VP Engineering Compensation in San Francisco

There used to be a fairly simple answer to the question "what should a VP of Engineering make in San Francisco?" A decade ago, the range was perhaps $400,000 to $700,000 in total compensation, the structure was relatively uniform across company types, and the meaningful negotiation happened at the edges. The single biggest predictor of a VPE’s package was company stage: a Series C company paid one number, a public company paid another, and the gap between them was maybe 40%.

That market is gone. In 2026, the same person — same résumé, same skill set, same target geography — can receive offers that range from $480,000 to over $2 million in total compensation, depending entirely on which kind of company is hiring. The packages are no longer different by degree; they are different by structure, and the differences are large enough that "VPE compensation in SF" is not really a single market anymore. It’s at least four different markets that happen to share a job title.

This piece is built on 47 VP Engineering placements we closed in the Bay Area between January 2025 and the first week of April 2026. Every number you’re about to read comes from a real signed offer letter for a real engineering leader who is now doing the job. The packages are reported gross, with equity valued at the most recent 409A for private companies and the trailing 30-day average closing price for public companies. The breakdown by company type tells the story more clearly than any aggregate average could.

The headline: a 3x range for the same job

Across all 47 Bay Area VPE placements in our 2025–2026 dataset, median total compensation at grant was $820,000. The 25th percentile was $525,000. The 75th percentile was $1.4 million. The single highest total comp in our entire VPE dataset was $2.3 million at grant.

That headline median is the number most engineering recruiters and most published benchmarks will quote when asked about VPE compensation in San Francisco. It is also the least useful number we could give you. The variance is the entire story, and the variance has a specific structural cause.

Take a moment with that table. Public-company post-IPO VPE compensation, in real dollar terms, is now lower than it was in 2022. AI-native company VPE compensation, in the same time frame, has roughly doubled. The same job title, the same labor pool, two and a half times the dollar range between the bottom and top of the structure. The gap isn’t closing — it’s widening.

Why this matters practically: a senior engineering leader navigating the Bay Area job market in 2026 needs to make a decision before they negotiate the offer, which is which segment of this market they want to be in. The negotiation happens within a segment. The segment determines the absolute level. The traditional career-management advice — "find the job that fits, negotiate hard" — misses that the most important variable is segment selection, not negotiation skill.

Big Tech / FAANG-equivalent

Public Big Tech VPE roles are interesting because the compensation structure is highly standardized. Nearly every major Big Tech company has internal compensation bands tied to leveling systems, and a VPE typically lands somewhere in L8–L9 territory (or the equivalent at companies with different leveling). The base salary is functionally capped by the band; the equity component is where the variability lives.

Our 8 Big Tech VPE placements in 2025–2026 had the following structure:

The single most important thing to understand about a Big Tech VPE package: the initial grant is not what you’re really being paid. By year three or four, the cumulative refresh grants will exceed the initial grant in dollar terms, and they’ll determine the actual realized compensation more than the headline at signing. A VPE who joins at year one with a $2.6 million initial grant will, by year four, typically be sitting on $4 million to $7 million of unvested equity if the company stock has performed at all reasonably. The refresh policy is where the real money lives.

Refresh policies vary more than candidates expect. Some companies have explicit policies (e.g., guaranteed refresh of 30% of initial grant value each year starting year 2). Some have discretionary policies tied to performance ratings. Some have nothing in writing and depend entirely on whether your manager pushes for you at the annual comp cycle. We always ask the refresh question explicitly in final-round negotiations — the answers reveal more about how the company actually thinks about senior retention than almost any other question.

The other thing to understand about Big Tech VPE packages: realized comp is heavily stock-price-dependent, and the recent past has been volatile. A VPE who joined a major public tech company in early 2022, near the post-COVID stock peak, signed an offer letter where the equity portion was sized at the then-current share price. By the time their first vest tranche hit in 2023, the stock had declined 35% to 50% at most companies. Their actual realized year-1 compensation was significantly below the headline number. The reverse has been true for VPEs who joined at trough valuations in 2023; their packages have grown 20% to 40% in realized dollar terms as stocks recovered. The headline at signing is one input; the trajectory of the stock determines the rest.

