Equity Compensation: RSUs, ISOs, and the Tech Wealth Engine

The Million Dollar Question: An employee exercises incentive stock options with a $2 million paper gain and sells nothing. What do they owe the IRS?
A) Nothing — no sale, no tax B) Ordinary income tax on the full $2 million C) Alternative minimum tax on the spread, potentially hundreds of thousands of dollars D) Payroll tax only

Read on for the answer.

Most first-generation technology wealth is not created by founders. It is created by employees who signed an offer letter, ignored the second page, and discovered years later that the second page was the part that mattered. This piece explains what those instruments actually are, when they get taxed, what they really cost, and why two people at the same company with the same paper gain can end up in completely different financial positions.

What it is

Equity compensation is pay in company stock rather than cash. That is the whole idea. Everything else — the acronyms, the schedules, the tax code sections — is detail about when the stock becomes yours, when the government taxes it, and whether you can turn it into money in between.

The scale is easy to underestimate. Alphabet recorded roughly $24.95 billion in stock-based compensation expense for 2025. Meta recorded about $20.4 billion, up more than 22% year over year. Nvidia disclosed that the total fair value of restricted stock units and performance stock units as of their vesting dates was $22.2 billion in the fiscal year ended January 25, 2026 — against $15.1 billion the prior year and $8.2 billion the year before that. That is not a perk line item. That is one of the largest ongoing private wealth transfers in the American economy, and it lands almost entirely on people who work for a living.

Four instruments do nearly all the work:

• Restricted stock units (RSUs). A promise of shares on a future date. Nothing to buy, nothing to exercise. When they vest, the value is ordinary income, taxed like salary.
• Incentive stock options (ISOs). The right to buy shares at a fixed strike price. Governed by Section 422 of the tax code, they get favorable treatment if you follow the holding rules — and they carry the alternative minimum tax problem that this article keeps returning to.
• Non-qualified stock options (NSOs). Same right to buy, none of the special treatment. The spread between strike and market value is ordinary income at exercise, with payroll tax attached.
• Employee stock purchase plans (ESPPs). A payroll-deduction plan that lets you buy company stock at a discount, capped at 15% under Section 423, with an annual ceiling of $25,000 measured in grant-date fair market value.

The single variable that separates these is the gap between the tax event and the cash event. RSUs put them close together. Options can put them years apart, in the wrong order.

Who uses it

“Tech workers” is too coarse a category to be useful here. Four distinct populations hold equity, and their situations barely resemble one another.

Public company employees. Someone at a large listed technology firm typically holds RSUs and nothing else. Their equity is liquid within days of vesting, priced by a public market, and taxed on a predictable schedule. This is the most benign version of the arrangement, and it is where the steady grind from a high salary into the $1M–$5M band actually happens. Levels.fyi puts the median software engineer total compensation at about $191,840 including equity — a good salary, not a fortune. It compounds into a fortune only across many years of grants in a rising stock.

Late-stage private company employees. These are the people at SpaceX, Stripe, OpenAI, Anthropic and the rest of the pre-IPO cohort. They may hold double-trigger RSUs, options, or both, at a company worth hundreds of billions on paper. Their wealth is real in a valuation sense and imaginary in a spending sense until a tender offer or a listing arrives.

Early startup employees. Option-heavy, strike-price-low, outcome-binary. The honest expectation for any single early grant is zero, because most startups do not produce a liquidity event. The rare exception is the one everyone hears about.

Founders and executives. A separate species. They hold restricted stock rather than options, file 83(b) elections at formation, and structure for qualified small business stock treatment from day one. Their tax planning starts before the company has revenue.

Note what is missing from this list: nobody reaches $100M+ purely on employee equity. Equity compensation is the engine of the $1M–$30M band. Above that, you generally have to have founded the thing, run it, or invested in it.

Why they use it

For the employer, the logic is straightforward and largely unromantic. Stock conserves cash, which matters enormously for a company that is growing faster than it is earning. It aligns behavior with the share price. And critically, it functions as a retention device: unvested equity is a hostage. An employee with two years of unvested grants and a refresh cycle ahead of them is expensive to poach, which is exactly the intended effect.

For the employee, the logic is arithmetic. On a W-2, taxed at ordinary rates, saving your way to seven figures from salary alone takes decades and unusual discipline. Equity changes the math in two ways: it gives you an asset that can appreciate independently of your labor, and — if held correctly — it converts a portion of your compensation from ordinary income into long-term capital gain. That is the difference between a top marginal rate in the high thirties plus state, and 20% plus the net investment income tax.

There is a third reason people rarely say out loud, which is optionality. A vested, liquid equity position buys leverage in the next negotiation, funds a sabbatical, or seeds a company. The money changes what you are willing to say no to. That effect shows up long before the balance sheet says “wealthy.”

