In any negotiation, you need to understand what leverage either side has. In this case, CA could block the conversion, and OpenAI could leave California. Both are possible but extremely unlikely outcomes! So the whole point is to take these unlikely outcomes to the table, negotiate in good faith, and come out with an agreement.
So California needs to believe that OpenAI will stay in California just as much as OpenAI needs to believe that CA won't block the conversion (or impose other onerous regulations around AI). So yes, it's possible to speculate about whether or not people are sincere in their motivations, but when you need to make a deal, there needs to be a measure of good faith and trust on both sides in order to make something happen.
And in this case, both sides are incentivized to make the deal. OpenAI wants to be a PBC in order to access more capital, and California wants OpenAI to be a PBC so that it can IPO so that all employees (all of whom are likely CA residents), will sell stock, which can then be taxed as CA income.
If I am understanding things correctly, OpenAI could leave in the future - but CA can’t retroactively block the conversion in that case.
California is home to 1 in 8 Americans, and likely a much higher fraction of AI researchers, users and partner organizations to OpenAI (like Nvidia). The California AG has plenty of leverage beyond blocking/reversing the conversion. What leverage will OpenAI have after "leaving"[1] the California?
1. They're guaranteed to have an engineering office in the SF Bay. Not many of those folk will agree to relocate to Texas/Miami.
Yup the Tesla treatment. Can change HQ's all you want but the main engineering work, brains, and talent will be at a HQ in California no matter what. Sam will probably do this for brownie points with the current administration, it will be politicized news for a cycle, but after the dust settles the majority of non-admin people will still be working unchanged out of CA.
That’s assuming Altman is sincerely going to keep trying to develop “AGI” and not try to turn OpenAI into a profitable business. You don’t need AI researchers if you get good enough video generation and pornbots to become immortally wealthy and say fuck the rest. If this is the case, OpenAI could be a completely done product, all that’s left to do is stop spending so much money on SG&A and get those revenue streams cranking.
What's stopping OpenAI from offering telework? AI isn't dealing with high security or particularly sensitive data (that developers need to actively access - training data is all public or stolen works) and it's not a hardware product.
Yup, but the IPO will still have happened in CA, and there's going to be a tax windfall from that.
It's about a moment in time, not an "in perpetuity" agreement.
Unless you're Universal and Marvel, leaving Disney unable to buy out Universal's contract with Marvel, and unable to use classic comic book characters because the parks too close to Universals. Still cracks me up.
Not if the big holders aren't residents, they can move away just before like Rogan with his Spotify deal, or Jonathan Blow just before a game release after developing in California and going to public college there, etc.
Since it's a non-profit still holding it any gains to the non-profit entity upon the conversion don't go to California, and principal stakeholders can move away. Other funds raised from the IPO can be invested in other states untaxed (long term datacenter leases instead of booking the capital of building one) until they move the company away I think.
There will probably be a lot of smaller stakeholders that stay with a lot of money for the state, and California at least doesn't do the $15 million QSBS so they may get a lot from that tail of employees. A large portion of this tail of lower compensated employees may get laid off due to AI replacement before IPO and lose a lot of unvested years, if we are to believe OpenAI's own claims about timelines for job replacement in that field at lower levels.
I'd recommend anyone expecting to profit from OpenAI stock and aiming to avoid California taxes to look into the subject in depth with paid advisors. The California FTB is not stupid or naïve and has a history of successfully getting paid for stock that vested with a California nexus, even if the beneficiary moves out of state. Good luck!
You're right, it's harder with vesting stock compensation than other things you build up over time like an audience or a developing a game over a long span.
Take a look at Shohei Ohtani's contract.