Late-stage private SaaS

Our 13 late-stage private SaaS VPE placements were our largest single sub-segment in 2025. The structure looks different from Big Tech because the equity isn’t liquid, the bonus structure is typically simpler, and the math depends entirely on the eventual exit.

A typical late-stage private SaaS VPE package in 2025 looks roughly like this:

• Base salary: $350,000 to $400,000
• Annual bonus: 20% to 30% of base, often discretionary based on engineering team performance and company OKRs
• Initial equity grant: 0.15% to 0.40% of fully diluted shares, vesting over 4 years with 1-year cliff
• Refresh policy: typically negotiable, increasingly granted annually starting year 2 at 25% to 50% of initial grant size
• Sign-on: historically uncommon, increasingly granted — we negotiated sign-on bonuses in 7 of 13 deals in our 2025 dataset

The math on the equity component is what determines whether a late-stage SaaS VPE role is a good deal. Take a company valued at $1.5 billion at the most recent round, granting a 0.25% equity stake to the incoming VPE. The face value at signing is $3.75 million. If the company eventually exits at $3 billion (a 2x return for late-stage investors), the equity is worth $7.5 million at exit, minus tax (typically capital gains given the long hold). If the company stays private and never exits, the equity is worth zero in liquid terms, with limited tender opportunities along the way.

That binary outcome is the key thing late-stage SaaS VPE candidates need to internalize. Of the 11 late-stage SaaS companies where we placed VPE candidates between 2021 and 2023, only 3 have had a liquidity event by mid-2026. Two were acquired, one IPO’d. The other 8 are still privately held, in some cases with valuations meaningfully below where the candidate’s grant was sized. The grant on signing day was a paper number; the realized comp has varied from 4x the original headline to roughly zero.

Late-stage SaaS VPE compensation has been roughly flat over the past two years in our data. Median total comp at grant of $680,000 in 2025 is essentially the same as $695,000 in 2023 (the slight decline is roughly within the noise of sample variation). The combination of harder capital markets, slower growth, and lower exit valuations has compressed what late-stage SaaS companies are willing to pay. The candidates who did well in 2021–2022 SaaS placements were the ones who joined companies with strong subsequent exits; the candidates who joined comparable companies with weak subsequent trajectories effectively took pay cuts in real terms.

Public, post-IPO software

The most challenging segment in our 2025 dataset. 11 placements at public post-IPO software companies, ranging from companies 3 years post-IPO to companies 15+ years post-IPO. The median total comp of $520,000 is materially lower than the broader VPE median of $820,000, and the year-over-year change versus 2022 is the only meaningfully negative number in our entire dataset.

The reasons are mechanical, not narrative. Public-company refresh grants are sized in dollar terms based on stock price at the time of grant. From 2022 through early 2024, software stocks broadly declined 30% to 60% depending on the company. New equity grants in 2023–2024 were sized at lower share prices, meaning the same target dollar value translated into a larger share count, but the realized dollar value over the vesting period has depended on stock recovery — which has been uneven. A VPE joining a public SaaS company in early 2024 may have signed an offer letter with $400,000 of equity value, only to see the actual realized dollar value of that grant fall further as the stock continued to decline.

The structure of post-IPO software packages also constrains negotiation leverage. Public-company comp committees are constrained by proxy advisor guidance and shareholder votes on executive compensation. The bands for VP-level engineering roles are typically narrower than at private companies, and the variability available in negotiation is limited. We’ve had multiple final-round negotiations at public software companies in 2025 where the comp committee explicitly stated that increasing the equity grant by even 10% would require committee approval that wouldn’t fit within the offer timeline.

The realistic comp scenario for a VPE joining a mid-tier public SaaS company in 2026 is:

• Base salary $340,000 to $380,000
• Cash bonus 25% to 35% of base, paid on annual metrics
• Initial equity grant $400,000 to $700,000 of RSUs, vesting over 4 years
• Annual refresh grants of $80,000 to $250,000 starting year 2
• Sign-on cash bonus of $50,000 to $150,000, usually with a 12-month clawback

Total comp at grant: $520,000 to $680,000. Realized comp over a 4-year tenure varies meaningfully based on stock-price performance.

