This isn't really true with ISOs (hence the point of them). You may get an AMT gain which is ugly, but you don't owe regular taxes on the spread unlike Non-Quals where you would.
Yes, you're absolutely correct. That's a bad job on my part.
ISOs have a chance of pushing you into AMT land. If you exercise ISOs, and the FMV is different than the exercise price, then you should consult with a CPA to find out if you have a tax liability.
When I was in this situation I had a CPA project my taxes for the current year. It turned out that my taxes under the traditional system were more than under AMT, so I didn't have to worry. But this easily could not be the case if the spread is sufficiently large.
The key is that you pay taxes on that spread. If you're early enough - I was roughly #25 - and do it early in your tenure, then you only have to come up with the cash to buy the shares and a minor tax bill. If I had waited until I left to execute, the spread would have been 12-15x. I know a few people who stayed 4 years to fully vest and then executed. I don't know detailed numbers but it sounded painful.
If/when Twilio eventually IPOs, then the ROI will be far better than any index fund.
(I don't know anything about the "if/when" as I haven't been inside in over 2 years.)
How do you know there will ever be a spread? If your startup fails, your shares are worthless. Or better yet, your shares are diluted out of most of their value by several subsequent rounds of private equity, which generally you have no control over whatsoever, but which will certainly go to enrich the founders. Resulting in even more direct transfer of wealth of your investment, to the founders and venture capitalists.
I'm having trouble understanding why any startup employee would do this, as opposed to exercising stock options when they actually have value and ideally some liquidity. Yeah you have to pay taxes, but that's because you came out ahead.
I don't see anything other than a massive gamble. You've already staked enough of your future on one speculative start-up as an employee; why would you then put a big chunk of your own money at risk? An index fund has reliable long-term returns.
But you are right, it is yet another risk. At Twilio, the pay was awful but I felt the longer term risk/reward was worth it.
If I was with $startup and the strike price was $texas-sized, I wouldn't do it while the shares were still illiquid because executing would be so much.
IMO the gold standard here is to issue actual founder shares as long as possible (up to and possibly past series a) and then to do options with early exercise and extended validity, and of course complete transparency on all the numbers.
http://blog.detour.com/introducing-progressive-equity/
Early exercise for the vast majority then is something you have to know to ask for in addition to number of shares.
If I exercise an option with no spread, I am paying (at least) as much as it is presently valued.
He also discusses the need for a change in tax treatment by the IRS. One of the fundamental issues is how options are taxed. Should you exercise an option, you will need to pay taxes on the spread (delta of strike price and current FMV, i.e. latest 409A valuation).
In many cases, the spread is so small or nonexistent, that the tax bill is irrelevant. But, in a few cases it's so large that most people can't possibly raise the capital to cover the tax bill.
I think the fundamental issue is the definition of FMV. When there's no public market, and employees are covenanting away any rights to sell their equity on secondary markets, is there really a fair market? I would say no.
[0] http://blog.samaltman.com/employee-equity