NSO exercise strategy: when to exercise, hold or sell, and the moves around it

What actually decides how much tax you pay on a pile of NSOs? Not the size of the grant. The size of the bill is set by how much spread you turn into income in any one tax year, and that is a decision you control. The spread between your strike and the share price is taxed as ordinary income the day you exercise, stacked on top of everything else you earn that year. Master that one fact and you have mastered NSOs, because every strategy below is a variation on keeping that spread from piling into your top bracket, on stock you actually wanted to own.

This is the long version on purpose, meant to be the only thing you need to read on the strategy side of NSOs: when to exercise, whether to hold or sell, how to spread a large grant, the pre-IPO exercise and its cash trap, the NSO-versus-ISO choice, how to fund the tax, and how to give the shares away. Come back to any single section when you are about to act.

A quick refresher, because everything hangs on it. An NSO, a non-qualified stock option, lets you buy company stock at a fixed strike price. When you exercise, the bargain element, the gap between strike and the market value that day, lands on your W-2 as wages and is taxed like salary. After that, the shares are just stock, with a cost basis equal to the value at exercise. Hold more than a year past exercise and the gain is long-term, taxed at the lower rate. Sell sooner and it is short-term, taxed like income. No AMT, no two-year-from-grant maze. The whole NSO game is timing.

The one lever: the calendar

Ordinary income is taxed in layers. Your salary fills the lower brackets, and the bargain element piles on top. The first slice of spread might be taxed at your current rate, but a large enough exercise climbs the ladder and the top of it gets taxed at the highest bracket you reach. That is the entire reason “when” matters so much more for NSOs than for almost any other equity.

So the goal is never to dodge the tax. The spread gets taxed as ordinary income either way. The goal is to keep as much of it as possible out of your top bracket, and to make sure the shares you end up holding are shares you actually want. Everything else is detail.

When to exercise: the windows worth using

The cheapest time to exercise is when the spread is small, your other income is low, or you can break the exercise into pieces. Four windows do most of the work, and three of them get their own section below: a low-income year (the single highest-return move), exercising early before a big run-up while the spread is small (the pre-IPO move), and slicing a large grant across several years. The fourth is an early exercise paired with an 83(b) election if your plan lets you buy unvested shares, which locks in today’s tiny spread before the stock climbs, on a hard 30-day deadline, covered in early exercise and 83(b).

Exercise-and-hold vs exercise-and-sell

Once you exercise, the next decision is whether to hold the shares or sell them. The honest answer is that holding chases a tax discount, and the discount rarely pays you back for the risk you take to get it.

You sell at or near exercise. There is little or no capital gain to tax because the price has barely moved from your basis. You walk away with cash, you have already paid the ordinary-income tax on the spread, and you carry zero single-stock risk. Clean and boring.

You keep the shares, hoping for a year-plus hold so future appreciation is taxed at the long-term rate. You get the tax discount only on the gain that happens after exercise, and only if the stock cooperates. In exchange, your money sits in one company’s stock the whole time.

This is where second-order thinking earns its keep. The long-term rate applies only to the gain after exercise, not to the spread you already paid full freight on. If the stock climbs 20% over your holding year, the discount applies to that 20%, not the whole position.

Now the risk side. To earn that discount you hold a concentrated, single-stock position for a year or more. One company. One earnings miss, one bad quarter, one sector rotation. I have watched people chase a modest tax saving and lose multiples of it when the stock dropped while they waited for the clock. That is the tax tail wagging the dog.

Two questions decide the hold honestly.

Exercising in a low-income year

When is the cheapest year to recognize a big spread? A year when your other income drops. The bracket ladder is the reason: your salary fills the lower rungs first, so in a year when the salary rung is half-empty, the spread has room to sit lower before it hits the top rates.

Four windows do this for you: a few months of unpaid or reduced-pay leave, the gap between leaving one job and starting the next, the early-retirement valley before pensions and withdrawals begin, and a thin year in a naturally lumpy income (founders, salespeople, anyone variable). In each, your salary rung is half-empty, so the spread has room to fill the lower brackets first. Just mind that a gap between jobs also starts the short post-termination clock to exercise.

Worked example: the same spread in a normal year vs a sabbatical year

Take a single filer holding $300,000 of NSO spread, using the confirmed 2026 single brackets 2026.

In a full-salary year, a $220,000 salary lands around $203,900 of taxable income after the standard deduction, which has already climbed past the top of the 24% bracket on its own. Pile the $300,000 spread on top and the whole spread is taxed starting from there, running through the 32% bracket (which for 2026 starts at $201,775 of taxable income) and deep into the 35% bracket (starting at $256,225). The spread sees none of the lower rates.

In a sabbatical year, half-year wages of about $110,000 land around $93,900 of taxable income, sitting in the 22% band (which ends at $105,700). Now the spread fills the rest of that 22% band, then the entire 24% band (up to $201,775), then 32% (to $256,225), and only the top portion reaches 35%. A large chunk of the same $300,000 gets taxed at 24% or less instead of 35%.

The exact dollar saving depends on the full return, but moving six figures of income from a 35% rate to a 24% rate is thousands of dollars, not a rounding error.

Spreading a large grant across years

What is the cheapest way to exercise a big pile of NSOs? Almost never all in one year. Take $300,000 of total spread: exercise it all at once and it lands as a single block on top of your salary, pushing the top of it into the highest brackets. Split it into three roughly equal exercises across three years, and each year absorbs a smaller block that may never reach those top rates. Same total exercise, lower average rate, just by using the calendar.

Building the plan is headroom math. Each year, figure out how much ordinary income you can absorb before you cross into a higher bracket, recognize roughly that much spread, and lay the chunks across the years you have, working backward from the option expiration and any post-termination window so the plan never outlives the options. If a low-income year is coming, load more into it. And fund each year’s tax, because the flat withholding falls short every time.

