ISOs and the AMT: the complete guide
Exercising and holding ISOs can hand you a cash tax bill on a gain you never sold, and this is the whole story of how that happens and how to plan around it.
ISOs · Taxation
Why do so many smart people get blindsided by a tax bill on stock they never sold? Because incentive stock options come with a tax break the regular system gives you and the alternative minimum tax quietly takes back, and almost nobody sees the second half coming. The AMT is the ISO surprise. This guide is the whole picture, start to finish: how the bill gets created, how big it gets, how you usually get most of it back, and how to plan the entire arc so it never lands on shares you cannot sell to pay it.
This is the long version on purpose, and it is meant to be the only thing you need to read on ISOs and the AMT. Every piece is here: the two tax systems, the trigger, the exact calculation, the second cost basis, the credit and how to track it and how to pull it home faster, the breakeven, the private-stock disaster, the state bill, and the charitable move that softens it. Read it top to bottom the first time. Come back to any single section when you are about to act.
The one sentence that explains the trap
When you exercise an ISO and hold the shares, regular tax sees nothing happen. The AMT sees income. That single disagreement is the source of nearly every ISO tax surprise there is.
You did not sell anything. You may not want to sell anything. You may not be able to sell anything. The AMT does not care. It taxes the paper gain on the day you exercise, in cash, due next April. People hear “tax break” and stop reading. The break is real, and so is the catch riding underneath it, and the catch is the thing that hurts.
Here is the good news I want you to hold onto through all of this: the AMT is one of the few tax problems you can measure to the dollar before you act. It feels like an ambush. It is actually one of the most predictable bills in the tax code. The people who get hammered are the ones who never bothered to measure it. By the end of this guide you will know how.
A quick map of where we are headed, because it is a longer road than most ISO explainers admit. First, the two tax systems and why they disagree about your exercise. Then the bargain element that triggers the bill, the exact calculation with the 2026 numbers, and a worked example you can follow line by line. After that, the parts that play out years later: the second cost basis your shares quietly pick up, and the credit that returns most of the cash. Then the planning, which is where the money is actually made or lost: how much to exercise, when in the year, the sell-or-hold fork, the cash trap, the worst-case private-stock version, and the second state bill that catches Californians. The arc is exercise, hold, sell, recover, and each section below takes one stop on it in turn.
The two tax systems you live in at once
The United States runs two income tax systems side by side, and you owe whichever one asks for more.
The regular tax system is the one you know: brackets, deductions, the number on your pay stub. The alternative minimum tax, the AMT, is a parallel calculation that strips out certain breaks to make sure high earners with lots of deductions still pay a floor. Most years, for most people, the regular tax comes out higher, so the AMT never enters your life. Exercising and holding ISOs is one of the few things that flips that, because the two systems flatly disagree about whether your exercise even happened.
In an exercise-and-hold year, the ISO bargain element is income for the AMT side only, so the tentative minimum tax usually comes out higher. The gap is the AMT you owe on top.
You do not get to pick a system. Your tax software runs both calculations and you pay the larger result. A big ISO exercise lifts the AMT side above your regular tax, and that gap is your AMT for the year. Below a certain line, the exercise costs you nothing extra, because the regular tax is still the higher of the two. Above it, the AMT takes over and starts billing you. That line is the whole game, and we will get to exactly where it sits.
The bargain element is the variable you control. Exercise a small block and stay under the exemption, and the two systems land in roughly the same place. Exercise a huge block in one year and the AMT result towers over your regular tax, and the distance between the two towers is the check you write. That is worth sitting with, because it reframes the whole problem from “will I owe AMT” to “how much spread do I want to recognize this year.” You are not at the mercy of the tax code here. You are choosing a number.
For reference, this is the regular side the AMT is being measured against: the ordinary brackets your salary and other income climb.
Step height tracks the rate. The label under each step is the top of that bracket, and only the dollars that land inside a bracket are taxed at its rate.
One reframe to carry through the rest of this guide: the AMT on an ISO exercise is mostly timing, not a permanent surcharge. Regular tax eventually catches the same gain when you sell, and to avoid taxing it twice the system hands most of the AMT back as a credit in later years. So the danger was never that the money is set on fire. The danger is owing the cash early, on a paper gain, sometimes on shares you cannot sell to fund the bill. Hold that thought. We come back to the credit in detail below.
