That 90-day expiry window is the most archaic bullshit. Can't come up with 200K cash? Too bad, and thanks for all that hard work and long hours you put in.
The way some companies get around it is that, after 90 days, they replace the expired ISOs with nonstatutory stock options which, as their name implies, are not recognized by the tax code. Tax code is complicated but NSOs are ultimately worth maybe 10%-20% less than "equivalent" ISOs.
But, at least until recently, there wasn't standard paperwork for doing this, so the only companies that did it were those willing to innovate on their stock plan. I think a lot of founders weren't trying to screw their employees, but their lawyers weren't comfortable going off the beaten path.
But I think that excuse is fading now. Enough companies have done it to provide precedent.
My circa-2015 startup looked at this and fretted a bit but ultimately sidestepped the problem by giving everyone restricted stock instead of options. That's only really possible at the very early stages when the 409(a) valuation is nearly zero. But for anyone joining a seed-stage startup today, that's what I strongly recommend: Ask for restricted stock, and if the purchase cost is non-zero, ask them to comp you for it as a signing bonus (which is "free" for them since they're buying the stock from themselves). You'll have to pay income tax on the value but that should be pretty small (maybe you can even get the company to "gross it up", but that's real dollars for them). Then you'll never have to worry about exercising stock options or paying AMT. Don't forget to file your 83(b) election within 30 days.
I'm at a company that gives NSOs. Can you explain what you mean when you say that NSOs are worth less than ISOs?
When you exercise an ISO, it is not considered regular taxable income. It is as if you legitimately purchased the stock on the market for that price. No tax is due until you sell. If you sell immediately then you pay regular income tax on the gain. But if you hold it for at least a year, then you pay long-term capital gains rate, which maxes out at 20% (vs. 37% for regular income tax).
When you exercise an NSO, the gain is considered income and is immediately taxable at regular income tax rates, whether you sell it or not. If the company is not public, then you can't sell share to pay the taxes, so if you're going to exercise NSOs, you'd better do it either before the valuation has gone up much (in which case you end up paying mostly long-term capital gains tax), or wait until after IPO (and pay regular income tax). Since investing in startups is very very risky, exercising early probably isn't the right choice for most regular people.
Extra complication: The US has two tax codes that exist in a sort of quantum superposition. Each taxpayer must pay whichever tax bill is larger of the two. The above describes the regular tax code, but there is also Alternative Minimum Tax (AMT). Under AMT, ISOs are not special; they are treated like NSOs. But, the maximum tax rate under AMT is 28% rather than 37%, and doesn't kick in until higher income levels. So if you exercise ISOs before the company has gone public, but after the valuation has raised significantly over the exercise price, you might again have trouble paying the taxes. But if you manage to pay them, and you manage to hold for a year, your overall gain will probably be 10%-20% more than what you would have gotten with NSOs.
Extra extra complication: Let's say you exercise your ISOs before IPO and you manage to pay the AMT. Years later, you sell the stock. Is your capital gains computed based on the ISO exercise price, or the valuation at exercise? BOTH! Under regular income tax, you owe capital gains on everything since the exercise price, but under AMT you already paid for the gains between the exercise price and the valuation at exercise. If you don't want to get double-taxed you have to learn what it means to take a credit for a timing-based AMT adjustment. This is well beyond the point where Turbo Tax gets pretty unhelpful and you probably need to get a CPA, but in the Bay Area there are so many rich clients that a decent CPA will charge thousands (maybe tens of thousands) of dollars to do your taxes. Have fun!
All that said, if your plan is to exercise-and-sell in one action sometime after IPO, then I think ISO vs. NSO doesn't make much difference.
I would still recommend folks to properly understand what it is and how it is calculated so they can double check the numbers, but that really applies to pretty much every tax form.
With this, it actually makes sense to exercise options and stay employed at the startup: you now may not be qualified as a short-term seller when you sell the stock, since you held on to it for a while.
I've seen it suggested that one should exercise on day 1. But that doesn't have to mean you pay for them immediately (after all, you don't get to keep them immediately - you have to wait for vesting...), so you work out a payment plan with deductions every paycheck so that you've pay fully for whatever is vested at the time it vests.
I've seen too many people exercise stock options trying to minimize tax implications down the road, only to see them eventually leave the company for any number of reasons and they end up get heavily diluted to almost nothing in future rounds. All that money basically down the drain trying to chase long term capital gains vs regular income tax rates when the likelihood that they'll make anything is extremely low.
Everyone's risk posture is different, but considering how few startups are actually successful, I would argue it's better to just assume the options are worthless and just deal with the regular income tax situation if you get lucky enough to be around during an IPO or acquisition.
This makes me curious: what's the total typically required to purchase one's options (not including the potential tax implications)? Does it vary wildly from a couple thousand to tens- or hundreds-of-thousands? Lottery ticket money, or significant fractions of annual income?