You Read the RSU Number in Your Offer Letter. Then What?

A recruiter sends you an offer. It says $180K base, $20K bonus, $300K in RSUs over four years. You do the quick math: $300K divided by four is $75K a year. Total comp looks like $275K. You accept the job, you put in your notice, you feel good.

Then year three arrives and you feel $44K less good.

That initial math is not wrong. It just skips the vesting schedule, the tax withholding gap, the compensation cliff nobody warned you about, and the fact that "four years" means something different at Amazon than everywhere else. The engineers who sort this out before signing negotiate better because of it.

RSUs Are Shares, Not Bonus Cash

An RSU (Restricted Stock Unit) is a promise that your employer will give you shares of company stock when certain conditions are met. At a public company those conditions are almost always just time. You stay employed, the shares vest, you own them outright.

The word "promise" is doing a lot of work in that sentence.

The key difference from a bonus: RSUs are compensation denominated in stock price, so the number in your offer letter fluctuates with the market. That $300K grant could be worth $350K or $220K by the time you're collecting it, depending on where the stock goes. You're being paid partly in a volatile asset issued by the same company you work for, whose performance you're rated on. This is called alignment of incentives. It is also called concentration risk.

Once shares vest, you own them like any other stock. You can sell immediately, hold for long-term capital gains treatment, or split the difference.

RSU Vesting Schedules: When the Money Actually Shows Up

Most tech companies vest RSUs over four years with quarterly or monthly distributions. A simple 25-per-year schedule with quarterly payouts gives you one-sixteenth of your grant every three months. Uniform. Predictable. Boring. Not everyone does this.

RSU vesting schedule comparison: Google 33-33-22-12, Amazon 5-15-40-40, standard 25-25-25-25

Google front-loads. Their standard new-hire schedule is 33-33-22-12: 33% in year one, 33% in year two, 22% in year three, 12% in year four. Your first two years look excellent. Year three looks like a pay cut. Year four looks like the company is slowly walking away from you while making eye contact.

Amazon back-loads. Their schedule is 5-15-40-40: almost nothing in years one and two, then the full weight of your equity in years three and four. Amazon compensates for this with cash sign-on bonuses (often substantial in year one, smaller in year two). The RSUs kick in right when your sign-on money has run out and you've been there long enough to have opinions about the codebase. The retention pressure is exactly where Amazon wants it.

Standard uniform vesting is mostly everywhere else: Meta, Apple, Netflix. Your grant splits evenly, roughly $75K worth per year on a $300K grant.

For more on how these companies structure the full offer and interview process, the Amazon software engineer interview guide and Google software engineer interview guide cover what to expect end to end.

The Cliff: A Gate Before the Money Starts

A vesting cliff is a minimum service period before any shares vest at all. The standard cliff in tech is one year. Nothing vests in months one through eleven. On your one-year anniversary, the first full chunk drops at once. If you leave on day 364, you leave with zero equity regardless of how close you are.

Cliffs exist to protect the company from paying out equity to people who leave in the first few months. From the employee side, they mean you are fully at risk until that anniversary. Both parties understand the arrangement. Neither party is being coy about it.

Major public companies have been moving away from cliffs. Google, Meta, and others increasingly vest on monthly or quarterly schedules with no cliff for at least some grant types. Startups almost universally keep the one-year cliff. Worth asking about explicitly before signing.

The Compensation Cliff Nobody Warns You About

Most engineers only discover this one during year three, and they discover it personally.

Your initial RSU grant is a four-year pool that winds down. But your salary grows, your performance reviews happen, your market value increases. Without refresh grants to compensate, your total comp can actually drop in year three or four even as your base salary goes up.

At Google, a new hire who receives $400K in RSUs is collecting $132K worth in years one and two, then $88K in year three, then $48K in year four. That is a $44K drop in RSU income in a single year, even if base salary went up $20K. You can do the net-of-tax math on that.

This is not a bug in the system. It is a retention mechanism. The company wants you to either earn a refresh grant through strong performance, get promoted with a new grant attached, or make a decision. Year three is when that decision happens naturally.

This is why engineers at well-paying companies often see total comp peak around year two to three and then need active management to sustain it.

Refresh Grants: The Compensation Stabilizer

A refresh grant is a new RSU award given to an existing employee, with its own four-year schedule. By years three and four at most major tech companies you're accumulating overlapping grants simultaneously: the original new-hire grant winding down, one or two refreshers mid-schedule, possibly a promotion grant beginning. The combined vesting of these overlapping pools creates something closer to a stable monthly floor.

Refresh grants are not guaranteed, and they are not in your offer letter. They vary significantly between companies.

Meta is widely considered the most generous, with formulaic refreshers tied to performance rating and level (E7 engineers can receive up to $400K annually). Google weights your last two performance review scores. Amazon folds refreshers into a compensation model that assumes 15% annual stock growth and counts that appreciation against new grants, which can work against you. Startups, especially early-stage, often provide no refresh grants at all. Only 30-40% have a consistent program.

