Picture this. You exercised NSOs with a $280,000 spread last year. Withholding came back at 22%, the same rate your bonus gets withheld at, and you assumed that was the tax, paid in full. Then your CPA calls in March: you owe another $42,000.
That's the most expensive surprise non-qualified stock options have to offer. It's nowhere near the only one. NSOs let you participate in a stock's appreciation without committing capital until you choose to, but only if you plan accordingly. Skip the planning, and the bill still arrives. You just won't know how big it is until you file.
This guide walks through what NSOs are, how they're taxed, and the decisions you'll face with every exercise.
NSO 101: What You're Holding
When your company grants you NSOs, they're basically saying: today our stock is worth $X. We think it'll be worth more later. As a thank-you for helping grow the company, here's the right, but not the obligation, to buy it at today's price for the next several years.
A few terms worth pinning down before we go further, since the rest of the post depends on them.
Grant date. The day you receive the option. Nothing happens to you on this day, but the date does two things you'll care about later: it sets your strike price, and it starts the clock states use to decide how much of the eventual income belongs to them.
Strike (or exercise) price. What you pay per share when you exercise. Almost always set equal to the fair market value (FMV) on the grant date. For public companies, FMV is the market price. For private companies, FMV is the 409A valuation.
Vesting. Options don't become exercisable the day you receive them. They vest, usually in pieces over three or four years, and only as long as you're still working for the company. Leave before an option vests, and it disappears.
In-the-money vs. underwater. If current FMV is higher than your strike price, the option is in the money: you can exercise and capture a discount. If FMV is below your strike, the option is underwater and worth nothing in the moment, though it might recover before expiration.
Bargain element (or spread). When you exercise an in-the-money option, the difference between FMV at exercise and your strike price is the bargain element. It's the IRS's favorite part of this whole process.
The Exercise Math: Where the Surprises Live
Here's the working example we'll use the rest of the way: you exercise an option for 1,000 shares with a $20 strike when FMV is $120. The mechanics:
You pay $20,000 to buy the shares (1,000 × $20)
The bargain element is $100,000: 1,000 × ($120 − $20)
That $100,000 lands on your W-2 as compensation income, taxed exactly like salary
The IRS treats the bargain element as ordinary wages. Your employer has to withhold federal, state, local, Medicare, and (if you haven't blown past the wage cap) Social Security on it. For 2026, the Social Security wage base is $184,500, and most senior tech professionals blow past that early in the year, which usually takes Social Security withholding off the table for a mid- to late-year exercise. Medicare doesn't have a cap. The base 1.45% is withheld from all wages, and your employer starts withholding the additional Medicare tax of 0.9% on wages above $200,000, regardless of how you file.
Here's where most people lose track of the actual tax bill.
The withholding gap. Federal supplemental withholding is statutory: 22% on the first $1 million of supplemental income, 37% above that. That's withheld, not taxed. Your actual federal tax rate on the bargain element is whatever your marginal bracket is. For a senior tech professional with household income in the $400k–$800k range, that's usually 32%, 35%, or 37%. So on a $100,000 bargain element, federal withholding picks up $22,000, but the actual federal bill is $32,000–$37,000. The $10,000–$15,000 gap shows up in April. Multiply across multiple exercises in a year and the gap scales right along with them. And exercises aren't the only thing feeding it. The same 22% applies to your bonus and to every RSU vest, so by the time you exercise, the gap has usually been building since January across three separate line items nobody is totaling.
This is the single most common reason high-earning tech employees walk into March owing money they didn't budget for.
Back to the working example. To take possession of $120,000 worth of shares, you need to come up with about $60,300 in cash: $20,000 to buy the shares, plus roughly $40,300 in withholding ($22,000 federal supplemental, $11,700 New York State at 11.70%, $4,250 New York City at 4.25%, and $2,350 in Medicare and surtax). Those are New York's statutory supplemental rates; the same exercise in a state with no income tax would need about $16,000 less. They also run higher than New York's actual brackets. At 11.70% against the 6.85% bracket most households in this range land in, the state over-withholds by roughly $4,900 here, and the city adds a few hundred more. That comes back at filing. You still have to fund it at exercise, and it doesn't close the federal gap: it just makes April cost $5,000 to $10,000 on this exercise rather than $10,000 to $15,000. It's a great trade, if you have $60,000 in checking.
The Cashless Exercise (And What It Trades Away)
Most people who don't have $60,000 sitting around use a cashless exercise. It's elegant: the broker fronts the cash, sells just enough shares at the moment of exercise to cover the loan and the tax withholding, and you keep the rest. No money out of pocket.
In our example, the broker would sell about 503 shares at $120 to cover the $60,300 ($60,300 ÷ $120 ≈ 503 shares). You'd walk away with 497 shares: currently worth roughly $59,640. Net result: you converted an option into about $60k of stock without writing a check.
The tradeoff is straightforward. At New York rates, covering the exercise and the withholding costs slightly more than half the grant, and those 503 shares stop being yours at $120. If the stock doubles over the next two years, you doubled the 497 you kept, not the 1,000 you were granted.
That isn't an argument against the cashless exercise. It's what the alternative actually costs: funding it yourself means $60,300 out of cash reserves, or selling something else in a taxable account and paying tax on that instead. You're choosing which asset pays for the exercise, and the cashless version pays for it with the upside of the shares themselves. Which one is right depends on how much of your net worth is already in this stock, what the cash is otherwise for, and whether you'd buy this many shares at $120 if you were starting from nothing.
Private Company NSOs Are a Different Animal
Public-company NSOs are mostly a tax-and-cash-flow puzzle. The shares are liquid: you can always sell some to pay the bill. Private-company NSOs are something else entirely.
