Software engineer salary by company stage
Software engineer salary by company stage
Software engineer base salaries in 2026 scale predictably with company stage. A mid-level engineer earns roughly $110k at a seed startup, $145k at Series B, $175k at Series C/D, and $205k at a public Big Tech firm in the US. Equity flips the math: pre-seed equity is lottery-ticket gambling, while public RSUs are a near-cash bonus. Stage drives the trade.
The reason the spread exists is not generosity. Each stage has a different ratio of cash, equity, risk, scope, and structure, and engineers price that ratio differently depending on career arc. A staff engineer joining a Series A for $160k base plus 0.4% equity is making a different bet than a staff engineer joining Stripe for $310k base, $220k RSU vest, and a 15% target bonus.
Below, the actual numbers, the structural reasons behind them, and how to think about which stage matches your hiring budget or your career stage.
The numbers, by stage and level (US, 2026)
These are total cash compensation (base + bonus) ranges for US-based engineers in major tech metros, drawn from typical patterns visible on Levels.fyi, Pave, and Carta benchmark data. Equity is shown separately because it is essentially a different currency.
| Stage | Junior (L3) | Mid (L4) | Senior (L5) | Staff (L6) | Equity range |
|---|---|---|---|---|---|
| Pre-seed (1-5 ppl) | $80-100k | $100-130k | $130-160k | $150-180k | 0.5-2.0% |
| Seed ($1-3M raised) | $90-110k | $110-140k | $140-175k | $160-200k | 0.25-1.0% |
| Series A | $105-130k | $130-160k | $160-190k | $185-220k | 0.10-0.50% |
| Series B | $120-150k | $145-180k | $175-215k | $210-260k | 0.05-0.25% |
| Series C/D | $135-170k | $165-205k | $200-250k | $245-310k | 0.02-0.10% |
| Pre-IPO / late stage | $150-185k | $180-225k | $225-285k | $280-360k | RSU, $80-250k/yr value |
| Public Big Tech | $165-200k | $195-240k | $250-320k | $340-450k | RSU, $150-500k/yr value |
| FAANG top of band | $180-220k | $230-280k | $310-380k | $450-620k | RSU, $300k-1M+/yr value |
A few notes on reading this table.
Bonus targets cluster at 10-15% for late-stage and public companies, and are often zero at pre-seed and seed. The mid (L4) column is the most useful benchmark because it represents the bulk of any engineering org and is where most founders set their first market reference point.
Geographic premium for SF/NYC is roughly 15-20% above national. Austin, Seattle, and Boston track within 5%. Fully-remote roles at public companies typically pay 10-15% below SF tier but 5-10% above national median. Remote at seed stage is often pay-flat regardless of location.
Why the spread exists
The pre-seed engineer is paid less in cash because the cap table is the comp. A 1.5% common stake in a company that exits at $400M is $6M pre-tax. The same engineer at FAANG would clear roughly $1.8M in RSU value over four years at L5. Both bets pay if they hit, but the variance is wildly different.
Stage compresses cash variance and inflates equity certainty as you move right on the table. Public RSUs vest on a calendar regardless of company performance, so they price as cash with a delay. Seed equity prices as a 1-in-15 lottery ticket, so engineers demand it in larger denominations to compensate.
Three structural drivers explain the cash deltas:
Burn discipline. Pre-seed and seed companies pay below market because their runway math requires it. A $2M seed with 4 hires at $180k blended is 14 months of runway. Founders who try to pay above market at seed run out of money before product-market fit.
Recruiter access. Series B+ companies can afford recruiters at 20-25% of first-year base. That access expands the candidate pool, which lets them hire on signal rather than urgency, which compresses the band toward the top quartile. Earlier-stage companies pull from network and pay accordingly.
Comp committee formality. By Series C, most companies have hired a People Ops lead and run a banded compensation system aligned to Pave or Radford data. That alone adds 8-12% to mid-level offers because it eliminates the seed-stage habit of low-balling new hires.
For a deeper read on how productivity itself shifts across stages, our engineering productivity benchmarks for 2026 cover the output side of the same equation.
What the salary numbers do not capture
Base + bonus + equity is roughly 60% of the actual cost of an engineer in 2026. The rest hides in benefits load, ramp time, tool spend, and turnover risk.
A typical fully-loaded multiplier for a US salaried engineer:
- Benefits load: 25-30% (health, dental, 401k match, payroll tax)
- Recruiter fee: 18-22% of first-year base if external
- Ramp to productivity: 3-6 months of half-output
- Equipment and SaaS: $300-600/month per engineer
- Manager overhead: roughly 10% of a manager's time per direct report
A senior engineer at a Series B with a $185k base, 12% bonus, $40k RSU value, 28% benefits load, and a recruiter fee amortized over 24 months is costing the company about $350k/year all-in for the first two years. After that, drop to $300k as the recruiter cost rolls off.
Turnover risk compounds. Roughly 35-40% of engineers in their first year at startups leave within 18 months according to Carta's hiring data; the real cost of a bad engineering hire is $600k to $1.2M once severance, replacement, and lost momentum are priced in.
The on-demand alternative
Weekly booking changes the math at every stage, but especially below Series B. A senior engineer on Cadence costs $1,500/week, which is $78,000 for a full year of continuous booking. The fully-loaded comparable at Series B is roughly $260,000. The delta funds two more engineers or six months of runway.
The honest framing: this works when the scope is well-defined and the engagement is project-shaped (3-26 weeks). For a 5-year strategic hire who needs to own a function, headcount still wins because equity, tenure, and institutional context matter.
