Entry-level AI roles now pay a median $156,000 in San Francisco — up 41% since Q1 2024. That figure is not for senior engineers. It is for new graduates with 0–2 years of experience, before the first promotion cycle. The compression between new-grad and mid-level compensation in AI has narrowed to under 18% at the frontier labs — a gap that was 35–40% across the rest of software engineering as recently as 2022. The 2026 AI graduate market is, by every measure this index tracks, the most competitive entry-level hiring environment in the history of the technology sector.
This is the first annual edition of the ENTRA AI Graduate Salary Index. It covers 18 roles, four regions, and 0–2 years of experience only. It does not duplicate the Top 20 Highest-Paid AI Roles 2026 (published January 2026), which covered all experience levels. This index isolates the graduate cohort specifically: new BSc and MSc hires, entry-level contractors placed through platforms such as Mercor and Scale AI, and early-career practitioners below the senior threshold.
Methodology: Salary figures are median total compensation (base + bonus + equity, annualised) for roles with 0–2 years of experience. Sources: Levels.fyi (Jan–Apr 2026 self-reported, n=4,200+), Mercor contractor rate surveys (n=1,800), LinkedIn Salary Insights (Apr 2026), ENTRA Talent Index proprietary panel (n=620 companies). All figures USD unless noted. UK figures converted at GBP/USD 1.28. EU at EUR/USD 1.09. UAE AED/USD 0.27. Equity is annualised over a standard four-year vest. Contractor roles (marked †) reflect annualised hourly-rate equivalents at median billed hours; they do not include equity.
The 2026 Salary Table
| Role | Median US | Median UK | Median EU | Median UAE | YoY % | |---|---|---|---|---|---| | ML Engineer (Foundation Models) | $168,000 | $112,000 | $88,000 | $74,000 | +38% | | AI Research Scientist | $175,000 | $118,000 | $92,000 | $68,000 | +44% | | Applied AI Engineer | $156,000 | $104,000 | $82,000 | $70,000 | +41% | | AI Safety Researcher | $171,000 | $115,000 | $89,000 | $62,000 | +52% | | Data Scientist (AI-Native) | $138,000 | $92,000 | $74,000 | $64,000 | +29% | | AI Product Manager | $148,000 | $98,000 | $78,000 | $72,000 | +35% | | Prompt Engineer / LLM Engineer | $132,000 | $86,000 | $68,000 | $60,000 | +61% | | RLHF Specialist (W-2)† | $128,000 | $82,000 | $64,000 | $52,000 | +48% | | Computer Vision Engineer | $152,000 | $100,000 | $80,000 | $66,000 | +33% | | NLP Engineer | $149,000 | $98,000 | $79,000 | $65,000 | +31% | | AI Infrastructure / MLOps | $144,000 | $96,000 | $76,000 | $68,000 | +27% | | Robotics AI Engineer | $158,000 | $104,000 | $84,000 | $78,000 | +36% | | AI Data Engineer | $126,000 | $84,000 | $66,000 | $58,000 | +24% | | Multimodal Systems Engineer | $162,000 | $108,000 | $86,000 | $72,000 | +43% | | AI Policy Analyst (Tech firm) | $118,000 | $82,000 | $66,000 | $54,000 | +19% | | Quantitative AI Researcher | $182,000 | $124,000 | $96,000 | $76,000 | +47% | | AI Ethics Researcher | $112,000 | $76,000 | $62,000 | $48,000 | +22% | | Autonomous Systems Engineer | $160,000 | $106,000 | $85,000 | $80,000 | +39% |
† RLHF Specialist figures reflect W-2 employment at AI labs and annotation platforms; contractor rates via Mercor/Scale AI can exceed $200,000 annualised for PhDs in quantitative disciplines.
Regional Breakdown
United States — The Frontier Premium Widens
The US market, specifically San Francisco and Seattle, remains in a category of its own. The median new-grad AI Research Scientist at a frontier lab (Anthropic, OpenAI, Google DeepMind, Meta AI) cleared $175,000 in total compensation for the Jan–Apr 2026 period — and the p90 figure, at $224,000, has now crossed a threshold that senior software engineers outside AI did not reach until their fourth or fifth year of tenure as recently as 2023.
The compression story is the under-reported one. At Anthropic and OpenAI, the gap between a new-grad ML engineer offer and a mid-level ML engineer offer has narrowed to 14–18%. The driver is straightforward: frontier labs have concluded that the supply of genuinely capable new graduates in foundation-model work is sufficiently constrained that underbidding them at the entry band produces unacceptable attrition within 12 months. The result is an entry-band that would have been competitive for L4–L5 Google engineering roles in 2021.
Outside the frontier-lab cohort, the applied-AI employer tier — AI-native startups, applied enterprise AI, fintech AI divisions — sits 20–28% below on base but frequently compensates with faster equity appreciation on smaller strike prices. The effective total-compensation gap at 24–36 months post-hire, accounting for equity performance, is narrower than the headline base differential suggests.