How it works

Vesting. The standard American schedule is four years with a one-year cliff: nothing vests for twelve months, then 25% lands at once, then the remainder vests monthly or quarterly. Variations matter more than they look. A front-loaded schedule — some large employers vest 40% in year one — pays more early but leaves less on the table if you stay. A back-loaded schedule does the opposite and is a retention tool wearing a compensation costume.

Refresh grants. The initial grant is not the whole story. Most large employers issue annual “refreshers,” so that a long-tenured employee has overlapping four-year schedules vesting simultaneously. This is why the sixth year at a company can pay dramatically more than the first, and why the “equity cliff” in year four — when the initial grant runs out and the refresher is smaller — is a real and widely felt event.

Single- versus double-trigger RSUs. At a public company, RSUs vest on time and that is that. At a private company, RSUs that vested on time would create a tax bill on shares nobody can sell — a disaster. So private companies use double-trigger RSUs: they require both a time condition and a liquidity condition (an IPO or acquisition). Nothing is taxed until both fire. When they do fire, they fire all at once, which is why newly public companies see a large withholding event and a share-price wobble at the first lockup expiry.

Exercising options. Exercising is buying. You pay the strike price in cash to receive shares. Historically employees had 90 days after leaving to exercise or forfeit — a rule that quietly transferred a great deal of value back to companies, because departing employees frequently could not raise the cash. A number of employers now offer extended post-termination windows of up to ten years. It is worth asking about in an offer negotiation; it is one of the few equity terms that is occasionally negotiable.

The 83(b) election. If you receive restricted stock (not units) that is subject to vesting, you can elect to be taxed on its value at grant rather than as it vests. At formation, when the stock is worth almost nothing, this is close to free and starts the capital-gains clock immediately. The election must be filed with the IRS within 30 days of the transfer. Miss it and there is no remedy. It is the highest-value deadline in the whole category and the most commonly blown.

The ISO $100,000 rule. Only $100,000 of ISO grant-date value can become exercisable in any calendar year. Anything above that converts to NSO treatment automatically. Large grants therefore arrive as a mix whether the paperwork says so or not.

Secondary sales and tender offers. This is the structural change of the last decade. Companies stay private far longer, so liquidity has moved from the IPO to the company-run tender offer, in which the company or an investor buys shares directly from employees at a set price. SpaceX runs these roughly twice a year; its December 2025 round priced shares at $421 apiece at a valuation near $800 billion. OpenAI’s October 2025 tender moved about $6.6 billion at a $500 billion valuation, with more than 600 participants and roughly 75 employees selling the maximum $30 million each. Anthropic followed with a tender at a $350 billion pre-money valuation in 2026. Between them, employees of the two companies have sold around $14 billion in private shares without either firm having gone public.

What it costs

The cost of equity compensation is tax and risk, in that order.

RSUs. Value at vest is ordinary income. The cost is that you have no control over the timing — the tax event is the calendar’s decision, not yours. Most employers withhold at the 22% federal supplemental rate on the first $1 million of supplemental wages. If your actual marginal rate is 35% or 37%, that withholding is badly short, and the gap arrives as a surprise the following April. On a $400,000 vest, the shortfall can run well into five figures.

NSOs. Ordinary income on the spread at exercise, plus payroll taxes, plus the cash required to buy the shares. Two outflows, no proceeds, unless you do a same-day sale.

ISOs and the AMT. Here is the trap. Exercising an ISO produces no regular taxable income — but the spread between strike price and fair market value is a preference item for the alternative minimum tax. You can owe a substantial cash tax bill on a gain you have not received, in a stock you cannot sell. If the shares then fall, you have paid tax on a peak that never materialized. This destroyed a meaningful number of paper fortunes in 2000 and again in 2022.

2026 made this materially worse. Under the One Big Beautiful Bill Act, the AMT exemption phaseout thresholds reset to $500,000 for single filers and $1,000,000 for joint filers beginning in January 2026, down from the far higher post-2017 levels, and the phaseout rate doubled from 25% to 50%. The 2026 exemptions themselves are $90,100 for single filers and $140,200 for joint filers, and they now disappear twice as fast. The practical effect is that households in the $500,000–$1,000,000 income range — precisely the band where large ISO exercises happen — are far more exposed than they were in 2025 on an identical transaction.

The offsetting upside: QSBS. Working in the other direction, the same 2025 legislation expanded Section 1202 qualified small business stock relief. For stock issued after July 4, 2025, the per-issuer exclusion cap rose from $10 million to $15 million, indexed for inflation after 2026, and the old all-or-nothing five-year holding period became tiered: 50% of gain excluded at three years, 75% at four, 100% at five. For an early employee at a qualifying company who exercised cheaply and held, this is one of the most generous provisions in the tax code — potentially eight figures of gain, tax-free at the federal level.

