
Software engineer total comp in 2026 is base salary plus target bonus plus equity (RSUs or options) plus signing bonus plus benefits. At a US public tech company, the typical split for a senior is 55% base, 10% bonus, 30% equity, 5% one-time signing. A FAANG L5 lands $380K to $520K all-in. A Series-B startup senior lands $190K cash plus illiquid equity worth $0 to $400K depending on outcome.
The components look simple on a one-page offer letter. The math underneath them (vesting schedules, refreshers, ESPP discounts, ISO vs NSO tax treatment, double-trigger acceleration) decides whether a $450K offer is actually worth more than a $330K offer. This is the anatomy guide.
Every tech package, from a 5-person seed startup to Google L7, decomposes into five buckets. Some are zero at small companies. None are negotiable in isolation; recruiters trade them off against each other.
| Component | Share (FAANG) | Share (Series-B) | Liquid? | Taxed when |
|---|---|---|---|---|
| Base salary | 40% to 55% | 70% to 85% | Yes | Each paycheck |
| Target bonus | 8% to 15% | 0% to 10% | Yes | Annually |
| Equity (RSU or options) | 30% to 50% | 5% to 25% paper | Partial | Vest (RSU) or exercise/sale (options) |
| Signing bonus | 2% to 6% (year 1 only) | 0% to 5% | Yes | Receipt |
| Benefits + 401k + ESPP | 8% to 15% (cash-equivalent) | 5% to 10% | Mixed | Varies |
Public companies skew toward equity because RSUs are liquid the day they vest. Private companies skew toward base because illiquid options are worth nothing until a liquidity event. Internalize that before reading any further.
Base is the only fully predictable, fully liquid number, unaffected by stock price or performance. It pays your rent. It is also the anchor for almost everything else: target bonus is a percentage of base, 401k contributions are a percentage of base, severance is usually weeks of base. When you negotiate, fight for base first.
Median 2026 US base by level (Levels.fyi):
Series-B to Series-D startup base typically lands 10% to 20% below FAANG for the same level, with the gap closed by equity that may or may not pay out.
Target bonus is a percentage of base, paid annually based on company performance and individual rating. A 15% target at company performance of 1.0x and individual rating of "meets" pays exactly 15%. At 0.8x performance with a "below" rating, it pays maybe 9%. Budget the target as a 50% probability number, not a guarantee.
Typical targets: FAANG IC 10% to 20%; hedge fund / quant 50% to 200%; Series-B 0% to 10% (often informal CEO discretion); big-bank engineer 25% to 75%, with 50% deferred in stock.
This is where most engineers get confused, and where most negotiation room lives.
Restricted Stock Units (RSUs) are a promise to deliver shares on a vesting schedule. You owe nothing to receive them. On vest, full market value is taxed as ordinary income (W-2). Public-company RSUs are effectively cash with extra steps. Private-company RSUs are usually structured with a "double-trigger" so they only vest after both time AND a liquidity event (IPO or acquisition); otherwise you would owe income tax on illiquid shares.
Stock options give you the right to buy shares at a fixed price (the strike). If the company is worth $100/share and your strike is $5, exercising captures $95 of value per share. If the stock is at $4, your options are "underwater." Two flavors:
Series-A through Series-C startups grant ISOs to early employees and convert to NSOs once the $100K cap is hit. Late-stage and public companies grant RSUs.
The default schedule across nearly all of US tech is 4 years with a 1-year cliff: nothing vests for 12 months, then 25% lands on day 366. The remaining 75% vests monthly or quarterly over 36 months.
Common patterns:
Amazon's headline TC is real, but year 3 is when equity income actually kicks in. Leave at month 30 and you get less than half what the offer implied.
Initial grants vest to zero by year 4. Without refreshers, TC drops ~25% every year. Refreshers are annual top-up RSU grants, sized 25% to 75% of the original. Strong refresher cultures (Google, Meta, NVIDIA) keep steady-state TC flat or growing. Weaker ones (Amazon, many post-IPO companies) let TC fall 10% to 20% from the year-1 peak.
The refresher is the most under-negotiated line on a senior offer. Always ask: "What is the typical refresher for someone at meets-expectations in this role?"
One-time cash, typically $20K to $100K for IC roles. Usually clawed back 100% if you leave before 12 months, 50% between 12 and 24 months. It exists to replace unvested equity from your prior employer and to front-load comp against back-loaded vesting (Amazon especially). It is the easiest line to negotiate because it does not affect anyone else's pay band.
