We Tariff Steel. Why Not Software Jobs?

The United States has tariffs on steel, aluminum, lumber, and semiconductors. The argument is straightforward: when critical industries get hollowed out by cheaper foreign labor, the long-term cost to the economy is worse than the short-term savings. So we protect them.

Software engineering is now one of the most critical industries in the country. It drives everything from defense systems to financial infrastructure to the tools every business relies on daily. And for the last 20 years, American companies have been shipping those jobs overseas at scale, with zero friction and zero consequences.

That should change.

The Double Standard Is Hard to Ignore

If a company wants to import steel from another country, they pay a tariff. The reasoning is that domestic steel production matters for national security, economic resilience, and American jobs. Fair enough.

But if that same company wants to ship 200 software engineering jobs to Bangalore, there is no tariff. No penalty. No friction at all. They just do it, pocket the labor arbitrage, and move on. The American engineers who built the systems get laid off, and the institutional knowledge walks out the door with them.

The logic does not hold up. If domestic steel production is a matter of national interest, domestic software engineering is ten times more so. Software runs the country. The people who build and maintain it should not be treated as a line item to optimize away.

The Cost Savings Are Not What They Used to Be

The original pitch for offshore engineering was compelling: get the same work done at a fraction of the cost. And for a while, the math checked out on a spreadsheet. But the hidden costs have been piling up for years.

Timezone friction slows everything down. Context gets lost in handoffs. Codebases accumulate technical debt because the teams building them rotate every six to twelve months. Quality issues compound. What looked like a 60% cost savings on paper often turns into a net negative when you factor in rework, coordination overhead, and the drag on velocity.

Companies kept doing it anyway because the savings showed up on one line of the P&L and the costs were spread across a dozen others. That made it easy for finance to justify and hard for engineering to argue against.

AI Changes the Equation Entirely

Here is where it gets interesting. AI coding tools have gotten good enough that a senior domestic engineer with the right tooling can now outproduce a team of four or five offshore developers. Not in theory. In practice, today, at companies that are paying attention.

The value proposition of offshore engineering was always about volume. More hands, lower rate. But when AI compresses the value of volume, the entire model falls apart. You do not need 150 developers in three timezones when 20 engineers in one timezone with AI tooling can ship more, faster, and with fewer defects.

This means the argument against a tariff on offshored jobs just lost its strongest leg. Companies can no longer claim they need offshore teams to stay competitive on cost. The technology exists right now to build leaner, faster, and cheaper with domestic talent. They just have to be willing to invest in it.

What a Tariff Could Look Like

Nobody is suggesting we ban offshore engineering. The proposal is simpler: make it cost something. If a US-based company with domestic revenue wants to replace American engineering jobs with offshore labor, that decision should carry a cost that reflects the economic impact.

It could be structured as a percentage of the offshore labor spend. It could be tied to the number of domestic positions eliminated. It could be a flat surcharge per offshore contractor above a certain threshold. The specifics matter less than the principle: if you are going to export American jobs, you should contribute something back to the economy you are pulling talent out of.

That revenue could fund retraining programs, subsidize AI tooling adoption for small and mid-size businesses, or support STEM education. The point is not to punish companies. It is to create a cost structure that reflects reality instead of pretending there are no consequences to hollowing out domestic engineering capacity.

The Companies That Stay Domestic Will Win Anyway

Here is the thing most executives have not figured out yet: the companies that invest in smaller, AI-augmented domestic teams are going to outperform the ones clinging to the offshore model. They will move faster, build better products, and spend less doing it.

A tariff would accelerate that transition. It would force companies to run the real math instead of the spreadsheet version that ignores coordination costs, quality issues, and the compounding drag of technical debt. And for the companies already investing in domestic talent and AI tooling, it levels the playing field against competitors who are still racing to the bottom on labor costs.

We protect steel because we decided it matters. Software engineering matters more. It is time the policy caught up.