AMT and Stock Options: 2026 Planning Guide for Tech Workers
I still remember the Slack message from a Coinbase engineer in early 2021: "I owe HOW much in taxes and I haven't sold a single share?" He'd exercised incentive stock options the year before, the stock had since dropped 60%, and he was staring down an AMT bill calculated on paper gains that had evaporated. That's not a hypothetical. That's Tuesday for tech employees who don't plan around the Alternative Minimum Tax. If you're holding ISOs heading into 2026, this is the year to get your AMT strategy right, before exercise decisions turn into tax regret.
Understanding AMT and Stock Option Triggers
The Alternative Minimum Tax is a parallel tax system. You calculate your regular tax, you calculate your AMT, and you pay whichever number is higher. It exists because Congress got tired of high earners using deductions and preferences to pay next to nothing back in the 1960s. Problem is, the mechanism built to catch wealthy tax avoiders also catches software engineers who did nothing more exotic than exercise their options early.
The part nobody warns you about is this. When you exercise incentive stock options and hold the shares (instead of doing a same-day sale), the spread between your strike price and the current fair market value counts as income for AMT purposes. Not for regular tax purposes. Just AMT. So you can owe real tax dollars on a gain you haven't actually realized, because you haven't sold anything yet.
Say you've got options with a $2 strike price, and the company's 409A valuation has climbed to $22. You exercise 10,000 shares. Regular tax code sees nothing taxable yet. AMT code sees $200,000 of "phantom income." Run that through the AMT rate structure (26% up to about $232,600 of AMT income in 2025, 28% above it, adjusted for inflation each year) and you're looking at a tax bill in the tens of thousands. On a stock you can't sell because it's private and illiquid.
This happened to employees at companies like Uber and Airbnb pre-IPO, and it's happening right now at every well-funded private company with a rising valuation. The AMT doesn't care that your shares are illiquid. It doesn't care that the stock might drop 70% before you can sell (see: nearly every 2021 IPO class marking to market in 2022). It cares about the spread on your exercise date. Full stop.
For a full breakdown of how ISOs work mechanically, our post on ISO stock options exercise strategies is a good place to start. Understanding the trigger is step one, though. The real game is in the planning.
ISO Exercise Strategies to Minimize AMT
You have more control here than you think. Most engineers treat their exercise date as an afterthought, something that happens whenever they get around to it or whenever a liquidity event forces the issue. That's a mistake. The timing and structure of your exercise is the single biggest lever you have to control AMT exposure.
One strategy: exercise early, when the spread between strike price and fair market value is small or nonexistent. This is why early exercise provisions (often paired with an 83(b) election) matter so much for founders and early employees. If you exercise the day you're granted, or shortly after, there's little to no spread, and therefore little to no AMT hit. I wrote a full guide on this in 83(b) elections and early exercise, worth reading before you sign any exercise paperwork.
Another approach: spread your exercises across multiple calendar years, exercising only enough each year to stay under or near the AMT exemption phase-out threshold. The AMT exemption for 2025 is $88,100 for single filers and $137,000 for married filing jointly, phasing out once your AMT income crosses roughly $626,350 (single) or $1,252,700 (married). Exercise too much in one year and you don't just pay AMT. You can lose your exemption entirely, which compounds the damage.
A third lever, and the one people forget: use a partial disqualifying disposition on purpose. If you exercise and sell some shares in the same calendar year (even if it's not your full position), that portion is taxed under regular income tax rules, not ISO/AMT rules. You give up favorable long-term capital gains treatment on those specific shares, but it can generate cash to cover the AMT bill on the shares you're holding. Sometimes that's the least-bad option in a year where you're exercising a big chunk of your grant.
None of these work alone. Strike price, current valuation, other income, filing status, state tax situation (California's added state AMT makes this whole calculus meaner for Bay Area employees) all factor in. This isn't a spreadsheet you build once. It's a projection you update every time your company's valuation moves or your income changes.
AMT Credit Planning and Carryforward Rules
Here's the part that gives people some relief once they understand it. AMT you pay on ISO exercises isn't gone forever. It generates a Minimum Tax Credit (MTC) you can carry forward and use in future years, when your regular tax liability exceeds your AMT liability.
Think of it as an interest-free loan you made to the IRS. Annoying, but not permanent.
The mechanics: in any year your regular tax exceeds your tentative AMT, you can apply some or all of your accumulated credit to reduce your regular tax bill, down to (but not below) your tentative minimum tax for that year. This typically happens in the year you sell the shares that triggered the AMT in the first place, since a qualifying disposition creates a big spike in regular taxable income (long-term capital gains) without a matching AMT preference item.
The catch, and it's a real one, is that there's no guarantee you'll use the full credit fast. One of our clients exercised a big batch of ISOs at a pre-IPO company back in 2021, then watched the stock drop 80% after it went public into a down market. He'll get his credit back eventually. But "eventually" has meant four years of carrying it forward while that cash sat with the Treasury instead of his brokerage account. The IRS has detailed guidance on Form 8801 (used to calculate and claim the credit) if you want to see exactly how the carryforward math works: IRS Instructions for Form 8801.