How exercise-and-hold pulls the trigger
The thing that triggers the AMT is not exercising. It is exercising and holding. The difference matters, so let me be precise.
The trigger is the bargain element: the value of your shares at exercise minus what you paid to exercise them, your strike price. Exercise 10,000 shares with a $2 strike when the shares are worth $12, and your bargain element is $100,000. Regular tax looks at that $100,000 and shrugs. That is the ISO benefit working as designed. The AMT looks at the same $100,000 and counts it as income for the year. Nothing about your shares changed. You sold none of them. But one of the two systems now has a much bigger number to tax.
Why does holding do it? Because if you sell the shares in the same year you exercise, that is a disqualifying disposition: the spread becomes ordinary income on the regular side, and the AMT adjustment generally washes out. The AMT is the toll you pay for holding the shares to chase the lower long-term rate. You hold to convert the spread into long-term capital gains, and the AMT is the risk you carry on the road there. That is the whole reason a same-year sale is the one move that makes the AMT preference disappear, which becomes the escape hatch we use more than once below.
The trap inside the trap
Exercise and hold a large pre-IPO grant, owe AMT on the spread, and then watch the stock stall or fall. The shares might be illiquid or worth less than your tax bill. You still owe the AMT, in cash. I have watched this wreck good plans. It is the failure mode I warn people about most, and it gets its own section later in this guide.
The AMT calculation, step by step
The AMT runs its own income calculation, subtracts an exemption, then applies its own rate to what is left. Three numbers do most of the work, and here are the confirmed 2026 figures for each.
The exemption is the slice of AMT income you get to subtract before the tax bites. For 2026 it is $140,200 for married filing jointly 2026, $90,100 for single filers 2026, and $70,100 for married filing separately 2026. If your AMT income lands under your exemption, the AMT has little or nothing to tax. This is why small exercises so often slip through untouched.
The exemption does not last forever. Once your AMT income climbs past a threshold, the exemption starts shrinking, and a smaller exemption means a bigger bill. For 2026 the phase-out begins at $1,000,000 of AMT income for married filing jointly 2026, $500,000 for single filers 2026, and $500,000 for married filing separately 2026. The exemption is gone entirely at $1,280,400 joint 2026, $680,200 single 2026, and $640,200 married filing separately 2026.
The rate applies to what is left after the exemption. For 2026 it is 26% on AMT income up to $244,500, and 28% on the amount above that, with the breakpoint halved to $122,250 for married filing separately 2026. The bargain element from a big exercise can easily reach into that 28% band, so as the exercise grows, the marginal cost of each extra share of spread climbs from 26 cents on the dollar toward 28, and faster than that once the phase-out starts clawing back your exemption on top.
So the order of operations is simple: build your AMT income, subtract whatever exemption survives at that income level, tax the rest at 26% and then 28%, and compare the result to your regular tax. You owe the AMT only to the extent it comes out higher. If your regular tax already exceeds the AMT result, the exercise triggers no AMT at all, which is the entire reason the size of the exercise is something you steer rather than suffer.
The double hit nobody warns you about
A large exercise does two damaging things in one move. It adds the bargain element to your AMT income, and that same income can push you into the phase-out, which shrinks the exemption that was protecting you. You get taxed on more and shielded on less at the same time. This is why people so often see the AMT bill come in far higher than they expected. They knew about the exemption. They missed the phase-out eating it.
How do I know if I am near the phase-out?
Add your regular income for the year to the bargain element from any exercise you are planning. If the total is approaching the phase-out start for your filing status, $1,000,000 of AMT income joint or $500,000 single for 2026, the exemption is already eroding and each additional share you exercise costs more AMT than the last. That is the signal to slow down or split the exercise across years.
A worked example, with the math shown
Let me run real numbers so the steps stop being abstract. The share counts and prices here are made up to keep the arithmetic clean. The exemption and the rates are the confirmed 2026 ones.
Show the math
Suppose you exercise 10,000 ISOs at a $2 strike when the fair value is $20. You file jointly, you have a normal salary, and you exercise and hold.
