In 2013 we moved a family to San Francisco to join a big data startup. The work was genuinely excellent. The city was alive in the way that only a city mid-boom can be. The team was sharp, the product had real customers, and the technical problems were the kind that get you out of bed early. None of that is the part worth writing about.
The part worth writing about is the visa.
The thing nobody quite tells you
An H1B sounds like a work permit. It is not. It is a tether. You are not employed in the United States — you are employed by one specific company in the United States, and the moment that relationship ends, so does your legal right to be there. There is no neutral ground. There is no "I'll take a few weeks and figure out what's next." The visa is the job, and the job is the visa.
California is an at-will employment state. There is no statutory notice period. The day a downsizing happens, you are technically out of status. Your spouse's work authorisation, your kids' school enrolment, the lease on the apartment — all of it sits on top of a status that evaporated at 9am on a Tuesday.
We were fortunate. When the downsizing came, we negotiated a bridge that kept us in status long enough to transfer employment to another sponsor. That negotiation happened under the kind of pressure that is hard to describe to anyone who has not lived inside it. The clock is not metaphorical. It is a real clock, and it is loud.
The green card was worse
If the H1B is a tether, the green card process is a maze where someone keeps moving the walls. It is long. We knew it would be long. What we did not fully appreciate was how exposed you remain throughout. At one point the government decided to re-review our existing H1B — not the green card application, the underlying visa — and that essentially meant restarting significant pieces of the process from scratch. Years of paperwork, evidence, letters, prevailing wage determinations, all suddenly back in play.
You learn to keep every payslip. Every tax return. Every employment letter on company letterhead. Every plane ticket in and out of the country, in case someone asks you to reconstruct your physical presence on a particular date in 2015. You become an archivist of your own life, because the alternative is being unable to prove you are who you say you are.
We came home in the end. The decision was not dramatic. It was the accumulation of a hundred small frictions, each one survivable on its own, none of them sustainable as a permanent condition. The day we landed back in New Zealand, the absence of that low-frequency hum of immigration anxiety was the most noticeable thing about being home. It took weeks to stop reaching for the folder.
The current moment
The US immigration system is in the middle of significant change. Green card categories are shifting, processing posture is shifting, and the political signal around skilled migration is, to put it generously, mixed. We are not going to predict where it lands. We will say this: anyone considering an H1B today is signing up for a category of uncertainty that did not exist at the same intensity a decade ago, layered on top of the uncertainty that was already baked in.
If the answer to "what happens if this goes sideways?" is "we figure it out then," that answer is much more expensive than it sounds.
Why this is actually an opportunity
Here is the part that matters more than our story. The implicit assumption in the H1B path is that the United States is where serious technical work happens, and that proximity to it is worth significant personal cost. That assumption was always weaker than it looked, and it is weaker still now.
New Zealand has a long history of outperforming its size. The number 8 wire instinct — make it work with what you have, ship the thing, iterate — has produced an unreasonable number of category-defining companies and inventions for a country of five million people. Rocket Lab. Xero. Weta. Halter. Soul Machines. The pattern is not luck. It is a particular kind of engineering culture that prizes pragmatism over ceremony.
That culture is well-suited to the current moment. The economics of building software have shifted in a way that disproportionately rewards small, high-context teams that can move quickly. AI tooling has compressed the work that used to require a fifteen-person team into something three engineers can ship in a quarter. The bottleneck has moved from headcount to taste — to product thinking, to knowing what is worth building and what the second-order consequences are. Those are skills that do not require a particular postcode.
What we would tell a younger version of ourselves
If you are a New Zealand engineer or founder weighing an H1B today, the question is not "is the work in the US better?" Sometimes it is. The question is "what is the actual delta, and what am I trading for it?"
The delta is narrower than it used to be. Remote work normalised cross-border collaboration. Capital is more globally distributed than it was in 2013. Customers in North America are entirely happy to buy from a company headquartered in Auckland or Wellington or Christchurch, provided the product is good and the team is responsive. The friction of being elsewhere is real but it is measurable. The friction of being on an H1B is not measurable until something goes wrong, and then it is the only thing that matters.
The trade is your optionality. On an H1B you cannot easily start a company. You cannot easily take a sabbatical. You cannot easily take a lower-paying role at a more interesting startup. You cannot easily say no to your employer when they ask for something unreasonable, because your employer is also your immigration sponsor. Every career decision routes through a lawyer.
For some people, in some roles, the trade is still worth it. We are not going to pretend otherwise. But it is a trade, and it should be priced as one.
Building from here
The more interesting bet is to build from New Zealand. Use the AI tools that have collapsed the cost of getting a v1 in front of customers. Sell into whatever market makes sense — most of them are reachable from a laptop in Mount Eden. Hire the people you actually want to work with, rather than the people who happen to live within commuting distance of an office you are paying too much for. Keep the team small enough that product thinking remains the binding constraint, because that is the constraint that AI cannot remove for you.
The US will keep being a large and important market. It will not be the only place where good work happens, and for a meaningful number of New Zealand engineers it is no longer the obviously correct destination. That is not a loss. It is a reallocation, and we think the country on the receiving end of it stands to benefit.
We spent years trying to stay in someone else's system. The work we are doing now, from here, is better. Make of that what you will.