The Software Engineering Job Market in 2026: New Grads Squeezed Out, AI Labs Win Big

TLDR: Anthropic and OpenAI now dominate candidate interest for interview coaching, together accounting for over half of all requests. Intern and new grad hiring has cratered even as overall engineering hiring recovers. Pure frontend roles are effectively vanishing, and AI engineers are commanding significantly higher compensation than their software engineering peers.

The signal that caught my attention most in this data is what candidates are actually preparing for. At Interviewing.io, 35% of coaching requests are targeted at Anthropic, with OpenAI at 16%. That's 51% of all requests going to two AI labs, while each individual Big Tech company picks up around 17%. That's not a coincidence or a fad. Claude Code sitting at the top of developer tool surveys for two years running, Anthropic's $65B fundraise at a near-trillion-dollar valuation, the filing to go public, and the hire of Andrej Karpathy — these are signals that the talent market has fully internalized. People go where the momentum is, and right now the momentum is at the AI labs.

The retention numbers tell the same story from a different angle. Anthropic has an 80% two-year retention rate, which stands out in an industry where switching jobs every 18 months used to be the default career strategy. Google DeepMind sits at 78%, OpenAI at 67%. I find the OpenAI number interesting given how much attention they attract. When 35% of candidates want in but a third of employees leave within two years, there's a gap between the outside image and the internal experience worth paying attention to.

The new grad situation is genuinely concerning. Intern hiring kept falling through 2024 and 2025 even while overall engineering hiring recovered — the first time those two numbers have moved in opposite directions. Only 1 in 10 engineering hires at larger companies in 2025 were recent graduates, down from nearly 3 in 10 in 2023. Companies are not taking on junior talent to develop; they're hiring experienced engineers who can contribute immediately, and in smaller numbers than before. The side effect is that the engineers who do get hired fresh out of school are increasingly coming from a short list of elite universities. MIT, Stanford, CMU, UC Berkeley, Harvard, Caltech, Georgia Tech, Cornell. If you're not from one of those schools, the narrow path just got narrower.

Job title distribution is shifting in ways that matter for anyone planning a career. AI engineering has grown explosively. Forward Deployed Engineers, who sit at the intersection of product, engineering, and customer success, are multiplying fast as enterprise AI adoption creates demand for technical people who can make the software work in real customer environments. On the other end, pure frontend engineers are disappearing. Full-stack is becoming the baseline expectation, not a specialization. Native iOS and Android engineers are also declining as cross-platform frameworks mature and eat into the justification for platform-specific specialists. Sales engineers are up slightly, which tracks with the growth of complex enterprise sales where the deal requires a technical person in the room.

The compensation picture for AI engineers is striking. At the 80th percentile in the US, senior engineers are earning over $300K base. AI engineers earn more than software engineers generally, with the gap widening especially in equity, and widening most at AI labs compared to Big Tech. There's also a structural shift happening in management: fewer engineering managers per engineer, fewer VPs and directors across Big Tech. Workers are staying put more than they used to, and when engineers do leave Google, Apple, or Meta, they're mostly heading to AI labs. Amazon is the exception — those engineers go everywhere. Microsoft alumni, interestingly, are the most likely to become self-employed.

Key takeaways:

Anthropic and OpenAI now capture more candidate interview prep interest than all of Big Tech combined, reflecting a genuine shift in where engineering talent wants to work
Intern and new grad hiring has collapsed relative to overall hiring recovery, concentrating entry-level opportunities at a small set of elite universities
Pure frontend engineering as a job category is effectively disappearing, full-stack is the expectation, and AI engineering carries a meaningful compensation premium over traditional software roles

Why do I care: As a senior frontend developer and architect, the disappearance of pure frontend roles is the data point I keep coming back to. The writing has been on the wall for a while, but this makes it concrete. Full-stack is no longer a nice-to-have on a resume — it's the baseline. The question worth asking is whether the frontend craft gets lost in the transition, or whether it gets absorbed and respected as part of a broader skill set. On the AI engineering premium: if you've been curious about whether building experience with AI tooling actually translates to compensation, the answer is clearly yes.

State of the software engineering job market in 2026, part 2