Step 1: the bargain element. Value at exercise minus strike, times shares: ($20 minus $2) times 10,000, which is $180,000. That is the income the AMT counts and regular tax ignores.
Step 2: build your AMT income. Start from your regular taxable income for the year, then add the $180,000 bargain element on top. Say your regular taxable income before the exercise is $200,000. Your AMT income is roughly $380,000.
Step 3: subtract the exemption. At $380,000 of AMT income, a joint filer is still well below the $1,000,000 phase-out start, so the full $140,200 exemption 2026 applies. $380,000 minus $140,200 leaves $239,800.
Step 4: apply the AMT rate. The whole $239,800 sits under the $244,500 breakpoint, so it is all taxed at 26% 2026. That is about $62,350. This is your tentative minimum tax.
Step 5: compare to your regular tax. You owe the AMT only to the extent the tentative minimum tax tops your regular tax. If your regular federal tax on that $200,000 of income is, say, around $35,000, then the AMT is higher by roughly $27,000. That $27,000 gap is the extra cash this exercise costs you this year, due next April, on shares you have not sold.
Change one input and the answer moves. A higher fair value means a bigger spread and a bigger bill. More other income can push you toward the phase-out and shrink your exemption. This is exactly why you run your own numbers before you write the check, not after.
Two refinements the napkin version skips. Long-term capital gains and qualified dividends keep their preferential 0, 15, or 20% rates inside the AMT, but they still raise your AMT income and can erode your exemption. And the real Form 6251 layers in other adjustments beyond the ISO spread. Treat the worked example as the shape of the calculation, not a filed return. To size your own before you act, use the ISO AMT estimator.
The dual-basis problem: your shares carry two cost numbers
Here is the quiet sequel to the AMT bill, and it bites careful people years later. The day you pay AMT on an exercise, you create a second cost basis the IRS never reminds you about.
Equal to your strike price, the cash you actually paid. The regular system never taxed the spread at exercise, so it does not let you add the spread to basis.
Equal to the value at exercise, which is strike plus the bargain element. The AMT already taxed that spread, so it lets you count it as cost. This number is higher.
Why does the second number matter? Because at sale you compute your gain twice, once for each system, and gain is sale price minus basis. The higher AMT basis means a smaller AMT gain at sale, smaller by exactly the spread you already paid AMT on. That is the mechanism that keeps the AMT from taxing the same appreciation a second time.
The most common ISO sale mistake I see
Someone pays a big AMT bill at exercise, holds for years, sells, and then files using only the strike-price basis the broker reported on the 1099-B. They pay full tax on the entire gain a second time and never claim the higher AMT basis or the credit it helps unlock. The money is gone unless they amend. Two cost numbers, and they used the wrong one. Your broker will not fix this for you.
Keep your Form 3921 from the exercise year and your own log of the bargain element, because the AMT basis will not show up on any broker statement later. The 1099-B is not wrong, it is incomplete: the broker reports the cash you paid, not the spread the AMT already taxed, and tax software that reads only the 1099-B will quietly overcharge you. The basis adjustment on the AMT side of your return is yours to make, from your own records.
What if I already filed and missed the AMT basis?
You may be able to amend the return and recover the overpaid tax, within the normal time limits for amending. The money is not automatically lost, but nobody hands it back unless you ask. If the AMT you paid at exercise was large, the amount at stake on the sale can be large too, so this is worth chasing rather than writing off.
Does a disqualifying disposition change the dual-basis problem?
Yes. If you sell early and disqualify, the AMT adjustment for that exercise generally washes out, because the regular system now taxes the spread as ordinary income instead. The dual-basis double tax mostly bites the qualifying sale, where you paid AMT at exercise and then held the shares.
The AMT credit: you usually get most of it back
This is the reframe that changes how the whole thing should feel. The AMT you pay on an ISO exercise is not a permanent extra tax. Most of it is timing.
When you exercise and hold, the part of the AMT driven by the ISO timing difference becomes a minimum tax credit. It runs on Form 8801, it lives in IRC Section 53, and it carries forward indefinitely with no expiration. In any later year where your regular tax runs higher than your tentative minimum tax, the credit pays down the difference. The cash leaves early and comes home over time.