Rough tiers. At the $1M–$5M band, most equity wealth is public-company RSUs, taxed as ordinary income, with tax planning that amounts to withholding correctly and selling on a schedule. At $5M–$30M, the mix shifts toward options and private shares, and the annual tax bill becomes something a professional manages. Above $30M, planning starts before the grant: QSBS stacking, exercise-and-hold decisions modeled years out, and charitable structures for concentrated appreciated stock. The advisory cost at that level typically runs from a few thousand dollars a year to well into six figures for a family office arrangement.

Hidden costs and tradeoffs

Concentration. This is the big one, and it is structural rather than optional. An employee with company RSUs has their salary, their equity, their career trajectory, and frequently their mortgage all levered to a single firm. When the stock falls, it usually falls in the same quarter the layoffs are announced. Diversification feels disloyal and looks like a bet against your employer, which is exactly why so few people do it on schedule.

The handcuffs actually work. Unvested equity distorts career decisions in ways people notice only in retrospect: staying eighteen months longer in a role that stopped teaching them anything, declining a better job to reach a cliff, accepting a bad manager because the refresher lands in March. The money is real. So is the cost of the years.

Illiquidity and the waiting. Private-company equity can sit unconvertible for the better part of a decade. Tender offers are periodic, capped, and sometimes priced below the last round. Third-party secondary marketplaces exist but typically transact at a discount, and many companies restrict transfers outright.

The cash to exercise. Options require money you may not have. Exercising a large early grant can mean writing a six-figure check for shares in a company that might not exist in three years — and, with ISOs, possibly a second check to cover the AMT.

The volatility, emotionally. Watching a paper position swing by hundreds of thousands of dollars in a quarter is a real psychological cost, and it tends to produce exactly the wrong behavior: holding through the top out of loyalty, then selling near the bottom out of exhaustion.

What people get wrong

“RSUs and options are basically the same thing.” They are opposites in the one way that counts. RSUs have no strike price and cannot be underwater — if the stock drops 60%, they are worth less, but they are worth something. Options with a strike above the market price are worth nothing at all. That is why mature companies grant RSUs and early-stage companies grant options.

“Exercising is when I get paid.” Exercising is when you pay. It is a purchase, often followed immediately by a tax bill. The payday is the sale, which may be years later or never.

“My strike price is what I’ll owe.” The strike is what you pay the company for the shares. The tax is calculated on the spread between the strike and the current fair market value — a separate, usually larger, number.

“We’re waiting for the IPO.” For most of the current large private companies, this framing is a decade out of date. The tender offer has become the primary liquidity mechanism, and it comes with its own rules: participation caps, blackout windows, and a price the company sets. SpaceX has said it is reserving up to 5% of shares in a planned offering for employees and friends, and OpenAI confidentially filed for an IPO in June 2026 — but employees at both had already been selling for years.

“I have 1% of the company.” One percent of what, after what? Preferred shareholders typically hold liquidation preferences that get paid first, and every subsequent funding round dilutes common stock. A grant described as 1% at hire can be a meaningfully smaller share of a much larger company by the time anything is sold — and in a modest exit, common stock can receive very little after the preference stack clears. The percentage is a headline; the capitalization table and the preference terms are the story.

Bottom line

The answer to the Million Dollar Question is C. Exercising incentive stock options generates no regular taxable income, but the spread between strike price and fair market value is an alternative minimum tax preference item in the year of exercise. Our hypothetical employee has received nothing, sold nothing, and may still owe a very large cash payment — and in 2026, with the phaseout thresholds reset to $500,000 and $1,000,000 and the phaseout rate doubled to 50%, the bill on that exercise is bigger than the identical exercise would have produced a year earlier.

Which is the honest summary of the whole category. Equity compensation is the most dependable route from a salary to seven figures available to someone who does not own a business — and the one most reliably mishandled, because the decisions that determine the outcome are made years before the money appears, in documents nobody reads at the moment they are signed. The people who do well are rarely the ones who picked the best company. They are the ones who understood the tax calendar, filed on time, and diversified on a schedule they set in advance rather than in the week the stock moved.

---

Related reading: Paths to Millions: How People Actually Get Rich · Tech Wealth: The Industry That Rewrote the Rules · Taxes: How the Wealthy Actually Pay · SpaceX, OpenAI, Anthropic, and the Next Gold Rush of Tech Wealth · Borrowing Against Wealth: The Buy, Borrow, Die Playbook