced, with construction underway and first-fab targets in 2025-2027. The technological gap with leading-edge fabs in Taiwan remains substantial; the US fabs will produce earlier-generation chips at scales below what TSMC ships from Tainan. The strategic-autonomy argument for these subsidies is less about leading-edge capability than about geographically diversified supply, which is a more defensible goal but a more modest one than some of the original rhetoric.

Where Industrial Policy Has Worked

The historical track record is mixed but not as bad as 1980s free-market critiques suggested. East Asian developmental states — Japan's MITI, Korea's industrial promotion, Taiwan's ITRI — produced industries that competed at the world frontier. The conditions for success in those cases included strong export discipline (industries had to compete in world markets, not just sell to captive domestic buyers), reciprocal commitments (subsidies in exchange for measurable performance), and competent administrative capacity. The US has weaker versions of these conditions, which is the most-cited risk for the current programs.

Mariana Mazzucato's The Entrepreneurial State documents that even apparently market-led innovation in the US — semiconductors, the internet, biotech — depended on substantial publicly funded basic research and procurement. The IRA and CHIPS are partly an attempt to restore that capacity at scale. Whether the institutions exist to do it competently is genuinely uncertain.

The Trade-Off with Free-Trade Orthodoxy

Industrial policy is, by definition, picking winners. That conflicts with the trade-policy framework — most-favored nation, non-discrimination, limits on subsidies — that the US itself wrote and championed for decades. The IRA's domestic-content requirements have generated WTO complaints from the EU, Korea, and Japan that may or may not be sustained. The political cover for ignoring these complaints comes from the framing that China's own industrial policy long since broke the rules; the US is now responding in kind. Whether this can be the durable framework for global trade or whether it is a one-shot realignment that requires a new set of rules to follow is the unresolved question.

The Risks That Are Specific to Industrial Policy

The historical failure mode is capture: the targeted firms become political constituents of the subsidy program rather than performers held to commercial discipline. The 1970s steel and shipping subsidy programs are the textbook examples. The IRA's structure tries to avoid this by routing most of the spending through tax credits with relatively mechanical eligibility rules, but the Treasury rulemaking has shown how susceptible even mechanical rules are to industry pressure. CHIPS Act spending, by contrast, is directly negotiated and includes clawback provisions, which is closer to the East Asian model but more dependent on administrative competence.

The Honest Reading

Industrial policy works in some institutional configurations and fails in others. The US is attempting it on a scale not seen since the Apollo program, with administrative capacity that has atrophied since the 1970s. The IRA's tax-credit structure is the most robust against capture; the CHIPS Act's direct grants are more vulnerable but potentially higher-leverage. The five-year mark — 2027 — will produce the first serious empirical test of whether US industrial policy can be done well or whether it inevitably reproduces the 1970s failure modes. Until then, the case for or against is heavily theoretical, and the political constituencies that the spending is building may matter as much for the program's survival as the economics of any specific project.

The Replicability Question

The IRA's success or failure has implications for whether the US can sustain ambitious policy across multiple sectors. The political-economy strategy of tying spending to specific congressional districts has been more durable than prior industrial-policy efforts. Whether the strategy can be replicated for other priorities — housing, healthcare, education, care infrastructure — depends on whether the same political-coalition-building can work in sectors that lack the industrial geography that drove the IRA's design. The next decade will reveal whether the IRA is a one-off success or the template for a new generation of place-based federal policy.

The Capture Watch

The historical failure mode of US industrial policy has been capture — the targeted firms become political constituents of the subsidy program rather than performance-disciplined participants. The IRA's tax-credit structure is more resistant to capture than direct grants would be, but not immune. The Treasury rulemaking on Section 30D and Section 45V has been progressively loosened under industry pressure since the initial 2023 guidance. Whether the program retains substantial performance discipline through its ten-year scheduled life, or whether it gradually shifts toward whatever the recipients prefer, is the long-run institutional question. The early signs are mixed.