Prepayment, not penalty
Think of the ISO-driven AMT as money you fronted the Treasury. The minimum tax credit is the mechanism that gives it back. The catch is that you only recover it in years where your regular tax exceeds your AMT, so the pace is not always fast, and it is not always recovered in full if you keep yourself in AMT territory year after year.
So how long does it take? Exactly as long as it takes the gap between your regular tax and your AMT to open up, because every year you recover credit equal to that gap and not a dollar more. Wide gap, fast refund. Narrow gap, slow drip. Zero gap, nothing comes back that year and the rest rolls forward untouched. The question “when do I get my AMT back” is really “how big is the spread between my two tax calculations each year,” and you can read that off your own return.
What opens the gap wide? A big salary, a bonus, ordinary income with no fresh AMT preferences. The sale year is often the single best year for recovery, and most people miss why: your AMT basis is higher than your regular basis by the bargain element you already got taxed on, so in the sale year that higher basis shrinks your AMT, which widens the gap, which lets more credit flow home. What drags it out is the mirror image. Stacking another large exercise on top feeds the AMT again. A sabbatical or early retirement lowers your ordinary income, and your regular tax falls with it toward your AMT until the gap closes. A move to a high-tax state pushes your AMT up year after year on heavy state-tax add-backs.
The sabbatical that strands the credit
I have watched someone exercise, owe a heavy AMT bill, then leave their job to start something new. Their ordinary income fell, their regular tax fell with it, and the gap that recovers the credit closed. The credit was real. It just sat there for years while their income was low. Plan the exercise around the income years that actually let it come back, not just the year the stock looks good.
This is the cleanest argument there is for staging your exercises instead of doing one giant one. Size each year’s exercise so you drop back under the AMT line afterward and you turn the credit on. Pile it all into one year and you can owe a large bill that then trickles back over a long stretch while inflation quietly eats it. Same total tax, very different timeline, and a dollar back in three years beats the same dollar back in twelve.
Track the credit, or the Treasury keeps it
People leave AMT credit on the table for one boring reason: nobody tells them they have it. The credit carries forward quietly, and if you do not track it, you forget to claim it. A one-page log fixes that. Build it the year you first pay AMT, not the year you finally remember it exists.
Year and credit generated
The year you exercised and paid AMT, and the credit it created. The timing-difference portion of that AMT, the part driven by the bargain element, is your starting credit balance. Pull it from the AMT forms for that year.
Credit used this year
In any later year your regular tax exceeds your tentative minimum tax, you can use some credit. Log the amount you claimed. Many years it is small, sometimes zero. That is normal.
Running balance carried forward
Starting balance minus what you used. This number rolls to next year. It is the single figure people lose track of, and it can be five or six figures after a big exercise.
AMT cost basis still outstanding
Track your separate AMT basis on the shares in the same log, since the basis and the credit unwind together. When you sell, that higher basis cuts your AMT gain and helps release the credit. Forget it and you overpay at sale.
The credit runs on Form 8801 and does not expire, so a balance you log this year stays claimable until it is fully recovered. After a large one-time exercise it can take a decade or more to drain, depending on your income and whether you exercise again, which is exactly why a running log beats memory. You are managing a balance that outlives most of the tax returns around it.
Pull the credit home faster
Tracking is half the job. The other half is making the credit flow back instead of idling. You do not speed up the credit by asking for it. You speed it up by widening the gap between your regular tax and your AMT in a given year, because every dollar of that gap is a dollar of credit that can come home.
The sharpest lever is a planned sale in a low-AMT year. When you sell ISO shares you exercised and held, your AMT basis is higher than your regular basis by the spread you already paid AMT on, so a higher AMT basis means a smaller AMT gain, which lowers your AMT that year, which opens room for the credit. A planned sale does double duty: it diversifies a concentrated position, and it widens the regular-versus-AMT gap that releases the credit.
Find your unused credit
Pull your last filed return and locate the minimum tax credit carryforward. That is the pool you are draining. If nobody ever tracked it, that is the first leak to fix.
