Every week brings a new headline declaring that AI will crush software engineer salaries. And every week, actual compensation data tells a more complicated story. Yes, tech salary growth hit a 15-year low of 1.6% in early 2026. Yes, 52,150+ tech workers have been laid off this year alone — roughly 815 per day — with 44% of those layoffs directly citing AI as the reason. But the engineers who know how to build with AI? They're pulling a 21-43% salary premium over their peers.
The real story isn't about salaries collapsing. It's about salaries splitting.
The Headline Numbers: Tech Salary Growth Hit a 15-Year Low
Let's start with the uncomfortable part. After years of aggressive compensation growth driven by zero-interest-rate hiring sprees, tech salaries have flatlined. The 1.6% year-over-year increase across the industry in Q1 2026 is the weakest growth since 2011, when the sector was still recovering from the financial crisis.
Here's how it stacks up against recent history:
| Year | Tech Salary Growth (YoY) | Context | |------|--------------------------|---------| | 2021 | 8.3% | Post-COVID hiring boom | | 2022 | 6.7% | Peak ZIRP competition | | 2023 | 3.1% | Post-layoff recalibration | | 2024 | 2.4% | Cautious hiring resumes | | 2025 | 2.0% | AI anxiety begins | | 2026 | 1.6% | 15-year low |
But averages hide the real dynamics. That 1.6% is being dragged down by specific segments — primarily entry-level and mid-level generalist roles — while specialized positions continue to see strong gains. The median software engineer salary in the US still sits at approximately $127,000, and in major hubs like San Francisco and Seattle, total compensation packages for senior engineers regularly exceed $250,000.
The question isn't whether software engineer salaries are falling. It's whether your specific type of software engineering work is gaining or losing value.
Entry-Level Developers Are Feeling It Most
If there's one group bearing the brunt of AI-driven salary pressure, it's junior developers. The numbers are stark: P1/P2 (entry-level) hiring has dropped 73% at major tech companies compared to 2023 levels. That's not a gradual decline. That's a structural shift.
The logic from the employer side is straightforward. AI coding assistants like Copilot, Cursor, and Claude can now handle a significant portion of what junior developers traditionally did: boilerplate code, simple bug fixes, documentation, basic feature implementation. A senior engineer paired with AI tools can now produce output that previously required a senior plus two juniors.
This doesn't mean entry-level positions have vanished entirely. CS graduates are still landing jobs — the average starting salary for a CS grad in 2026 is $81,535, which is actually up 7% year-over-year. But fewer of those jobs exist, and the bar to get them is considerably higher.
Key takeaway: The entry-level developer market has contracted by 73%, but those who do land roles are earning more than ever. The filter has gotten tighter, not the pay.
Companies are hiring fewer juniors but paying the ones they hire more, because those hires now need to demonstrate competency with AI tools from day one. The job description has changed. "Junior developer" in 2026 means something different than it did in 2023.
What This Means for Career Starters
If you're a new grad or bootcamp graduate entering the market right now, the path forward requires a different strategy than it did three years ago:
- Portfolio projects need to demonstrate AI-augmented development, not just raw coding ability
- Internship experience matters more — the 73% drop in entry-level hiring means companies overwhelmingly convert interns rather than hire cold
- Specialization beats generalism earlier in your career than it used to
Use our salary calculator to benchmark where entry-level roles in your target city actually land, because the national average obscures massive regional variation.
The AI Skills Premium: Who's Actually Earning More
Here's where the narrative flips. While generic tech salaries stagnate, engineers with demonstrable AI skills are commanding significant premiums. And the data breaks down in an interesting way:
| AI Skill Level | Salary Premium | |----------------|---------------| | IC-level AI competency (using AI tools effectively) | +12% | | Single specialized AI skill (ML, NLP, computer vision) | +21% | | Multiple AI skills (full-stack AI/ML capability) | +43% |
That 43% premium for multiple AI skills is not a marginal difference. On a base of $130,000, that's an additional $55,900 in annual compensation. These aren't theoretical numbers from job postings — they're drawn from actual offer data across major tech employers.
The premium is highest at the senior and staff levels, where the ability to architect AI systems (not just use AI tools) commands the biggest multiplier. But even at the mid-level, engineers who can fine-tune models, build RAG pipelines, or deploy inference infrastructure are seeing offers $15,000-$30,000 above comparable non-AI roles.
The Skill Stack That Pays
Not all AI skills carry equal weight. The market currently values:
- LLM application development — building products on top of foundation models
- ML infrastructure and MLOps — the plumbing that makes AI systems production-ready
- AI safety and evaluation — testing, red-teaming, alignment work
- Domain-specific AI — applying AI to healthcare, finance, legal, or manufacturing
General "prompt engineering" as a standalone skill has already commoditized. The premium goes to engineers who combine deep software engineering fundamentals with applied AI capability.
Which Roles Are Most at Risk (and Which Are Safe)
The Harvard Business Review published a finding in late 2025 that deserves more attention than it got: companies are laying off workers based on AI's perceived potential, not its actual demonstrated performance. That's a critical distinction. It means the current wave of AI-driven layoffs is partially speculative — firms are cutting roles they believe AI will replace, even when the technology hasn't fully proven itself in those domains yet.