AI-native: the new top of the market

The single most disruptive force in our VPE compensation data over the past 18 months has been AI-native companies. We placed 9 VPE-level hires at AI-native companies in 2025–2026, and the median total comp of $1.4 million is more than double what we saw at comparable Series C companies just three years ago.

"AI-native" is a category we’ve used internally since 2023 to describe a specific kind of company: foundation-model labs, AI infrastructure companies, and applied-AI companies where AI is the core intellectual property rather than a feature on top of an existing product. The list of companies that meet this definition has grown rapidly — from perhaps 15 in early 2023 to over 80 today — but the talent pool that can credibly lead engineering at these companies remains small.

A typical AI-native VPE package at a Series C company in 2025 looks roughly like:

• Base salary: $375,000 to $425,000
• Annual bonus: 25% to 40% of base, often discretionary
• Initial equity grant: 0.3% to 0.7% of fully diluted shares
• Refresh policy: annual refreshes increasingly common, sized at 25% to 75% of initial grant per year
• Sign-on bonus: $150,000 to $500,000 in cash, typically with 12-month clawback

The face value math on the equity component is what produces the headline. Take a Series C AI-native company valued at $5 billion at the most recent round, granting a 0.5% equity stake. Face value at signing: $25 million. Over a 4-year vest, that’s an annualized $6.25 million in equity comp at the headline level. Total annual comp at grant ($425K base + ~$140K bonus + $6.25M equity = ~$6.8M annualized at grant) is the kind of number that makes Big Tech VPE compensation look modest.

The realized number is, of course, dramatically different in expected-value terms. The 409A valuation is typically 30% to 50% below the most recent round valuation, which reduces the realized grant value at vesting. The company may not exit (or may exit at a lower valuation than the most recent round). Vesting is subject to continued employment, which means real risk if the company’s strategy changes or the VPE is unable to deliver on the technical mandate. And the time horizon to any liquidity event is at minimum 2–3 years and often longer.

Still, even after risk-adjustment, the expected value of AI-native VPE packages is well above any other segment we cover. The market is rational from the candidate side: the small pool of qualified candidates is genuinely being competed for, with multiple offers per finalist, and the companies are willing to pay because the alternative is not staffing the role at all.

Early-stage (Series A/B)

The smallest sub-segment in our 2025 dataset (6 placements), but worth covering because the structure is fundamentally different from later stages. Early-stage VPE roles — Series A or early Series B companies — pay meaningfully lower base salaries but offer materially larger equity stakes on a percentage basis.

Median early-stage VPE package in our 2025 data:

• Base salary $260,000 to $320,000
• Annual bonus usually none, sometimes structured around specific performance milestones
• Initial equity grant 0.8% to 2.5% of fully diluted shares
• Vesting typically 4-year with 1-year cliff, sometimes 6-year if the company is aiming for a longer runway

The realized comp at this stage is essentially a binary outcome. If the company succeeds, a 1.5% grant at a $50M valuation that exits at $2B is worth $30M to the VPE. If the company fails, the grant is worth zero. There’s no middle outcome to speak of. Most early-stage VPE candidates we work with are joining for the upside specifically; they’ve usually had a successful previous exit and can afford to take 30% to 40% lower base salary for a meaningful equity stake.

Why the AI premium is so large

The most common question we’ve gotten from VPE candidates in the past 18 months is some variant of: "is the AI premium real or is it bubble pricing?" The honest answer is both, in specific ways.

The premium is real for now, driven by two specific structural factors that haven’t been priced into traditional compensation benchmarks.

First, the talent pool is genuinely small. The set of engineering leaders who have credibly led a team building production-scale ML infrastructure or who have managed the engineering organization of a foundation-model company is, by our count, fewer than 800 people in the entire United States. Of those, fewer than 300 have actually managed engineering organizations larger than 50 people. The candidates we put forward for VPE-AI searches typically have 4 to 6 simultaneous competing offers within 30 days. The market is not deep enough to apply normal supply-demand dynamics.