Map your low-AMT years
Look ahead for years where your other income is lower or your AMT exposure is light. A job change, a sabbatical, or a low-bonus year are all candidates. Those are the windows where a sale or a recovery actually pays you back.
Size a sale that stays under the AMT line
Sell enough ISO shares to use the higher AMT basis and shrink your AMT gain, but not so much that the sale itself pushes you back into heavy AMT. The goal is a wide regular-versus-AMT gap, not a new AMT bill.
Confirm the sale qualifies before you trigger it
A qualifying sale, more than two years from grant and more than one year from exercise, keeps the gain at the long-term capital gains rate. Selling early disqualifies the shares and changes the whole calculation, so check both clocks first.
Recover the credit on the return
In the lower-AMT year, the credit pays down the gap between your regular tax and your AMT. Make sure whoever prepares the return actually claims it. This is the step people skip.
Can I recover the credit without selling anything?
Sometimes. Any year your regular tax simply runs above your AMT, the credit comes back on its own, no sale required. The selling strategy is for people whose income stays high enough that the gap never opens by itself. If your income naturally swings, you may just need to remember to claim the credit in the down years.
What if I never get back into regular-tax territory?
Then the credit can carry forward unused, sometimes for a long time, and the cost is the time value of money: a credit you recover a decade from now is worth less than the cash you paid today, even though the dollar figure matches. This is why I treat the AMT as cash tied up rather than cash gone, and why a plan that drops you back under the AMT line in later years beats waiting passively.
Planning the exercise year
Everything above sets up the move that actually matters: deciding how much to exercise, and when. The plan is simpler than it sounds.
There is a point each year where your AMT exactly equals your regular tax. Below it, exercising ISOs costs you nothing extra. Above it, every additional share adds to a bill due next April. That gap is your AMT-free room, and it resets every January. The whole plan is to find the room, fill it, stop, and repeat next year.
Find your breakeven for the year
Estimate your regular tax with no exercise, then find the amount of bargain element that lifts your AMT up to meet it. That dollar figure is your AMT-free room. Divide it by the per-share spread and you get roughly how many shares you can exercise and hold for free this year. Leave a margin, since values move.
Exercise to the line, not past it
Buy enough ISOs to use the room and not one share more. A position too big for a single year is a position you spread across several. Recompute the line every year, because your income, your deductions, and the exemption all shift.
Exercise early in the calendar year
Exercise in January and you get eleven months to watch the stock before the bill is locked. If it craters, you can sell the same shares in the same year, disqualify on purpose, and undo the AMT exposure before it ever lands on a return. Exercise in December and you have surrendered that escape hatch. Same shares, very different optionality, and the timing trick only fully works for public stock you could actually sell.
Start the clock early where the spread is thin
At a low valuation, the bargain element is small, so the AMT cost is small. Exercising soon after grant, with a timely 83(b) election where it applies, can start both holding clocks while the spread is tiny. The earlier you go, the cheaper the tax and the higher the risk.
The second-order effect everyone misses: staging is not only about staying under the AMT line. When you exercise and hold a private company’s shares, you wire cash into a single illiquid stock and start a tax clock on top of it. Staging the exercise stages the bet. You learn more about the company each year, and you risk less capital while you learn.
The breakeven controls your tax, not your risk
Staying under the AMT line is a tax win, but the shares are still a concentrated bet. Exercising AMT-free into a stock that then drops still leaves you holding a stock that dropped. Find your line and respect it, and then weigh the investment risk separately. The breakeven tells you how much spread is free this year. It does not tell you whether owning this much of one company is a good idea.
The disposition fork: qualifying vs disqualifying
Once you hold the shares, the next decision is when to sell, and it splits into two outcomes that tax completely differently.
You hold more than two years from the grant date and more than one year from the exercise date. Clear both clocks and your entire gain over the strike is a long-term capital gain at the lower rate. This is the outcome the whole ISO structure is built to reach. The cost is time: you stay invested in one stock to earn it. If you paid AMT at exercise, the qualifying sale is often where your higher AMT basis and your credit start coming back to you.
You sell before clearing both clocks, even by a day. The bargain element becomes ordinary income, taxed like salary, and only the rest is capital gain. You give up the ISO rate advantage on the spread. But a same-year disqualifying sale generally pulls the spread out of the AMT entirely, so it can erase an AMT problem on a gain that is shrinking. The “worse” treatment is sometimes the better outcome.