Real money, often $25K to $40K of cash-equivalent per year, that almost every offer-comparison spreadsheet ignores.
Here is the trap. When RSUs vest, full market value is W-2 income. The company sells (typically) 22% of shares to cover federal withholding. Your actual marginal rate is closer to 35% to 45% (federal + state + Medicare + Additional Medicare). The shortfall lands on your April tax bill.
Worked example: Meta senior, 2026, vesting $180K of RSUs. Withholding at 22% = $39.6K. Actual liability at 32% federal + 9.3% California + 1.45% Medicare + 0.9% additional Medicare = 43.65% marginal. Real tax owed: $78.6K. Shortfall due April 15: $39K cash. Engineers who do not set this aside get a brutal surprise.
Second trap: post-vest, you hold a concentrated position in your employer. The optimal move (per nearly every flat-fee CFP) is to sell on vest and diversify. Most engineers do not, ride the stock, and either get rich or get crushed.
Three advanced items that matter at early-stage startups and on senior offers.
83(b) election. Applies to restricted stock awards (actual shares granted with vesting restrictions), not RSUs. An early employee granted 100,000 restricted shares at $0.01 strike when 409A FMV is also $0.01 can file 83(b) within 30 days of grant. This pays income tax on $1,000 instead of ordinary income on the full appreciated value as it vests. If the company exits at $50/share, your gain is long-term capital gains. Savings can be 7-figure. Miss the 30-day window and there is no recovery.
Pre-IPO secondary. Late-stage private companies (Series-D and later) increasingly run tender offers letting employees sell 5% to 25% of vested shares to incoming investors. Stripe, Databricks, Anthropic, and SpaceX have run multi-billion-dollar secondaries. Ask in the offer process: "Have you run a secondary in the last 24 months?"
Acceleration on change of control. Single-trigger (all unvested vests on acquisition) is almost extinct outside founders. Double-trigger (vests on acquisition AND termination without cause within 12 to 24 months) is the standard senior-IC ask. For L5+, asking for double-trigger on at least 50% of grant is reasonable and increasingly granted. Most engineers never ask, and lose 5-figure to 6-figure value in M&A scenarios.
Meta E5 offer, May 2026. Stock at $620.
| Component | Year 1 | Year 2 | Year 3 | Year 4 |
|---|---|---|---|---|
| Base | $230,000 | $240,000 | $250,000 | $260,000 |
| Target bonus (15%) | $34,500 | $36,000 | $37,500 | $39,000 |
| RSU vest ($620K grant, 25/25/25/25) | $155,000 | $155,000 | $155,000 | $155,000 |
| Refresher RSUs | $0 | $40,000 | $90,000 | $140,000 |
| Signing | $50,000 | $0 | $0 | $0 |
| Cash + RSU TC | $469,500 | $471,000 | $532,500 | $594,000 |
| ESPP + 401k + benefits cash-equiv | $32,750 | $32,750 | $32,750 | $32,750 |
| All-in TC | $502,250 | $503,750 | $565,250 | $626,750 |
Year 1: ~$470K cash and stock, $500K all-in. RSU vest assumes flat stock price; real-world value moves with the share price.
60-person Series-B, post-money $250M, 50M shares fully diluted, last 409A $5 / preferred $10.
| Component | Year 1 | Year 2 | Year 3 | Year 4 |
|---|---|---|---|---|
| Base | $190,000 | $200,000 | $210,000 | $220,000 |
| Bonus (informal, ~5%) | $9,500 | $10,000 | $10,500 | $11,000 |
| Options (50,000 ISO at $5 strike) | $62,500 paper | $62,500 paper | $62,500 paper | $62,500 paper |
| Signing | $15,000 | $0 | $0 | $0 |
| Cash TC | $214,500 | $210,000 | $220,500 | $231,000 |
The $250K of paper equity is worth $0 unless the company exits. Realistic distribution: 40% chance of $0, 30% chance of $200K to $500K (acquisition), 20% chance of $500K to $1.5M (favorable IPO), 10% chance of $1.5M+. Risk-adjusted expected value is roughly $400K over 4 years.