Exercising before an IPO

Should you exercise your NSOs before the company goes public? Only if you can afford to be wrong, because exercising early means paying real cash tax on stock you cannot sell yet. The appeal is honest: while the company is private the spread is small, so the tax is modest, and exercising starts the holding-period clock so future appreciation can be long-term capital gain. How small the spread is depends on the 409A valuation that fixes your strike; check that gap first, because it is the size of the tax you are choosing to trigger now instead of later.

Spread today is small, so the ordinary-income tax is small. The holding clock starts, so more future gain can be long-term capital gain. But you pay strike and tax in cash now, on illiquid stock, and you carry the full risk that the company stumbles or never lists. Best when the strike is low, the spread is small, and you can lose the money without it hurting.

No cash out of pocket until you can actually sell. You only exercise once there is a market for the shares, so the bet and the tax happen when the outcome is known. The cost is a larger spread, taxed as ordinary income, and less of the gain qualifying for long-term rates. Best when you are not willing or able to risk cash on a private, illiquid position.

Strip the decision to one question: can you lose every dollar you would put in, the strike and the tax, and be fine? If no, the early exercise is not for you, no matter how good the tax math looks on paper. If yes, size the bet to what you can truly afford to lose, not the whole grant. I have watched people get so focused on minimizing the tax that they tie up cash they needed elsewhere, in stock they cannot sell, right before a downturn. Decide on the risk first. Let the tax be the tiebreaker, never the driver.

When you can choose between an NSO and an ISO

Sometimes a plan lets you pick between an ISO and an NSO. The real question is not which is “better” in the abstract. It is whether you actually intend to hold the shares, because an ISO can turn the whole gain into long-term capital gain if you clear two years from grant and one year from exercise, but the spread it shelters from regular tax is fully counted by the alternative minimum tax, so an exercise-and-hold can hand you a bill on a paper gain. An NSO has no AMT angle and no holding-period reward, just ordinary income at exercise and rules you can reason about on a napkin.

You plan to exercise and hold, you can clear both holding periods, and you have the cash to carry both the strike and any AMT the exercise triggers. The reward is the long-term capital gains rate on the whole gain instead of ordinary rates on the spread. You are betting on the hold, so the hold has to be real.

You plan to sell at or near exercise, you do not want to manage an AMT bill or a separate AMT basis, or the grant is large enough that the ISO would spill over the limit anyway. You pay ordinary tax on the spread and you are done. No second tax system, no holding-period tripwire, no surprise bill on shares you never cashed.

There is also a structural limit that can make this choice for you. For 2026, only the first $100,000 of grant-date value that becomes exercisable in a year can get ISO treatment 2026; the excess is taxed as an NSO regardless of the label. So on a large grant, part of your award is effectively NSO no matter what you pick.

Funding the tax: closing the withholding gap

How do you keep any NSO exercise from blowing up your April? Close the withholding gap on purpose, in the same year you exercise. Your employer withholds at the flat supplemental rate, that rate sits below most high earners’ real marginal rate, and the difference is yours to pay. Pay it early as a quarterly estimate and the surprise never forms.

The fix is arithmetic plus a calendar. Multiply the full spread by your actual marginal rate (plus state, if any), subtract what was withheld at the flat 22%, and what is left is your gap. Move that gap into a separate account the day you exercise, before it gets spent, then send it to the IRS as an estimated payment in the quarter you exercised. Paying in the right quarter is what dodges the underpayment penalty; waiting until April pays the tax but not the penalty.

Giving the shares away: charity and family

Once you exercise, the shares are plain appreciated stock, and that opens two clean moves. The piece people miss is that the giving has to happen after exercise, not before, because the option itself is compensation and gifting it does not move the income off your return.

Donating to charity skips the gain entirely. Donate shares you have held more than a year and you generally deduct the full fair market value and never pay capital gains on the built-in appreciation. Exercise at $20, the stock is now $80, and you hand the long-term shares to a donor-advised fund or a charity: you skip the capital gains tax on the $60 of appreciation and take a deduction for the $80. Selling first and donating the cash would tax that $60 on the way. The exercise tax is unavoidable and donating later does not refund it. What the donation saves is the second layer, the capital gains tax on everything the stock earned after exercise.

Gifting to family moves the future growth. Here you are not chasing a deduction, you are moving the asset and its future appreciation off your own balance sheet. For 2026 you can give up to $19,000 per recipient per year 2026 with no gift tax and no use of your lifetime exemption; above that, the excess draws against the lifetime estate and gift exemption, $15,000,000 for 2026 before the 40% rate applies 2026. Give early in a company’s growth and you move the upside out of your estate at today’s value, not tomorrow’s.

What this means for you

The size of your NSO tax bill is not fixed, and the lever is the calendar. Recognize the spread when your other income is low, while the spread is still small, in slices rather than all at once, and you keep more of it out of your top bracket. Once the shares are yours, decide the concentration question before the tax question: holding past exercise is a bet on one stock dressed up as a tax move, and most of the time selling at exercise and diversifying beats the chase. Fund the gap the flat withholding leaves with a quarterly estimate in the right quarter. And if you are giving, give after exercise, so the appreciated stock does the work the option never could.

That is the whole arc: exercise on your schedule, hold only what you mean to hold, spread the big grants, fund the tax on time, and give from the shares, not the option. None of the pieces is hard. The mistake is treating a multi-year decision as a one-time click, on gains that sometimes no longer exist. If you are sitting on a meaningful grant and an exercise or a liquidity event is in view, a fit check is what that conversation is for.