The two clocks are statutory and they run separately: two years from grant, one year from exercise, and the later of the two is your real finish line. Mark both dates the day you exercise.
Here is the part the “never disqualify” crowd misses. The qualifying rate is the prize on paper, but the path to it carries AMT risk and concentration risk. When the stock turns against you, the slower, lower-rate route can cost more than the fast, higher-rate one. Picture exercising and holding a stock that then starts to fall. Hold, and you owe AMT on a paper gain shrinking by the day. Sell in the same year, disqualify on purpose, and you pay ordinary income on the smaller realized gain while killing the AMT exposure. The “worse” tax treatment can be the better outcome. Holding a concentrated position past sense because selling triggers a higher rate feels like patience. Often it is just concentration risk wearing a tax costume, and the market does not care that you were one month from the lower rate.
Disqualifying on purpose is a tool, not a mistake, and you reach for it when diversification, liquidity, or a soured outlook matters more than the tax break. This is second-order thinking in practice: decide on the money first and let the tax fall out of that, not the other way around. You also do not have to pick one fate for the whole position. Sell enough now to cut your concentration to a level you can sleep with, even if those shares disqualify, and hold the rest for the qualifying rate.
Exercise-and-hold vs exercise-and-sell, and the cash trap
Step back from the tax for a second, because the bigger decision is whether to hold at all.
Holding chases the qualifying long-term rate, and the cost is time spent owning a single stock plus exposure to the AMT on the spread. Selling kills both the AMT risk and the concentration, and the price is a higher tax rate on the gain. Everyone runs the tax math on this. Few run the risk math, and the risk math is the one that decides it.
The tax tail wagging the dog
I see people hold a volatile single stock for a year purely to convert ordinary income into capital gains. Then the stock drops more than the tax they were trying to save, and they are worse off after optimizing than if they had just sold and paid. The tax saving is real. So is the chance the stock erases it. This is the principle I keep coming back to: I do not manage money for people who can live forever. A lower rate years from now is worth less than it looks if the path there runs through a stock that can cut your position in half.
There is a cash trap hiding in here too. If holding triggers AMT and you would have to sell shares to pay it, some of the holding logic unravels before you start. Know the bill before you commit to the hold. A cashless exercise solves the funding problem by selling shares the same day, but selling the same day usually disqualifies the shares and converts the ISO into something taxed like ordinary wages. Convenient, and quietly expensive. There is a middle path, sell-to-cover, that sells only enough shares to pay the strike and the tax and keeps the rest, but the shares it sells still disqualify. One exercise, two tax fates, so track which shares are which.
Three questions decide the hold honestly. Run them before you let a tax rate make the call.
How big is this position relative to everything you own?
If these shares are a small slice of your net worth, holding for the rate is a smaller risk. If they are most of it, the concentration usually outweighs the tax saving, and selling to diversify is the stronger move. The tax break on a qualifying sale is small next to the risk of riding one position through a long holding period.
How volatile and how liquid is the stock?
A steady, liquid public stock is one thing. A single high-flying name, or a private stock you cannot sell at all, is another. The more the stock can swing and the harder it is to exit, the more selling earns its keep, because the qualifying clock can force you to hold straight through real danger.
Can you cover the AMT without selling?
If holding triggers AMT and you would have to sell shares anyway to pay it, the holding logic unravels. The honest comparison is not qualifying rate versus disqualifying rate. It is the after-tax dollars from selling now versus the probability-weighted after-tax dollars from holding a concentrated, possibly illiquid position for another year. Framed that way, holding for the rate often looks worse than it does on a spreadsheet.
The worst trap: AMT on illiquid private stock
If I could get you to remember one danger from this entire guide, it would be this one. The single worst ISO mistake you can make is owing a real, cash tax bill on shares you cannot sell to pay it.
That is the private-stock AMT trap. You exercise a hot private company, the AMT counts a fortune you have not received, and there is no market to turn those shares into the cash the IRS wants. With public stock you can at least sell some shares to cover the bill. With private stock, often you cannot sell a single one. The bill is set by a 409A valuation, the company’s official private-share price, and you are funding it entirely from savings.