AMT credit planning isn't a file-it-and-forget-it exercise. Track your credit balance year over year. Project when you'll actually be able to use it. Factor that into decisions about when to sell shares or realize other income. We cover this kind of multi-year projection work in our broader tax planning strategies guide, but for concentrated stock positions specifically, it deserves its own model.
Timing Exercises Across Multiple Tax Years
Damn, this is where the real money gets made or lost.
Most people think about ISO exercises as a single event: I have options, I exercise them, done. Wrong mental model if you've got a large grant and multiple years before expiration. Typically that's ten years from grant date, or 90 days after leaving the company, whichever comes first. Know your window.
The better approach treats your option pool like a multi-year drawdown project. Model out exercising a portion each year, calibrated to keep your AMT income under the phase-out threshold, while watching your regular income too (salary, RSU vesting, spouse's income, bonus). RSU vesting cliffs matter a lot here. If you're at a company where a big chunk of RSUs vest in Q4, exercising ISOs that same quarter can push you into a much higher regular tax bracket and trigger a bigger AMT hit at the same time. Splitting the ISO exercise into a lower-income year can change the math substantially.
There's also a market-timing angle, and I want to be careful here. Nobody knows what a stock will do next year. But you do know your strike price and expiration date, and you know exercising when the spread is smaller costs less in AMT than exercising after a run-up. If your company's valuation has been climbing steadily (like a lot of AI infrastructure names have since 2023), waiting doesn't necessarily help you tax-wise. It can make the AMT bill worse even if the underlying investment thesis is still sound.
For anyone eyeing a 2026 IPO or expecting a liquidity event, this timing question gets urgent. Pre-IPO companies often see 409A valuations jump right before a public offering, and employees who wait to exercise "until it's real" sometimes get hit with a big AMT bill the same year, right as lockup restrictions keep them from selling anything to pay for it. If you're in that spot, have this conversation now. Not in December.
AMT vs. Regular Tax Calculation Examples
Two scenarios make this concrete fast.
Scenario one: the early-stage exerciser. Priya joins a Series B startup as employee 40. Strike price: $1.50. She exercises immediately upon grant, when fair market value is also roughly $1.50 (no spread). AMT preference item: basically zero. Four years later the company IPOs at $40 a share. She sells. Because she held over a year from exercise and two years from grant, the entire gain qualifies for long-term capital gains treatment. No AMT ever entered the picture, because there was no spread to tax when she exercised.
Scenario two: the wait-and-see exerciser. Marcus joins a similar company at the same strike price, $1.50, but waits four years to exercise, right before the IPO, when the 409A has climbed to $28. He exercises 20,000 shares. Spread: $26.50 per share, or $530,000 of AMT income. Depending on his other income and filing status, that can generate an AMT liability well into six figures, hitting in a year when the stock might still be under lockup, meaning he can't sell to cover the bill.
Same company. Same strike price. Wildly different tax outcomes, driven entirely by timing.
The AMT isn't punishing you for making money. It's punishing you for the gap between when you exercise and when fair market value was measured. Small gap, small tax. Big gap, potentially brutal tax. That's why "exercise early, exercise often, exercise in small doses" tends to beat "wait until I'm sure" as a general approach, though every situation has its own wrinkles: cash available to exercise, confidence in the company, personal risk tolerance.
None of this guarantees a better outcome, by the way. Stock prices can go down as easily as up, and exercising early means putting real cash into an illiquid, risky asset. This is a tax optimization discussion, not investment advice about whether you should hold concentrated stock in the first place.
Working with Tax Professionals on AMT
You can read every article on this topic, including this one, and still get the numbers wrong. AMT calculations depend on how dozens of variables specific to your return interact: state taxes, other income, timing of RSU vests, prior-year AMT credit balances, phase-out thresholds that shift with inflation adjustments each year. It adds up to something that really needs a projection, not a rule of thumb.
A good tax professional, ideally one who's actually handled ISO exercises for tech clients before, should be able to run a multi-year projection showing you exactly how many shares you can exercise in 2026 before you cross into meaningful AMT territory. They should also track your Minimum Tax Credit balance year over year so it doesn't just disappear into a filing cabinet somewhere.
A lot of generalist CPAs fall short here, honestly. Not because they're bad at their jobs, but because ISO/AMT planning for concentrated tech equity is a specialized skill, closer to actuarial modeling than tax prep. I've sat in on calls where a client's longtime accountant, sharp guy, great with small business returns, had genuinely never modeled a multi-year ISO exercise before. If your current preparer's answer to "how much should I exercise this year" is a shrug, find someone who's done this before.
We work with clients across financial planning engagements specifically on this kind of equity comp modeling, coordinating directly with tax preparers so the exercise strategy, the AMT credit tracking, and the broader portfolio picture pull in the same direction instead of fighting each other. If you want to see whether that kind of coordinated planning fits your situation, take a look at who we help or check our pricing to see how the engagement works.
The difference between Priya and Marcus up there wasn't luck. It was planning, done a year or two ahead of the deadline that actually mattered. If you're sitting on a meaningful ISO grant heading into 2026 and you don't have a multi-year exercise plan mapped out yet, now's the time, before another 409A valuation bump or vesting cliff makes the decision for you. Schedule a consultation and let's map out what your exercise strategy should actually look like.