Second, the equity grants at AI-native companies are large in face-value terms because the underlying valuations are large. AI-native companies at Series C+ are typically valued in the $1B to $50B range. A 0.3% to 0.7% grant carries a face value of $3M to $350M. Even after applying realistic discount factors for 409A pricing and liquidity risk, the headline numbers attract candidates in a way that pure cash compensation cannot.

The premium is also fragile in specific ways. The math depends on AI-native companies maintaining their current valuations, which in turn depend on continued capital inflows into AI infrastructure and applied AI. If that capital cycle slows — for any of several reasons including more efficient model training, market saturation of foundation models, or capital reallocation away from AI — the premium will compress quickly. We’ve advised candidates throughout 2025 to think of AI-native VPE offers as 18- to 24-month bets on the company’s ability to continue raising at increasing valuations, not as locked-in compensation.

For deeper coverage of the broader sector dynamics behind these numbers — how they fit into the overall picture of US senior compensation in 2026 — see our 2026 Executive Compensation Report, which puts the tech-specific bifurcation in context with finance, healthcare, and other sectors.

Anatomy of a VPE equity package

The equity component is where Bay Area VPE compensation really lives, and where most candidates leave money on the table because they don’t understand the structure as well as the company does. Five variables determine the realized value of an equity package, in approximately this order of importance.

Grant size. The most visible variable. Measured in shares for public companies and percentage of fully diluted for private. Easy to compare across offers superficially — harder to compare meaningfully because of all the variables below.

Vesting schedule. The traditional 4-year vest with 1-year cliff is being challenged at both ends. Big Tech companies moved toward 3-year vesting starting around 2023–2024 (Meta was the first major company to do this systematically). AI-native companies have experimented with front-loaded schedules — 33% in year one, 33% in year two, 34% over years three and four — to make day-one offers more attractive in the negotiation moment. Late-stage private companies have generally stuck with the traditional 4-year schedule but with smaller cliff periods.

Refresh policy. At public companies, annual refresh grants are standard. At late-stage private companies, refreshes are negotiable and increasingly granted starting year 2. At early-stage startups, you usually get one grant and that’s it. The cumulative dollar value of refresh grants over a 4-year tenure typically exceeds the initial grant by 50% to 100% — meaning the refresh policy is, in expected-value terms, more important than the initial grant size. We push refresh policy to be in writing in every VPE offer letter we negotiate.

Acceleration triggers. What happens to unvested equity on change of control? The strongest candidate position is single-trigger acceleration (full vesting on change of control alone). The most common compromise is double-trigger (acceleration only if change of control and termination without cause). The company default position is no acceleration. The cost to the company of agreeing to double-trigger in any non-exit scenario is zero; the value to the candidate in an exit scenario can be substantial.

Tax treatment and indemnification. Particularly relevant for IPO scenarios where you may owe income tax on shares you cannot yet sell due to lock-up. Major companies sometimes offer tax indemnification or cash gross-ups to cover this exposure. If your equity grant has tax exposure in a year you cannot fully realize it, ask for indemnification. This is rarely granted at private companies but is a meaningful component at public-company offers in the year of an IPO.

The refresh math nobody talks about

The most underdiscussed aspect of VPE compensation is the cumulative value of refresh grants over a 4-year tenure. We model this for every candidate we work with, and the numbers consistently surprise candidates more than any other component of the package.

Take a Big Tech VPE offer with the following structure: $420K base, 30% bonus, $2.6M initial equity grant over 4 years, and a "guideline" annual refresh of $650K starting year 2. At first glance, the annualized package is roughly $1.1 million ($420K base + $126K bonus + $650K equity amortized).

The math over a full 4-year tenure looks meaningfully different. Year 1 vests 25% of the initial grant: $650K. Year 2 vests 25% of the initial grant plus year-2 refresh: $650K + $163K = $813K. Year 3 vests $650K + $163K + $163K = $976K. Year 4 vests $650K + $163K + $163K + $163K = $1,139K. Add base and bonus across all four years: roughly $546K per year. Total realized comp over 4 years: roughly $5.9 million, or an annualized $1.48 million — significantly higher than the headline $1.1 million.

