For decades, industrial policy was treated as an embarrassment — something the advanced economies had left behind, along with dirigisme and five-year plans. The consensus was that governments should set rules and let markets pick winners.
That consensus is gone. Governments are back in the business of shaping industry, with budgets that would have seemed inconceivable a decade ago. But the new industrial policy does not look like the old one. It is quieter, more defensive and aimed at a different target.
Why the old consensus collapsed
The retreat from industrial policy was based on a clear theory: governments cannot pick winners, so they should not try. The theory had real evidence behind it — plenty of state-backed projects had failed spectacularly.
But the world changed around the theory. Supply chains proved fragile. A pandemic, a war and a series of trade disruptions showed that some industries mattered more than their market prices suggested. And the transition to clean energy made clear that some transitions require coordination that only government can provide.
The result is that the question is no longer whether governments should shape industry. It is how, and for what purpose.
The defensive character
The most striking feature of the new industrial policy is that much of it is defensive.
It is less about pushing new industries into existence and more about protecting ones that exist — semiconductors, batteries, critical minerals, pharmaceuticals, medical supplies. The logic is not that these industries will conquer the world, but that depending on a single foreign source for them is a strategic risk.
This is a different ambition from the old industrial policy. The old version reached for national champions. The new version reaches for resilience — ensuring that key goods can be produced at home, or at least somewhere friendly, even if it costs more.
The supply-side tilt
The new policy also has a distinctive approach: it works on the supply side, not the demand side.
Rather than subsidizing demand (as with consumer incentives), it funds supply — factories, infrastructure, research, training. It builds domestic capacity directly, using grants, loans, tax credits and guaranteed purchase agreements. The instruments are financial, but the target is industrial capacity itself.
This tilt has a practical logic. When the goal is resilience, you need the physical capability to produce, not just the demand for the product. Capacity is the asset, and policy is buying capacity.
The subsidies question
None of this is free, and the bill is arriving at a delicate moment for public finances.
Governments are borrowing heavily, interest rates are elevated, and every subsidy program competes with other priorities. The question of how much subsidy is enough — and how much is too much — is being argued in every treasury.
There is also a coordination problem. When several countries subsidize the same industry at once, the result can be overcapacity — more factories than demand, a subsidy war that benefits nobody. The recent rush to subsidize semiconductor and battery plants is already producing this risk.
The market’s response
Markets are responding to the new industrial policy with a mixture of enthusiasm and caution.
Companies in subsidized sectors have seen their valuations and investment plans grow. Suppliers and contractors are benefiting. But investors are also asking hard questions: what happens when the subsidies end, and will the demand be real without them?
The history of industrial policy suggests the answers will be mixed. Some subsidized industries will mature into genuine competitiveness; others will need continuing support, and some will fail despite it. The test is not whether every project succeeds, but whether the portfolio as a whole adds to resilience and capability.
The skills bottleneck
One of the least-noticed constraints on the new industrial policy is people.
Building a semiconductor fab requires thousands of highly trained engineers. Building a battery industry requires chemists, materials scientists and factory technicians. The money to build the plants is available; the workforce to run them is not. In several countries, the skills gap has become the binding constraint.
This is a reminder that industrial policy is, at bottom, an investment in human capability. The countries that pair capital with training — that treat education and apprenticeships as part of the industrial program — will be the ones that convert subsidy into lasting industry.
The honest assessment
Industrial policy is back, and it will be with us for a long time. The circumstances that brought it back — fragile supply chains, strategic rivalry, the energy transition — are not going to disappear.
The question is whether governments will practice it well. The best versions of the new industrial policy are transparent, time-limited, and tied to measurable outcomes. The worst are opaque, permanent and politically captured. The difference is not between policy and no policy; it is between policy that builds capacity and policy that builds dependence.
The old industrial policy promised to pick winners. The new one has a more modest and more honest ambition: to make sure the essential industries do not vanish, and that the capacity to adapt survives the shocks that are coming.
That is a less heroic goal, and a more defensible one. And it will require discipline, transparency and patience to get right — qualities that are scarcer than subsidy dollars.