Nobody Negotiates the Refresh, and That's Where Your Comp Dies

Your initial RSU grant is negotiated once and gone in four years. Your equity refresh grant decides what you earn every year after that. At a company with no real refresh program, year-five comp is structurally lower than year four no matter how well you performed, because the grant expired and nothing replaced it. The cliff isn't a failure. It's the default. And if you're already in the habit of treating equity like a number the company prices like cash, the refresh is the part of that number you never see coming.

So everyone bargains hard over a depreciating asset and stays silent on the one that compounds. You push the $300K sign-on grant to $350K, feel like you won, and never ask the question that actually decides the next decade of your pay: what does year five look like? That silence is the most expensive thing in your offer letter.

What actually happens to your equity refresh grant at year four?

A four-year grant divided by four is not a salary. It's a countdown. Each year a slice vests, and at the end the asset is exhausted. From that point, your equity line is whatever your refresh grants add up to. If those refreshes are small, or non-existent, your total comp drops.

How small? Sequoia found that in 2024, only 5% of non-executive refresh grants were worth more than 50% of a new hire's grant size, down from 49% in 2023. For executives, the same collapse: 6% in 2024, down from 48% in 2023. Read that again. A couple of years ago, roughly half of refresh grants matched or beat the new-hire benchmark. Now almost none do. The replacement equity that's supposed to catch you at the cliff has shrunk to a fraction of what brought you in.

It's not that companies stopped doing refreshes. 79% offer them to at least some non-executives, 81% to executives. They exist. They're just smaller than they used to be, and that's exactly the detail you can't see from the offer letter.

How does the math work for the person who asked?

Equity is not a side dish at the senior level. At senior engineer, it's 25 to 31% of total compensation, averaging around $125,000 a year, and it grew 42% year-over-year. When that quarter of your pay depends on a grant that's depleting, the refresh isn't a nice-to-have. It's the load-bearing wall.

Two engineers, same offer, different question.

Weak: Engineer A negotiates the initial $300K grant up by $50K and signs. Win.

Strong: Engineer B takes the same grant at par and asks, "Walk me through the annual refresh cycle for someone performing at target." She finds out there's no structured program, only ad-hoc retention grants. So she takes a different offer at a company that refreshes on a schedule.

By year five, Engineer A is collecting $350K split over four years, roughly $87.5K of equity that year, plus whatever ad-hoc refresh happened to materialize. Engineer B is stacking three overlapping refresh cohorts worth about $120K that year, and rising. The $50K Engineer A "won" is noise against a delta that widens every year. This is the whole Praxy worldview in one offer letter: compounding beats the one-time win, every time. It's the same trap as accepting a counter-offer that feels like a raise and bleeds out in slow motion, a number that flatters you today and costs you for years.

Which companies front-load the pain, and how?

The cliff isn't one shape. The vesting schedule decides when it bites, and the big names all do it differently.

Company	Initial grant vesting	What it does to year-by-year equity
Amazon	5% / 15% / 40% / 40%	Back-loaded. 80% of value lands in years 3 and 4, then drops off a shelf in year 5.
Google (Alphabet)	33% / 33% / 22% / 12%	Front-loaded. The grant shrinks every single year by design.
Wealthfront	25% of initial grant per year, refresh from year 2	Additive stack, then a drop-off in year five when the initial grant runs out.

Amazon's 5/15/40/40 is the one people misread. On a $400K grant you get roughly $20K, $60K, $160K, $160K. Comp feels like it's climbing. Then year five arrives, the initial grant is gone, and refreshes issued at maybe $100K a year on a slower vest can't fully cover the gap. Equity can fall from $160K to around $100K. A real cut, in the year your experience is most valuable.

Google's 33/33/22/12 inverts it. The same $400K pays about $132K, $132K, $88K, $48K. You can take a pay cut in years three and four from the initial grant alone, before you ever reach the post-year-four cliff, unless refreshes have stacked fast enough to fill the trough.

How much does one review cycle actually move?

At Meta, more than you'd guess. The IC5 refresh base target runs $175,000 to $200,000 a year, with a performance multiplier from 0% to 250%. A "Greatly Exceeds" rating applies roughly 1.75x and lands near $350K. "Meets Most" applies about 0.65x and lands around $115K to $130K.

That's a spread of more than $190K, decided every single year by where your manager lands you in the band. It's wider than most initial grants people fight over. One soft review cycle costs you more than a failed sign-on negotiation ever could. And it stacks: senior-level refresh stacking is what carries Meta E7 total comp to around $1,455,000 on Levels.fyi data, not the original grant. For context on where the ladder sits, median software-engineer total comp was $226K in 2025, with Staff at $457K. The gap between those rungs is mostly refresh.

This is consistency over intensity in raw dollars. The dramatic sign-on negotiation is a one-night event. The refresh rewards the boring, repeatable thing: hitting target, year after year, inside a program that actually pays for it. Same logic that governs how the promotion clock starts two quarters before the review you think decides it: the payout lands on work you did long before anyone wrote it down.

What are the five questions nobody asks before signing?

You don't have to be brave to ask these. You have to be specific. Put them to the recruiter or hiring manager before you sign:

1. Is there a structured annual refresh program, or are refreshes ad-hoc retention grants?
2. What's the target refresh amount for this level for someone performing at target?
3. What does a refresh look like for a strong performer versus someone meeting the bar?
4. How far ahead of grant expiry does the refresh cycle run, so I'm not exposed at year four?
5. Has the refresh program changed in the last two years?

That last one matters more now than it used to. 20% of surveyed companies plan to shift refresh vesting from time-based to performance-based in the next 12 to 24 months. Even a company with automatic refreshes today may bolt a performance gate onto it tomorrow. The mechanism you're joining isn't guaranteed to be the one you live under, which is exactly why you ask before you sign.

What can't you negotiate here, and what do you do with that?

Be honest about the limit. A recruiter genuinely cannot promise you a future grant. Refreshes are manager-discretionary and approved in annual cycles, so asking for a guaranteed refresh number is asking for something nobody can give. That request makes you look like you don't understand how equity works.

So here's the trade-off, named plainly. Asking about refresh policy is legitimate. Asking for a contractual refresh amount is not. You're not negotiating a term. You're gathering the information for a go/no-go decision. The answers tell you whether staying past year four is financially rational before you've spent four years finding out the hard way.

A few honest caveats. At FAANG with strong structured programs, the refresh is largely automatic for performers, so this matters less there than at a growth-stage or mid-market company. If you're early-career, getting the level right at hire is the bigger lever, because refresh amounts are set relative to level. And the cliff is worst where there's no program at all: only 30 to 40% of startups grant refreshers to long-term employees. Join a Series B on a $250K four-year grant, hit year five with no IPO and no refresh, and you do the same job for $0 of new equity while the company spends its talent budget on new hires negotiating fresh grants at market. Tenured people fall permanently behind. Not because they underperformed. Because nobody negotiated the replacement.

What to do now

Before you sign anything, run the five questions. Build the year-by-year equity waterfall for the offer in front of you: take the grant, apply the real vesting schedule, and look at what year five holds once it's exhausted. If the refresh answer is vague or "we handle that case by case," price that as the cut it probably is, and weigh it against an offer that pays on a schedule.

The initial grant is a known, depreciating asset the moment you sign. The refresh policy is the actual comp contract for years five through fifteen. Ask about it first. The candidate who does isn't being presumptuous. They're asking the only question that decides whether staying is worth it.

Sizing up an offer and not sure what year five really looks like? Send me the grant on WhatsApp and I'll help you build the equity waterfall and the five questions to send your recruiter before you sign.