A software engineer I know accepted a Google L5 offer last spring and turned down a Meta E5 counter-offer that was $67,000 higher in total compensation. He said the Google role had less performance anxiety. Whether that trade-off was correct is something only he can evaluate, but the fact that he had to make it at all illustrates something worth understanding: FAANG salaries are not a single number. They are a set of bets on your own performance, risk tolerance, and how you value cash versus equity.
The figures here draw from Levels.fyi’s 2025-2026 aggregated data and self-reported compensation in Blind and LinkedIn communities. I’m not going to round anything to a tidy number, because that’s how you know someone is estimating rather than reporting.
Median total compensation by company
Across all engineering levels, Meta and Netflix lead. Meta’s median total compensation sits around $455,000. Netflix comes in similarly, though Netflix almost never grants stock options and instead pays a very high base salary, which makes their packages look different structurally even when the total number is comparable. Google’s median is lower, around $308,000 total comp, largely because they hire at higher volume and have more engineers at mid-level bands pulling down the median.
Amazon is consistently the lowest among the five for comparable levels. Their base salaries are lower, their RSU grants vest on a back-loaded schedule (roughly 5% in year one, 15% in year two, 40% in year three, 40% in year four), and the cash compensation in year one of an Amazon offer can be significantly lower than an equivalent Google offer even when the headline numbers are close. A lot of engineers figure this out after accepting.
Apple falls somewhere between Google and Amazon on total comp, and their equity grants have historically been more consistent and less volatile than Meta’s.
Level-by-level breakdown
| Level (approximate) | Meta | Amazon | Netflix | |
|---|---|---|---|---|
| New grad / entry | $213K | $247K | $173K | $350K |
| Mid-level (SWE II / E4) | $285K | $371K | $221K | $415K |
| Senior (L5 / E5) | $371K | $486K | $298K | $495K |
| Staff / Principal | $547K | $788K | $412K | $620K |
These are medians, not ceilings. Top-of-band offers at Meta and Netflix for principal engineers exceed $1 million in total comp. A small number of staff-level Google engineers receive similar packages. These are real numbers, not outliers, but they require sustained high performance ratings to maintain because equity refreshes at Meta are heavily performance-gated.
The vesting schedule is part of the salary
Most engineers don’t model vesting schedules carefully when comparing offers, and this is one of the most common negotiation mistakes. Amazon’s back-loaded vesting means your first-year cash will feel genuinely tight if you’re used to a higher salary, even if the four-year total looks competitive. Google’s schedule front-loads significantly more in years one and two (approximately 38% and 32% respectively), which matters a lot if you’re not sure you’ll stay four years.
Netflix’s model is entirely different: no stock options, much higher base salary, and the expectation that you manage your own investment decisions. Their engineers often earn over $300,000 in pure base. This is great if you want predictability and questionable if you believe Netflix’s stock will outperform your own investment choices.
What “refreshes” mean and why Meta wins on paper
Every FAANG company except Netflix grants equity refreshes annually to retain engineers. At Meta, the refresh pool for high performers is large, and an E5 engineer who consistently receives strong performance ratings can see their total equity package grow substantially year over year without changing roles. The catch is that Meta’s performance culture is genuinely intense. “Meets all expectations” at Meta is effectively below average. Engineers who thrive there will likely out-earn comparable Google engineers significantly over a 5-year period. Engineers who don’t find the culture energizing will likely burn out before the refreshes accumulate.
This is where I’d push back a bit on pure total-comp optimization. The Stack Overflow 2024 survey found that “work-life balance” ranked above “higher salary” as a reason to leave a job for the first time since they started asking the question. Whether that reflects a genuine shift in priorities or just what people say publicly, I’m not sure. But compensation optimization that ignores environment tends to end badly.
Negotiating FAANG offers
Every FAANG offer has room. Google’s initial offers are negotiable, often significantly. Meta tends to offer closer to ceiling in initial offers but has more flexibility on equity split between RSUs and signing bonus. Amazon’s base salary has a hard cap (roughly $350,000 at most levels due to an internal policy), so negotiation there mostly works through equity and signing bonuses.
A few practical notes: competing offers are the single strongest negotiation lever, and the companies know each other’s packages well. Saying you have a competing offer without providing specifics weakens your position. Saying “I have an E5 Meta offer with $X in RSUs” gives the recruiter something concrete to work with internally.
The worst FAANG salary outcome usually isn’t a low offer. It’s accepting without negotiating, which leaves money on a table where the amounts are genuinely significant.