What percentage were created by H1Bs? I think you're conflating "immigrant" with "H1B visa holder". The two are not the same. Many of us who are critical of H1B and similar visa programs are very much in support of immigration in general.
Well H-1Bs are generally prohibited from doing any paid work outside their current employer. So it would be a miracle if that number was non-zero. But it's still not. Zenefits co-founder Laks Srini had to be hired as a database administrator on an H-1B so that he could leave his previous employer [1].
It seems unfair to say people on visa programs don't become entrepreneurs when, in most cases, they legally aren't allowed to.
1. https://www.computerworld.com/article/2852250/it-outsourcing...
Of course, H1Bs can't start companies, so that's not a very good way to look at things - if you don't legally allow category x to do y how many xs are going to do y? A better question is how many could.
But let's just talk about employees. I know I've worked at companies in the past in which the backbone of the work (not all of it, but an important part) was done by H1Bs, and we wouldn't have been able to complete that work and be solvent without the (brilliant) folks that had come over on H1Bs doing it. Those, say, 25 H1B jobs created jobs for hundreds of other folks.
So I'm just wary of thinking that everything is a zero sum game or that immigration is "taking jobs" of "Americans."
I run a school, and we have had employers pay for training of dozens of employees. It happens quite frequently.
The reason they do not do so now is because employees, for obvious reasons, have no loyalty to companies, and tend to jump ship after being trained. But again, this is because US companies changed from a lifetime model of employment to layoffs and offshoring for quarterly stock bumps.
Force companies to use domestic labor, create a culture where employees aren't constantly worried they'll be canned the next time sales are low, and I bet you'll get good employees who would be willing to learn and stay for the long-term.
Can that startup afford to relocate or remotely train a couple hundred out of work Americans for say, a year, pick the top half of those, and offer them packages?
Maybe? Would a year's training be enough? I don't know, but I doubt it's cost effective for spiky demand in labor.
This sort of supply issue should best be solved by the government, which has a vested interest in the employment of its taxpayers, or an NGO or some other organization. Not the companies in question.
Start ups are largely built on sweat equity. The kind of monthly salary most H1Bs generally work on. That comes with a risk. Most immigrants take it anyway because they start from 0 and you can only grow from there. If enough people do this enough number of times, eventually the system will always product a Google or Microsoft.
So ultimately its about who is ready to work for cheap.
Typically you create it with an E2, which let you issue E2 visas for 3-4 other people. Then either you die, sell, or start issuing J1, later converted in H1B. H1B is generally too risky, being bound to a lottery, and that's the last thing you want at an early stage startup.