That said, the risk isn't evenly distributed:
| Risk Level | Roles | Salary Trajectory | |------------|-------|-------------------| | High Risk | QA testers, junior frontend, data entry, basic DevOps scripting | Flat to declining | | Medium Risk | Mid-level backend, general full-stack, basic mobile development | Stagnant (0-2%) | | Lower Risk | Staff/principal engineers, security specialists, infrastructure architects | Growing (3-6%) | | Benefiting | AI/ML engineers, AI platform builders, AI safety researchers | Surging (8-15%) |
The pattern is consistent: roles that involve routine, well-defined tasks face the highest pressure, while roles requiring system-level thinking, ambiguity tolerance, and cross-domain judgment remain insulated.
Software developer roles as a whole are still projected to grow 17.9% through 2033 according to BLS projections. But that growth will be concentrated in the "lower risk" and "benefiting" categories above, not evenly spread across all types of developers.
The Roles AI Actually Struggles With
AI is remarkably good at generating code. It's remarkably bad at:
- Understanding business context and translating fuzzy requirements into technical decisions
- Navigating legacy systems with undocumented dependencies and tribal knowledge
- Making architecture tradeoffs that balance technical debt, team capability, and business timelines
- Debugging production incidents that span multiple services with incomplete observability
- Security engineering — identifying novel attack vectors and building defense-in-depth strategies
If your daily work centers on these activities, your salary trajectory looks very different from someone whose primary output is CRUD endpoints.
The Dallas Fed Data: AI-Exposed Sectors Have Higher Wages
This is the statistic that confuses people most, and it's the most important one in this entire discussion. Research from the Federal Reserve Bank of Dallas found that wages in AI-exposed sectors grew 16.7%, compared to the 7.5% national average. That's more than double.
Read that again. The sectors most exposed to AI disruption are seeing faster wage growth, not slower.
How does this make sense? The mechanism works like this:
- AI tools make individual workers significantly more productive
- Higher productivity per worker means each worker generates more revenue
- Companies compete for workers who can leverage AI effectively
- Competition drives up wages for those workers
- Meanwhile, workers who can't leverage AI become relatively less valuable
This is a productivity-driven wage premium, and it's the same pattern we saw with previous waves of automation. When ATMs were introduced, bank teller employment actually grew — but the role changed from cash handling to relationship management and sales, which commanded higher pay.
Key takeaway: AI exposure doesn't automatically mean salary decline. Workers who adapt to AI-augmented workflows are capturing the productivity gains. Workers who don't adapt are being displaced. The technology is a lever, and which direction it pushes your salary depends on how you use it.
You can compare cities to see how this dynamic plays out geographically — tech hubs with higher AI adoption tend to show stronger wage growth than markets that are slower to adopt.
How to Future-Proof Your Salary Against AI Disruption
Based on the data, here are concrete moves that correlate with salary resilience:
1. Move Up the Abstraction Ladder
The engineers seeing salary growth are those who work at higher levels of abstraction — system design, architecture, technical strategy. AI can write functions. It can't yet decide which functions should exist, how systems should interact, or when to take on technical debt strategically. Invest in system design skills and seek roles with "architect" or "staff" in the title.
2. Stack AI Skills, Don't Just Learn One
The difference between a 12% premium (basic AI competency) and a 43% premium (multiple AI skills) is enormous. Don't stop at learning how to use Copilot. Learn how to fine-tune models, build evaluation pipelines, and deploy inference at scale. The compounding premium rewards depth.
3. Specialize in a Domain
AI engineers who understand healthcare regulations, financial compliance, or manufacturing processes are far harder to replace than generalist AI practitioners. Domain expertise is a moat that pure technical skill alone doesn't provide.
4. Build in Public
With hiring down 73% at the entry level and competition fierce at every level, your portfolio and public work carry more weight than ever. Open-source contributions, technical writing, and conference talks create asymmetric optionality — they give you leverage in negotiations that a resume alone can't.
5. Benchmark Relentlessly
If you haven't checked your compensation against current market data in the last six months, you're flying blind. Use our salary percentile tool to see where you actually stand, and our cost of living comparison to evaluate whether a geographic move could increase your real purchasing power.
For engineers considering relocation to optimize their compensation, our relocation guides break down the true cost of moving between major tech hubs.
The Bottom Line: Polarization, Not Collapse
The data doesn't support the "AI will destroy software engineer salaries" narrative. It also doesn't support the "everything is fine" narrative. What it supports is a story of polarization.
The software engineering profession is splitting into two tracks:
-
Track A: Engineers who integrate AI deeply into their workflow, build AI systems, and operate at high levels of abstraction. Their salaries are growing 8-15% annually, they have multiple competing offers, and their skills become more valuable as AI adoption accelerates.
-
Track B: Engineers whose work consists primarily of well-defined, routine implementation tasks that AI tools can increasingly handle. Their salaries are stagnant to declining, hiring for their roles is contracting, and the competitive pressure only increases with each model generation.
The 52,150 layoffs this year are real. The 73% drop in entry-level hiring is real. The 1.6% salary growth floor is real. But so is the 43% AI skills premium, the 16.7% wage growth in AI-exposed sectors, and the 17.9% projected job growth through 2033.
Where you land depends on which track you're on — and the good news is that switching tracks is still very much possible. The AI skills premium exists precisely because demand for these skills outstrips supply. The window to build these capabilities, while still relatively uncrowded, is open now. It won't stay that way forever.
The engineers who treat AI as a career threat will likely see their fears confirmed. The engineers who treat it as a career accelerant will likely see their bets rewarded. The data, as of March 2026, is unambiguous on this point.
For a deeper dive into how AI is reshaping compensation across every role and experience level, read our Complete AI & Future of Work Salary Guide.