That math depends on the refresh actually materializing at the "guideline" level. If the refresh is discretionary and you receive 75% of the guideline, the 4-year total drops to roughly $5.5 million. If the refresh is 50% of guideline, total drops to $5.2 million. The structure of the refresh policy — specifically whether it’s contractual or discretionary — is therefore worth several hundred thousand dollars in expected value.

For private-company offers, the equivalent math is harder because the underlying equity has no immediate liquid value. But the same logic applies in expected-value terms: the refresh policy, if granted in writing, materially affects the total package over a 4-year tenure, and most candidates focus on the initial grant to the exclusion of the refresh.

Acceleration, lock-ups, and the IPO scenario

Two scenario-specific structural items that VPE candidates routinely overlook in offer evaluation.

Change of control acceleration. If the company is acquired during your tenure, what happens to your unvested equity? At Big Tech and late-stage private companies, double-trigger acceleration is increasingly standard but not universal. At early-stage and AI-native companies, the answer varies widely. The cost to the company of granting acceleration is zero unless an acquisition actually happens; the value to the candidate in an acquisition scenario can be $500K to $3M+ depending on the size of the unvested grant.

We’ve had several candidates in our follow-up data where the acceleration negotiation paid out within 18 months of signing: the company they joined was acquired, and the acceleration clause we’d negotiated converted unvested grants into realized cash. The clause was added to the offer letter at zero marginal cost to the candidate at the time of negotiation; the realization was substantial.

IPO lock-up scenarios. If you join a pre-IPO company that goes public during your tenure, what happens to your equity during the lock-up period (typically 180 days post-IPO)? The standard answer is: nothing changes — the equity continues vesting on the original schedule, and you can’t sell shares until the lock-up expires. Some companies offer accelerated vesting at IPO; some don’t. Some offer cashless exercise of options at IPO; some require you to come up with the exercise price yourself. The IPO scenario is sufficiently company-specific that you should ask explicitly in the final-round negotiation and document the answer.

Where to negotiate (and where it’s pointless)

Three patterns we see repeatedly across VPE offer negotiations in San Francisco.

Negotiate equity, not base. At public companies, base salaries are band-anchored and typically have little flex. At private companies, base salaries are constrained by internal compa-ratio considerations and have somewhat more flex but still less than equity. Equity grants, by contrast, have meaningful flex at most companies — we typically negotiate 15% to 30% increases on the initial equity number without affecting base salary at all.

Negotiate refresh policy explicitly. Most VPE offer letters are silent on refresh. Get it in writing. A guaranteed minimum refresh of $500,000 per year for years 2 through 4 adds $1.5 million to your expected realized compensation over a 4-year horizon. This is, in dollar terms, often the single largest negotiable item in the package — and the one most candidates fail to address.

Don’t over-negotiate sign-on at the expense of base or equity. A $200,000 sign-on with a 12-month clawback is, in expected-value terms, less valuable than a $150,000 increase in annual base salary — especially because the sign-on is taxed in one year while the higher base spreads tax and compounds for subsequent annual increases. Sign-on is the most flexible single component of the offer for the company to grant, which is exactly why companies prefer to grant it — it’s short-term in their P&L.

For the specific dynamics of negotiating equity packages in offers where you’re also weighing a counter-offer from your current employer, see our piece on why most counter-offers fail — the counter-offer dynamics are especially sharp for VPE candidates because most current employers can’t move quickly enough on equity to match an outside offer.

What we’re seeing in 2026

Three concrete trends from our 2025 and Q1 2026 placements that are noticeably different from prior years.

First, the AI-native premium is sticky. We were unsure entering 2025 whether the AI premium would compress as more candidates retrained into ML infrastructure and more companies adjusted hiring strategies. So far, the premium has held: our 2025 AI-native VPE placements pay 27% more than 2024 (already up significantly from 2023). The talent pool is growing but not faster than demand. We expect the premium to narrow but not disappear over the next 12–18 months.