Where weekly booking is the right call in 2026:
- Pre-seed founders who need shipping velocity but cannot dilute another point
- Seed companies bridging between funding rounds without locking in burn
- Series A teams shipping a new surface area (mobile, ML, internal tools) without a permanent hire
- Series B+ teams with seasonal scope (year-end migrations, audit prep, integration buildouts)
Every engineer on Cadence is AI-native by default, vetted on Cursor, Claude Code, and Copilot fluency through a voice interview before they unlock bookings. The platform has 12,800+ engineers in the pool, with a 27-hour median time to first commit on a fresh booking.
The shift driving 2026 numbers
Three things changed since 2023 that compressed the salary table for shippable scope.
AI-native tooling moved the productivity floor. A mid engineer using Cursor and Claude Code on a defined feature ships 2-3x faster than the same engineer without AI assistance, per Cadence's internal benchmarks across roughly 4,000 weekly engagements. That increase concentrates in feature work, bug fixes, and integrations; architecture and judgment-heavy work compress less. The market effect: mid-level cash bands flattened year-over-year while staff/principal bands kept climbing.
Remote-first is default at every stage. In 2026, 78% of engineering roles posted on US job boards are remote-eligible according to LinkedIn's labor report. That has collapsed the geographic premium for everything except top-of-band FAANG roles. A senior engineer in Austin or Raleigh now earns within 5% of a senior in SF.
Booking emerged as a third category. Beyond full-time hiring and traditional agency contracts, weekly engineer booking is now a real budget line at startups under 50 engineers. This puts downward pressure on agency rates (which used to charge $200-400/hour for senior work) and creates a new comp ladder for engineers who prefer project work over salaried employment.
For the international view, our breakdown of engineering team cost by country shows how the same engineer profile prices in Eastern Europe, LatAm, and South Asia.
How to use this data (decision framework)
Five questions to ask before you set a number on an offer or accept one.
- Is this a 12-week project or a 5-year strategic capability? If under 12 months, weekly booking almost always wins on cash math. Above 12 months, the equity case for a hire reasserts itself.
- Have I validated the scope, or am I hiring to figure it out? Hiring to figure it out is the most expensive way to validate. Book a senior for 6 weeks first, then hire once the shape is clear.
- Am I paying senior money for mid-level scope? This is the most common founder error. Senior engineers cost 40-60% more than mid for work that mid handles cleanly. Match the tier to the actual scope, not to perceived status.
- What is my replacement cost if this hire does not work out? At seed, a bad hire costs ~6 months of runway. At Series B, it costs $600k+. Both numbers shape how you should structure the trial period.
- Does the equity story make sense to a numerate engineer? If your offer relies on equity to be competitive, the engineer will price it. Run the math: dilution, exit probability, vesting cliff. If it does not pencil, fix cash.
If you want a fast pass on whether headcount or booking makes sense for a specific role, you can run the numbers on Cadence's ROI tool in about 90 seconds. It compares fully-loaded annual cost against a weekly booking equivalent for the same tier.
What this means for retention and refreshes
Salary at a single point in time is half the story; the other half is how comp evolves. By Series B, most engineers expect a refresh grant in year 2 and a merit bump of 4-8% annually. Skipping the equity refresh grant for engineers at year 2 is the most common reason senior engineers leave a Series B or C company. They benchmark the new hire's offer against their own and find a 15-20% gap.
The same is true for cash. The compounding effect of 4% annual merit on an engineer hired three years ago at $145k yields $163k today. A new hire into the same role gets offered $180k off the market. That spread is the resignation letter.
Sources
Numbers in this post draw from publicly visible compensation data at Levels.fyi, Pave, and Carta's compensation benchmarks, plus the Stack Overflow 2026 Developer Survey for stage-level patterns. Cadence's internal stats (12,800-engineer pool, 27-hour median time to first commit, 67% trial-to-active conversion) come from platform telemetry across roughly 4,000 weekly engagements in 2026 to date.
FAQ
What is a typical senior software engineer salary at a Series B startup in 2026?
A senior (L5) engineer at a US Series B startup typically earns $175k to $215k base, plus a 10% bonus target and 0.05-0.25% equity. Fully-loaded cost to the company is roughly $250-290k/year including benefits, equipment, and amortized recruiter fees.
How much equity should a software engineer get at a seed startup?
A senior engineer joining a seed-stage US startup in 2026 typically receives 0.5% to 1.0% in common stock with a four-year vest and one-year cliff. Junior hires receive 0.1% to 0.4%. The right benchmark is the cash gap: if base is $30k below market, equity should compensate at expected value, not at upside.
Do remote engineers get paid less than in-office engineers in 2026?
At pre-Series-B companies, remote and in-office engineers are usually paid the same band because there is no central office. At public companies and pre-IPO, fully-remote engineers typically earn 5-15% below the SF/NYC tier but still above national median. The gap is closing every year.
Is it cheaper to book a weekly engineer than to hire full-time?
For engagements under 12 months, yes, almost always. A senior engineer on Cadence at $1,500/week annualizes to $78k, compared to $250k+ fully-loaded for a Series B salaried senior. For permanent strategic roles where institutional knowledge compounds, full-time hiring still wins because equity and tenure cannot be replicated weekly.
How do I benchmark my own salary against company stage?
Pull comp data from Levels.fyi for your level (L3-L6) and your specific company stage, then adjust for geography (-15% for non-tier-1 metros, +15% for SF/NYC). Compare your total cash compensation, not base alone. If you are 10%+ below the stage median for your level, request a market-rate adjustment cycle, not a one-off raise.