It gets worse. You can pay AMT on a high 409A value, and then the company stalls, raises a down round, or never goes public. Now you have paid a large tax on a gain that shrank or vanished, holding shares you still cannot sell. You took on concentrated, illiquid risk and prepaid the tax on the optimistic case. I have a bias against illiquid assets, and this is exactly why. The higher AMT basis and the credit are real consolation, but a credit you recover slowly is poor comfort against cash you needed today.
An acquisition does not automatically rescue you, either. If the deal is cash and it closes, you finally get the liquidity to cover the tax and walk away. If it is stock-for-stock, or it drags, or it falls through after you already exercised, you can be left holding the AMT and no cash all over again. Exercising in anticipation of a deal that has not closed is a bet, not a plan, and the difference between the two is whether the money to pay the tax actually shows up.
Never fund an exercise with money you cannot lose
Treat the strike and the AMT as money that could go to zero with the company. If losing it would touch your security, the position is too big. Run the AMT on the current 409A spread before you exercise a private grant, exercise early while the spread is thin, stage it to keep each year small, and assume no liquidity until a real one exists. Counting on selling private shares to fund the tax is the assumption that fails people.
What if I cannot pay the AMT I already triggered?
This is the emergency the trap creates. If any liquidity exists, a tender or an approved secondary, a same-year disqualifying sale can unwind the exercise’s AMT preference before the year closes. Otherwise you are covering the bill from other resources. There is no clean fix once the year ends and the cash is not there, which is the entire reason to model the bill before you exercise, not after.
Is there ever a way to sell private shares to cover it?
Sometimes, through a company-approved secondary sale or tender offer, but these are not guaranteed and many companies restrict them. Counting on selling private shares to fund the tax is exactly the assumption that fails people. Assume no liquidity until a real one exists.
To see this play out start to finish, the case study the AMT trap when the stock crashed walks through someone who owed AMT at the peak and then watched the shares fall 70%, and what saved him.
State AMT: the second bill
You can clear the federal AMT and still owe a separate state alternative minimum tax on the same bargain element. Two bills, one exercise, and most people budget for one.
Only four states impose an individual AMT in 2026: California, Colorado, Connecticut, and Minnesota 2026. For the vast majority of equity holders, the rest of the country has none. California is the one most people in tech run into, because so much equity is granted there. California runs its AMT at a 7% rate 2026 on income above its own exemption, lower than the top federal AMT rate but applied to the same spread, and it has its own minimum tax credit you recover separately from the federal one. The California exemption is indexed each year and shields a chunk of AMT income before the 7% rate applies, smaller than the federal exemption.
The exact 2026 California exemption is not published yet
California indexes its AMT exemption each year, so it shields a slice of AMT income before the 7% rate applies, smaller than the federal exemption. The Franchise Tax Board has not yet published the exact 2026 California figure, so treat the California exemption qualitatively until the FTB releases it.
The state bill stacks, and so does the state credit
People budget for the federal AMT and forget the state runs a parallel system. The California bill is smaller than the federal one, but it is real cash, due on the same paper gain, in the same April. Size both before you exercise. The good news is the mirror image: a state AMT you pay can generate a state minimum tax credit you recover in later years, separate from the federal credit, with its own rules and its own paperwork. Tracking two credits is tedious, and money gets left behind when people track neither.
The mobility wrinkle worth knowing
California sources a capital gain to where you live when you sell, so a former California resident who moved to a no-tax state before selling held ISO shares pays no California tax on that gain, including a qualifying-disposition gain. California does not recharacterize the exercise spread as compensation the way New York does. The catch sits in the exercise year: if you exercised while a California resident, you triggered the California AMT that year, and that exercise-year AMT stands even after you move. Leaving California before the sale can drop the tax on the gain entirely while the exercise-year AMT remains. The mechanics turn on residency dates, so get them checked for your facts.