Second, public-company comp is rebounding modestly. The 2024–2025 stock recovery in major public tech has translated into modestly higher refresh grants for 2025–2026 packages. The post-IPO VPE median we reported above ($520K) is up 3% from the 2024 figure, after declining for two consecutive years. This trend should continue if tech stocks remain at or above current levels.

Third, vesting schedule innovation continues. The shift toward 3-year vesting at Big Tech is spreading to late-stage private companies. We placed 4 VPEs at late-stage private SaaS companies with 3-year vesting in 2025 (versus zero in 2023). The candidate-favorable structure is being offered as a competitive differentiator against companies that still use 4-year schedules.

Looking ahead to H2 2026

Three predictions for the next 12 months of Bay Area VPE compensation, based on the conversations we’re currently having with both candidates and clients.

The AI premium will narrow modestly but not collapse. As the talent pool grows and more candidates train into ML infrastructure roles, the absolute scarcity will ease. But the 1.5x to 2x premium over non-AI VPE roles is likely to persist for at least another 2 to 3 years, even if compressed somewhat. We’re seeing the first hints of this compression in negotiations conducted in Q1 2026 — AI-native equity grants are still large, but the gap between the highest and lowest offers within the segment is narrowing.

Public-company comp will continue to rebound, but slowly. If the 2025 tech-stock recovery holds through 2026, refresh grants will be larger in dollar terms, boosting realized comp at Big Tech and post-IPO firms. The aggregate effect on median public-company VPE comp will be 5% to 10% over the next 12 months, in our estimation.

Equity vesting will continue trending candidate-friendly. The 3-year vesting trend will spread further. We expect to see 2 to 3 major late-stage private companies move to 3-year schedules during 2026. The original 4-year cliff vest is increasingly seen as an artifact of an earlier era when the labor market favored employers; the current market favors candidates, and the vesting schedules are catching up.

For VPE candidates who want a confidential read on a current offer or on how a hypothetical offer might compare to our 2025–2026 dataset, drop us a note. I’m happy to discuss specifics: rachel.mendez@crimsontalent.com. The conversation is confidential, free, and useful whether or not you’re actively in market.

Methodology & caveats

This report is built from 47 verified, signed-and-accepted offer letters from VP Engineering placements in the Bay Area made by Crimson Talent between January 2025 and the first week of April 2026. We exclude offers that were extended but rejected, withdrawn, or never finalized. "Bay Area" includes San Francisco, the East Bay (Berkeley to Oakland), the Peninsula (San Mateo to Palo Alto), and the South Bay (Mountain View to San Jose).

All compensation figures are gross (pre-tax), in nominal US dollars. Equity is valued at grant using the company’s most recent 409A valuation for private companies and the trailing 30-day average closing price for public companies. Annual bonuses are reported at target. Refresh grants are reported at the "guideline" level documented in offer letters; actual refresh outcomes vary based on performance and company circumstances. Sign-on bonuses are reported as full one-time amounts.

"AI-native" is our internal categorization of companies whose core intellectual property is AI rather than companies that use AI as a feature. The list of companies we classify this way grew from approximately 15 in 2023 to over 80 in 2026 — for the boundary cases, we use a combination of self-description, primary revenue source, and engineering team composition to make the classification.

This report does not constitute legal, financial, or compensation advice. Past performance and historical compensation ranges are not predictive of future outcomes for any individual offer or search. Specific compensation outcomes vary based on candidate qualifications, company budget, role specifics, and market conditions at the time of negotiation.

This piece is authored by Rachel Mendez, Talent Partner for our Technology & AI practice, with data assembly and review by the Crimson Talent research team. Rachel leads our Bay Area technology executive search practice from our San Francisco office at 555 California Street. Direct contact: rachel.mendez@crimsontalent.com. For broader context on US senior compensation across all sectors, see our 2026 Executive Compensation Report. For deeper coverage of compensation in other major US tech markets, see our piece on why Austin and Dallas are outpacing the coasts.