$100,000 H-1B Visas: What Does it Mean for AI?

President Trump turned the employment world upside down this week by announcing a $100,000 application fee on H-1B visas for highly-skilled immigrants. Let’s talk about what that actually means… and, given that the tech sector in particular is highly dependent on the visas, what the implications are for the AI revolution.

As you might expect, there’s good… there’s bad… and there’s ugly.

We’ll start with the basics.

Up until the announcement, H-1B visas were relatively cheap, with application fees ranging from $1,700 to $4,500. And they were in high demand. The government only issues 65,000 per year, with an additional 20,000 set aside for recent foreign graduates of American masters or PhD programs. There can be hundreds of thousands of applicants in any given year, so the visas were generally awarded via a lottery system.

Among other requirements, the employer applying for the visa has to demonstrate that they are paying a prevailing market salary for the position and that they’re not using the visa as a way to undercut more expensive American labor.

Raising the price to $100,000 changes the calculus. (Note that the initial messaging was somewhat muddled; Commerce Secretary Howard Lutnik implied it was an annual fee, though it has now been clarified to be a one-time application fee.)

A would-be employee that was “worth” a $4,500 application fee might not be worth an extra $100,000 on top of their salary and benefits. Or for smaller employers, it may simply be a price they can’t afford. As a practical matter, the companies in search of labor will have to be a lot more selective about who they choose to sponsor. It’s a much larger investment now.

A more selective immigration policy isn’t bad, per se. It only makes sense to want to bring in the best of the best. But the massive bump in cost also means that a lot of quality would-be immigrants will get passed over… and that their skills will be used to make other countries wealthy rather than ours. It also means that larger, cash-flush companies like Nvidia or Meta will have their pick of highly-skilled immigrants, whereas smaller employers, start-ups or companies without political connections in Washington will find themselves priced out.

Let’s think through this.

In the last round, there were 339,000 applicants for H-1B visas. That means there were 339,000 jobs that American businesses were looking to fill and that they felt they couldn’t fill with American workers. These aren’t would-be immigrants looking for a job. These are unfilled job postings that already exist… just waiting for the right person to fill them.

American workers are the most productive in the world. But the reality is that there simply aren’t enough of us. Our working-age population has barely budged over the past 10 years and is expected to grow at barely 0.2% per year through 2030. Retiring Baby Boomers aren’t being replaced at the rate we need them to be.

If anything, we should be making it easier and cheaper to bring in highly-skilled foreign workers. We need the manpower.

What Makes Silicon Valley Special?

Not all H-1B visas go to tech workers. The US imports plenty of other specialists like doctors, skilled nurses, engineers and even professors. I know American aficionados of Peruvian paso horses who used H-1B visas to bring specialized horse trainers from Peru.

But let’s focus on technology, since the lion’s share of the H-1Bs go to that sector.

What makes Silicon Valley special?

It certainly isn’t California’s business climate. If anything, Silicon Valley succeeds in spite of its location rather than because of it.

Silicon Valley is the world’s preeminent innovation center because it attracts the best, brightest and most ambitious from around the world. It’s the people, not the place.

Now, the Trump Administration’s H-1B visa changes aren’t going to change that, at least not immediately. But given that we’re in a life or death struggle for AI supremacy with China, we should be bringing in more tech workers, not fewer. The hypothetical Indian AI developer that would have gone to Silicon Valley is still going to produce code. He’s just going to do it for a company in India, Dubai or London rather than America. And American tech companies will be incentivized to set up innovation campuses on foreign soil to utilize the talent they can’t find at home.

Considering Trump’s business reshoring efforts, this is just silly. It’s handicapping Trump’s own efforts to bring business to American shores.

In a typical rock-and-a-hard place situation, keeping the very people who could help move forward our efforts in the AI race also slows the positive impact AI can have on our dismal demographic outlook.

Speaking of AI…

There’s another wrinkle here too.

With or without the changes to the H-1B program, America is facing a labor shortage. The demographics are already baked in. Decades of low birthrates have left us with a workforce that simply isn’t growing at the speed we need it to.

This is why the AI revolution is so critically important. In the absence of more workers, we need to get more productivity out of the workers we have. Otherwise, the economy doesn’t grow.

It really is that simple.

The stricter immigration policy becomes, the more critically important AI development is. The work still needs to get done. If there’s no human to do it, it will need to be done by a machine.

AI-powered agents like ChatGPT are already transforming white collar work, but this is only the beginning. The next major leap will be in more localized AI applications like autonomous driving and robotics. For all of the hypo about robo taxis popping up in American cities, this revolution is still in its infancy. It hasn’t even started yet. Not really. That massive growth explosion is yet to come.

For an easy one-stop shop to take advantage of this trend, consider the Xtrackers Artificial Intelligence and Big Data ETF (XAIX). It’s a collection of some of the largest players in the AI space, including Oracle, Palantir and Nvidia. I recommended the ETF back in December, and my readers are already up about 21%.

Until next time,

Charles Sizemore