The charitable move: bunch giving into your exercise year
Can a charitable gift cancel out the AMT from an exercise? Partly, and only if you pick the right deduction. The AMT strips out most of the deductions you are used to. State and local taxes, gone. Miscellaneous deductions, gone. That is half the reason a big exercise tips high earners in high-tax states into the AMT in the first place. The charitable deduction is the rare one the AMT still lets you keep, which makes it the natural lever in a year you are already exercising and holding ISOs.
Lower your AMT income and you lower your AMT. A charitable contribution reduces your taxable income on the regular side and your alternative minimum taxable income on the AMT side, because the AMT does not add it back the way it adds back your state taxes. So in a year where the ISO spread has pushed you into the AMT, a gift pulls your AMT income back down and trims the bill.
Here is the second-order move. If you give every year anyway, doing it in dribs and drabs can waste the deduction once the standard deduction is in play, because small annual gifts may not clear the bar to itemize. Bunch several years of intended giving into your exercise year and two things happen at once: you clear the itemizing threshold, and you land the whole deduction in the exact year your AMT is highest, where it does the most work. A donor-advised fund is the usual tool. You contribute a lump in the exercise year, take the deduction now, and grant the money out to charities over the following years on your own schedule.
Confirm the exercise year actually puts you in AMT
The pairing only helps if you owe AMT. Size your exercise first and check whether you cross the line. If you are exercising under your AMT-free room, there is no AMT to offset and no reason to force a gift.
Decide how much you would give over the next several years
Add up your realistic giving for the coming years. That total, not a number invented for the deduction, is what you consider bunching into this year.
Consider gifting appreciated stock, not cash
Donating long-term appreciated shares can hand you the deduction and skip the capital gains tax on the built-in gain. For a concentrated equity holder, this doubles as a way to trim the position you are overexposed to.
Run the gift through the AMT math before you commit
Model the exercise with and without the gift, on both tax systems, so you see how much AMT the deduction actually removes. A deduction inside the AMT is worth less per dollar than under regular tax, so the offset is partial, not full.
Do not let the deduction drive the gift
A charitable deduction inside the AMT does not erase the bill dollar for dollar. It reduces your AMT income, and the saving is roughly the AMT rate on that reduction, so each dollar of deduction inside the AMT is worth about 26 to 28 cents for 2026 2026. Giving an extra $100,000 to save AMT on part of it leaves you poorer, not richer. Bunching only pays off if the bunched total clears the standard deduction, which for 2026 is $32,200 married filing jointly and $16,100 for single filers 2026. Give what you were going to give, then time it well. Generosity that also lowers a tax bill is a good deal. Generosity manufactured to lower a tax bill is just spending.
Filing and tools
When the forms arrive, two of them decide whether your exercise turns into a surprise. Form 3921 is the receipt your employer sends for the exercise year, reporting your strike, the value at exercise, and the share count. Form 6251 is where that receipt becomes a number on your return, where the bargain element gets added to your AMT income and where you find out whether the AMT is your tax for the year. No withholding happens on an exercise-and-hold, so if you never run Form 6251, you can owe the AMT and not learn it until the return is due.
Form 6251 is also the form to run, or have a preparer run, before you commit cash, because it layers in any other AMT adjustments beyond the ISO spread and tells you the real number. At sale, the same ISO sale shows up on Form 8949 in different versions depending on whether it was qualifying, disqualifying, or being run through the AMT, and the basis adjustment is the line that stops you paying tax twice. To size the bill before you act, use the ISO AMT estimator.
What this means for you
The AMT is not a trap if you can see it coming, and you can. It punishes the surprise exercise and rewards the planned one. Measure your room, exercise into it early in the year, spread a large position across years, claim your AMT basis at sale, track the credit until it comes home, and never let an AMT bill land on stock you cannot sell to fund it. Do that and the tax that ambushes everyone else becomes a line item you chose on purpose.
That is the whole arc: exercise, hold, sell, recover. None of the individual pieces is hard. The mistake is treating a multi-year plan as a one-time event, and it is a mistake that costs real cash on gains that sometimes no longer exist.
This is also exactly the kind of decision worth pressure-testing before you write the check, because the numbers are large, the moving parts are many, and the worst version of this is the one you cannot undo after the year closes. 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.
More in ISOs
Still have questions about your equity?
Join the community to ask directly, or take the two-minute fit check to see if a planning call makes sense.