Series-D AI infra company, 400 people, last round at $4B post.
| Component | Year 1 | Year 2 | Year 3 | Year 4 |
|---|---|---|---|---|
| Base | $310,000 | $325,000 | $340,000 | $355,000 |
| Bonus (10%) | $31,000 | $32,500 | $34,000 | $35,500 |
| RSU (double-trigger, $1.6M grant) | $400K (locked) | $400K | $400K | $400K |
| Signing | $75,000 | $0 | $0 | $0 |
| Liquidity (year-3 secondary, ~25% of vested) | $0 | $0 | $200,000 | $0 |
Most equity is vesting but unsellable until IPO or secondary. The company runs an annual tender at year 3 letting principals sell 25%. Stay through an $8B IPO and the full $1.6M lands liquid; on a stall or down-round, the $1.6M compresses to $400K to $800K.
If you are a founder reading this to size an engineering offer, three uncomfortable conclusions:
1. You cannot match FAANG TC in cash. A FAANG L5 making $470K wants $230K cash plus $240K liquid stock. You can match $230K cash on a Series B. You cannot match $240K of liquid stock. Your equity is illiquid lottery tickets.
2. Be honest about equity volatility. Do not pretend a $400K paper grant is "the same as $400K liquid RSUs." Sophisticated engineers know it is not. Pretending damages trust faster than offering less.
3. For project work, the TC math does not apply. A 12-week mobile rebuild does not need someone you have to outbid Meta for. It needs a senior for 12 weeks. A senior on Cadence at $1,500/week × 12 weeks is $18,000. The fully-loaded FAANG senior is $115,000 of comp for the same 12 weeks, plus 30 to 60 days of recruiting you do not have.
If you are sizing for a 12-week or 12-month project rather than long-term headcount, run the booking-vs-headcount numbers before you post the JD. The math flips earlier than most founders expect: we routinely see senior in-house roles sit open for 90+ days while the same scope ships in 4 weeks via booking.
The practical takeaway is not "FAANG pays more." It is: TC is a bundle, and the right bundle depends on the work.
For strategic 5-year hires (see how senior, staff, and principal comp actually compare in 2026), pay headcount with real equity and competitive base. For project scope, hire by the week. Fully-loaded cost of a US senior is roughly 1.6x base ($350K all-in for a $220K base), and you replace none of it for 6 months while they ramp.
Cadence's pool of 12,800 vetted engineers, every one AI-native by default (Cursor, Claude Code, Copilot fluency vetted in a voice interview before bookings unlock), absorbs scope that does not justify a headcount conversation. Book a senior at $1,500/week with a 48-hour free trial; swap the next week if fit is wrong, no notice period. Engineers get paid Friday for the week's work, so the ones we attract could be at Meta but want optionality.
For deeper benchmarks see how AI fluency is bifurcating developer pay 12% to 56%.
Try Cadence. Book a senior engineer in 2 minutes, run the numbers on booking versus headcount, and use the engineer 2 days at no cost. Weekly billing, swap any week, no notice period.
Total compensation (TC) includes base salary, target bonus, equity (RSUs or stock options), signing bonus, and benefits like 401k match, ESPP discount, and health insurance. At public tech companies, equity is typically 30% to 50% of TC. At Series-B startups, equity is mostly illiquid paper value with a 40% chance of going to zero.
The standard schedule is 4 years with a 1-year cliff: nothing vests for 12 months, then 25% vests on your one-year anniversary, with the remaining 75% vesting monthly or quarterly over the next 3 years. Amazon uses a back-loaded 5/15/40/40 schedule, paired with sign-on bonuses to make years 1 and 2 competitive.
ISOs qualify for long-term capital gains if held 1 year post-exercise and 2 years post-grant, but trigger AMT on exercise. NSOs have no AMT but tax the spread between strike and FMV as ordinary income immediately. Startups grant ISOs first, then NSOs once the $100K-per-year ISO cap is hit.
Companies withhold at the 22% supplemental rate, but your actual marginal rate is 35% to 45% (federal + state + Medicare). Set aside an additional 15% to 20% of vest value for the April shortfall. In high-tax states like California or New York, the gap is even larger.
Double-trigger vests remaining equity if (1) the company is acquired AND (2) you are terminated without cause within 12 to 24 months. For senior engineers and above, double-trigger on at least 50% of grant is standard and increasingly granted. Most engineers never ask, and lose 5- to 6-figure value in M&A.
Discount startup equity by probability: 40% chance of $0, 30% partial outcome, 30% full or above. Compare cash directly. A FAANG L5 at $470K vs a Series-B senior at $215K cash plus $250K paper equity: risk-adjusted startup TC is ~$315K, well below FAANG, but with a 10% chance of $1.5M+.