Your shares might be worth a fortune on paper and zero in actual cash. Most pre-IPO companies don't allow open-market sales, and many restrict private secondary sales as well. You exercise, you owe tax on the bargain element, and you have no way to sell shares to fund the tax. You write the check yourself, or you don't exercise.
The good news: at early-stage companies, the strike price and 409A are often close enough that the bargain element is small. Exercising 10,000 options at a $0.50 strike when the 409A is $0.75 means a $2,500 bargain element, almost trivially small in tax terms.
The interesting case is the early exercise. Many private companies let employees exercise options before they vest. If the strike price equals FMV at the time you exercise (typically the case if you exercise right after grant), the bargain element is zero, and so is the tax. You've converted future appreciation from ordinary income to potential long-term capital gains.
This is where the 83(b) election matters. When you early-exercise unvested shares, you have 30 days to file an 83(b) election with the IRS. The election locks in today's FMV as your basis and starts your long-term capital gains clock at the exercise date. Skip the election and the IRS treats each tranche as taxable when it vests, potentially years later, at much higher FMV, fully as ordinary income. The election isn't complicated, but the 30 days run from the exercise, not the grant, and there are no do-overs. Filing or not filing is the difference between having future appreciation taxed as capital gains or as ordinary income.
If the company qualifies, an early-exercised, 83(b)'d position can also stack with the QSBS exclusion. Under the new rules from the One Big Beautiful Bill Act, stock acquired after July 4, 2025 can shield up to $15 million (or 10x basis) of gain from federal tax: 100% with a five-year hold, with partial exclusions at three and four years. That's the kind of stacked-strategy outcome that doesn't happen by accident. It happens because someone modeled it before the exercise.
The catch: early exercise is a bet. You're paying real cash today for shares you might never be able to sell, in a company that might not make it, in exchange for tax savings that only matter if it does. Leave before vesting, and most plans force a buyback at the original strike price. Your money was tied up for nothing, and the tax payment was wasted. The lottery-ticket framing is honest: you're paying actual dollars today for an asymmetric bet on future equity value.
Sell or Hold? The Concentration Question
Once you exercise, your basis in the shares is the FMV at exercise. From that moment forward, future gain or loss is capital: short-term if you sell within a year, long-term after that.
Which raises the real question: should you sell the shares immediately, or hold?
The reflexive answer for most tech employees is to hold. The stock has been good to you. You believe in the company. Selling feels like betting against yourself.
Here's the math problem with that instinct. Your salary comes from this company. Your stock-based comp comes from this company. Your bonus comes from this company. If a substantial chunk of your investment portfolio sits in the same stock, then one company writes your paycheck, funds your retirement, and sets your net worth. Concentration risk is the polite term for it. The stock doesn't have to do anything dramatic for that to be a problem. A company priced for extraordinary results only has to deliver ordinary ones to underperform.
Once a single position runs past roughly 10% of liquid net worth, the conversation shifts from "should we trim?" to "how do we trim without handing a third of it to taxes?" The diversification toolkit at this level includes 10b5-1 plans, exchange funds for very large concentrated positions, and direct indexing for tax-efficient rebalancing, each its own conversation.
The deeper frame: the hold-or-sell decision isn't really about this exercise. It's about what each share is for. If the point of these shares is to make your financial life less dependent on this one company, every undiversified share works against that. Selling some isn't betting against your employer. It's the only way the shares can do the job you gave them.
The Wider Board
The hardest part of NSO planning isn't the math on any single exercise. It's that NSO decisions touch everything else: this year's tax bill, your cash flow for the next quarter, your estate plan, your charitable giving timing, your residency state at the moment you exercise.
A short list of decisions a real plan would coordinate around an NSO exercise:
Bracket management. If you're already pushing the 35% or 37% federal bracket, exercising more NSOs this year stacks on top. Splitting an exercise across two tax years can save real money if the math works out.
State tax mobility. States don't agree on how to source this income, and the two dominant methods land in different places: New York allocates on workdays between grant and vesting, California on workdays between grant and exercise. Either way it follows where you worked while you earned the option, not where you live on the day you exercise, so a move from California to a no-tax state part-way through doesn't erase the California share. The date you change residency drives all of it, which is why a move and an exercise in the same year are worth modeling together.
Charitable timing. Donating appreciated stock you've held longer than a year, instead of cash, gets you a deduction at FMV without realizing the gain. If you give regularly, coordinating gifts with NSO exercises is one of the cleanest ways to bring effective tax rates down.
Estimated taxes and the safe harbor. The withholding gap is a penalty problem, not just a cash problem. You avoid underpayment penalties by paying in 90% of this year's tax or 110% of last year's, and a large exercise can leave you below both without anyone noticing until the return is filed. Projecting your tax before each exercise guides your estimated payments to avoid penalties and unwanted surprises.
Coordination with ISOs and RSUs. If you hold NSOs alongside ISOs and RSUs, common at most tech companies, the NSO exercise doesn't land in isolation. An ISO exercise creates AMT exposure. The rest of your income from bonuses, RSU vests, and NSO exercises determines how many ISOs you can exercise before triggering AMT. What counts is the tax year everything lands in, not the order within it, which means the balancing can happen any time before December 31.
Each of these is a real lever. Most NSO holders pull approximately none of them: partly because nobody told them to, and partly because the levers only matter when they're coordinated across years and decisions. Pulling one in isolation usually doesn't move the bill. Pulling all of them, at the right moments, is where the real work lives.
Portfolio management is the visible piece. Coordinating tax, equity comp, estate, and cash flow against your specific goals, across years rather than at the moment you press the exercise button, is the rest of it.