In 2024, Meta’s median annual total compensation for software engineers crossed $400,000 for the first time, according to data collected by Levels.fyi and reported by Business Insider. That number includes base salary, annual bonus, and stock refreshes, and it’s for the median, not exceptional outliers. The gap between top-paying tech companies and everyone else has gotten harder to ignore.
This post covers what the compensation landscape actually looks like in 2026, which companies are at the top, and what the differences between them mean for a job search decision.
The companies paying the most, ranked roughly
Total compensation data for software engineers varies depending on level, location, and the current stock price of the company. These figures represent mid-to-senior level (L5/E5 equivalent) software engineers in the US, using publicly available data from Levels.fyi, the Stack Overflow Developer Survey 2024, and job posting disclosures where available.
| Company | Approximate total comp (mid-senior SWE) | Notable structure |
|---|---|---|
| Meta | $380K-$550K | High base, large equity grants, aggressive refresh |
| Google / Alphabet | $320K-$480K | Strong base, predictable equity, good bonus structure |
| Apple | $280K-$420K | Strong base, equity is smaller than Meta/Google but reliable |
| NVIDIA | $350K-$500K+ | Stock price appreciation has made NVIDIA equity exceptional recently |
| Amazon / AWS | $250K-$400K | Base is lower, equity is front-loaded with longer vest schedules |
| Microsoft | $260K-$380K | Steady; OpenAI relationship has created new high-comp pockets |
| OpenAI / Anthropic | $350K-$600K+ | High base, but equity is illiquid (private); risk profile is different |
| Stripe / Airbnb | $280K-$420K | Competitive; both still paying at or above FAANG base levels |
These ranges are approximate. Level inflation and company-specific job titles make direct comparisons genuinely hard. An “L5” at Google does not map cleanly to a “Senior SWE” at Stripe. The ranges are wide partly because they’re real.
Why NVIDIA jumped to near the top of this list
NVIDIA’s compensation ranking changed dramatically between 2022 and 2024. The underlying reason is straightforward: NVIDIA’s stock price increased by roughly 700% between January 2023 and January 2025. Engineers who received equity grants in 2022 saw their total comp for that year retroactively balloon.
Going forward, the picture is less clear. NVIDIA still pays competitive base salaries, but the equity value going forward depends heavily on whether the stock continues to appreciate. Engineers who joined in late 2024 and 2025 are holding equity at much higher strike prices than their predecessors. Whether that pays off depends on things that are genuinely uncertain. I’d rather say that plainly than pretend the projection is clean.
The Amazon comp structure deserves its own explanation
Amazon consistently confuses candidates because its base salaries look lower than peers. An L5 software engineer at Amazon might have a base of $175K-$185K, while the equivalent level at Meta might be $210K-$230K. The difference is made up in equity and signing bonuses, but the equity structure at Amazon front-loads differently.
Amazon RSUs vest at 5% / 15% / 40% / 40% over four years. This means your first two years, you’re receiving less equity value than your grant might suggest. Year three and four compensate, but candidates who leave within two years (which happens frequently given Amazon’s attrition rate) often feel they underearned relative to peers at other companies. It’s worth modeling this explicitly before accepting an offer.
The private company question: OpenAI, Anthropic, and others
Salaries at top AI labs are high by any standard. OpenAI and Anthropic have publicly offered base salaries above $300K for senior researchers and engineers, with equity on top. The catch is liquidity: that equity is in a private company, and there’s no predictable timeline for when it can be sold. Secondary markets exist but are illiquid and subject to company approval.
The expected value calculation depends on your personal risk tolerance and your view of those companies’ trajectories. If you’d asked me this question in 2022, I’d have said the probability of both companies reaching the valuations they’ve reached since was higher than the market priced. Whether the trajectory continues is genuinely uncertain, more uncertain than the public discourse usually acknowledges.
What the BLS data tells you (and what it doesn’t)
The Bureau of Labor Statistics reports median annual wages for software developers at $132,270 as of its most recent data. That figure is accurate for the median across all employers in all geographies, which means it includes a lot of roles at mid-size companies, regional employers, and non-tech industries that employ software developers.
For the specific set of top-paying tech companies covered here, the median is roughly 2 to 3 times higher than the BLS figure. The BLS number is useful for understanding the full distribution; it understates what the top 10 to 15 tech employers pay by a wide margin.
How to use this information in a job search
A few things that actually matter when using compensation data in a search:
- Total comp is what matters, not base salary. Don’t negotiate base in isolation without also pushing on equity grant size and refreshes.
- Level placement varies by company and matters a lot. Getting placed at L4 vs. L5 at Google is a six-figure annual difference. It’s worth pushing back on level placement during the offer stage if you have evidence you should be higher.
- Cost of living adjustments are real. Meta and Google have remote-work policies now, but many roles are still adjusted for location. A $400K offer in San Francisco and a $340K offer in Austin may have very similar purchasing power.
- Signing bonuses are one-time. Companies use them to close candidates, but they inflate year-one numbers significantly. Build your financial model on base plus ongoing equity, not total-year-one including signing.
If you’re actively interviewing at multiple companies from this list, keeping your performance consistent across rounds is harder than the prep guides suggest. Craqly’s AI interview copilot is built for exactly that scenario, running live practice sessions tailored to specific company interview styles so you can work through the differences in real time, not just read about them.
The gap between the highest-paying companies and the rest isn’t narrowing. If you’re optimizing for total compensation, the list above is where to focus.