If you're joining a company, ask specifically what the refresh policy is. Formulaic or discretionary? What happens at your level? Get concrete numbers if you can.

The Tax Trap That Catches Engineers Every Year

April. Every year. Thousands of engineers.

When RSUs vest, the fair market value of those shares is taxed as ordinary income. Your employer withholds taxes automatically at vest through sell-to-cover, where they sell a portion of your vesting shares to pay the withholding.

The federal withholding rate on RSUs defaults to 22%, the IRS supplemental wage rate. If you're in the 32% or 37% marginal bracket (which most senior tech employees are), that leaves 10 to 15 percentage points of federal tax un-withheld. Add state taxes in California (up to 13.3%) or New York (up to 10.9%) and the gap compounds.

Tax withholding gap diagram showing $44K withheld vs $74K owed on a $200K vest event, with $30K surprise balance due in April

On a $200,000 vest event, 22% federal withholding is $44,000. Your actual federal obligation at a 37% marginal rate is $74,000. That is a $30,000 gap. Due in April. On top of any state shortfall. The IRS will not be moved by the argument that you thought the company was handling it.

The fix is either quarterly estimated tax payments throughout the year, or adjusting withholding on your regular paycheck to compensate. Most engineers learn this exactly once.

After shares vest, you have a decision. Sell-to-cover gives you cash minus taxes. Same-day sale converts everything immediately. Holding means you now have a stock position: any appreciation is taxed as capital gains on sale (short-term at ordinary income rates if held less than a year, long-term at 0-20% if held more).

Holding is not obviously better. It adds concentration risk on a position you didn't choose to open. Plenty of engineers held through a 40% drawdown because they thought of RSUs as free money and had no sell plan. "I'll sell when it goes back up" is a sentiment, not a strategy.

How to Actually Value an Offer

The useful number for comparing across companies is annualized total compensation: base salary plus target cash bonus plus annualized equity value.

For equity, divide the four-year grant by four. Simple. But also think about the vesting structure. A 5-15-40-40 schedule means your year-one equity income is vastly different from a 33-33-22-12 or a flat 25%. If you're comparing two offers and planning to stay four years at each, the totals are the same. If you might leave in two years, the schedules are very much not the same.

Don't count refresh grants when evaluating an offer. You haven't earned them yet. They're real and they matter for long-term planning, but they're contingent on performance and company policy, neither of which you control at signing.

Also account for stock price risk. A $300K RSU grant at today's share price might be worth $180K or $450K when it vests. For volatile or early-stage companies, discount the stated value accordingly when making comparisons.

Levels.fyi is the practical benchmark tool. Use it to compare total compensation at your level across companies, including real refresh data from community submissions.

Pre-IPO Equity: A Different Game

If you're joining a late-stage private company, the equity mechanics change in ways that matter.

Standard single-trigger RSUs at a private company create a tax-before-liquidity problem. The shares vest, they're taxed as income at fair market value, and you can't sell them because there's no public market. You owe cash taxes on stock you cannot convert to cash. This is as fun as it sounds.

Well-run private companies solve this with double-trigger RSUs, which require both time-based vesting and a liquidity event (IPO or acquisition). No liquidity event means no vesting, no tax event. Facebook used this structure before its IPO. Many late-stage startups do too. Ask explicitly.

If you're looking at options rather than RSUs, the valuation math gets more complex. Your strike price is the 409A valuation, the IRS-mandated fair market value of common stock. The 409A is typically 25-60% of what investors paid for preferred shares, because common stock lacks the liquidation preferences preferred shares carry.

In any acquisition or wind-down, preferred shareholders get paid first. If the company's liquidation preference stack totals $50M and the acquisition price is $40M, common shareholders get nothing. The headline says "company acquired for $40M." Your bank account says otherwise. This happens more often than the startup success stories suggest.

Companies expect senior candidates to come in informed. That means knowing your equity terms before the offer conversation, and walking into the technical rounds ready. SpaceComplexity runs voice-based mock interviews with rubric-based feedback, so you're not figuring out how to think out loud under pressure for the first time on interview day.

The Practical Checklist

Before signing an offer with significant equity, get clear on:

Vesting schedule: Uniform quarterly, front-loaded, or back-loaded? Any cliff?
Refresh policy: Formulaic or discretionary? Triggered by performance review? Any data on typical refresher sizes at your level?
Grant type: RSUs at a public company, double-trigger RSUs at a late-stage startup, or options? What's the 409A vs. preferred spread for options?
Tax plan: What's your effective marginal rate and what will the withholding gap be at your expected vest events? Do you need to make estimated payments?
Sell strategy: Will you hold, sell-to-cover, or same-day sell? What's your concentration limit in a single stock?
Year 3 model: Sketch out your projected total comp in years one through four accounting for vesting schedule and realistic refresh estimates.