New York, Austin, and Boston round out the US tier. New York carries a 4–7% premium on AI Policy and Quantitative AI Researcher roles (driven by hedge fund and financial-services demand). Boston's cluster — MIT, Harvard, and the Route 128 biotech-AI corridor — sustains a 3–5% premium on Computer Vision and Robotics AI roles tied to the MedTech adjacency. Austin lags SF/SEA by 18–22% on median but has closed 6 percentage points of that gap since 2024 as Tesla AI, Apple Austin, and AMD AI have scaled local headcount.
United Kingdom — The London Floor Rises, the Gap Narrows Slowly
UK AI graduate salaries rose 28% in GBP terms since Q1 2024 — significant in isolation, but masked in USD comparison by sterling's 2025 softness. The median Applied AI Engineer in London clears £81,000 ($104,000), a figure that makes London competitive with non-SF US markets but still trails San Francisco by 33% on USD-equivalent total comp.
The structural constraint in the UK is equity. UK AI graduates receive materially less equity than their US counterparts at equivalent firms: the Levels.fyi 2026 panel shows UK new-grad equity grants averaging $18,000 annualised versus $42,000 annualised in the US for matched roles. The gap is partly structural (UK EMI scheme limits, different option pricing conventions) and partly cultural (UK startups grant equity later in tenure). For graduates who price their compensation on a four-year horizon rather than the signing year, this equity gap outweighs the base-salary headline.
London's AI Safety Researcher market is the exception. The density of AI safety-focused employers — DeepMind, Anthropic UK, the Alignment Forum-adjacent research cluster, and the UK AISI — has created a local premium on AI safety and AI ethics roles that does not exist at this magnitude in any other non-US city. AI Safety Researcher median total comp in London ($115,000) now sits within 33% of the San Francisco equivalent, the smallest regional gap in this index.
European Union — Germany and Netherlands Lead; Equity Gap Is Structural
EU entry-level AI compensation rose 31% in EUR terms since 2024, with the strongest cluster in Berlin, Amsterdam, and Paris. The Germany-Netherlands corridor accounts for 58% of the ENTRA panel's EU AI graduate placements by volume and carries a 12–15% premium over France and 22–28% over the EU periphery.
The headline EU constraint is identical to the UK: equity. EU new-grad AI engineers receive annualised equity grants averaging $9,000–$14,000 — less than one-third of the US equivalent. At the base-salary level, top EU employers (ASML AI, SAP AI, Mistral, the European frontier-lab cluster) are increasingly competitive with non-frontier US employers. At total-comp level, the gap reasserts itself. The EU graduates who accept US-employer remote offers — a cohort that has grown materially since 2024 — do so almost entirely for equity access, not base salary.
The Mistral effect is measurable in the Paris market. Mistral's 2025 and 2026 graduate hiring rounds set a local benchmark for base salary and option grants that pulled up the Paris AI market by an estimated 8–12% in the trailing 12 months. The Iliad Group / Free AI initiative and Google DeepMind Paris have followed the benchmark upward.
Middle East / UAE — Highest Non-Salary Compensation, Fastest Trajectory
UAE AI graduate salaries in USD terms remain 50–55% below San Francisco medians — but the comparison requires the full compensation picture. Dubai and Abu Dhabi AI employers (G42, Microsoft UAE, AWS MENA, Oracle Cloud Infrastructure UAE) offer a suite of non-salary benefits that the USD comparison does not capture: housing allowance (median $18,000–$24,000 annually for graduate roles at Tier 1 UAE AI employers), annual flight allowance, and school-fee support for dependents. Adjusting for these, effective total compensation narrows the gap to 35–40% versus San Francisco.
The more significant number is the trajectory. UAE AI graduate salaries rose 39% in USD terms from Q1 2024 to Q1 2026 — the fastest two-year growth of any region in this index, edging out the US (38%) and substantially ahead of the UK (22% in USD terms). The driver is the combination of G42's expansion under Microsoft's technical partnership, the ADGM and DIFC financial-services AI buildout, and the UAE government's AI talent attraction program, which subsidises visa processing and relocation for verified AI specialists.
The Robotics AI Engineer and Autonomous Systems Engineer roles show the strongest UAE figures ($78,000–$80,000) — a structural effect of the UAE's sovereign investment in autonomous logistics, smart-city infrastructure, and defense-adjacent robotics programs through entities including Presight, Edge Group, and the MBZUAI research cluster.
Forecast
Three forces will determine where these numbers sit at the May 2027 refresh. First, frontier-lab hiring restraint: Anthropic, OpenAI, and Google DeepMind each entered 2026 with hiring bars materially higher than 2024, compressing graduate offer volume even as offer values remain elevated — any meaningful expansion of frontier-lab graduate programs will push the US median above $185,000. Second, EU equity reform: the European Commission's pilot programs on startup equity taxation, if enacted as currently drafted, would narrow the EU-US total-comp gap by an estimated 8–12 percentage points by 2027, with Paris and Berlin as primary beneficiaries. Third, the UAE's Phase 2 AI investment cycle — tied to a reported $12B commitment from Abu Dhabi sovereign vehicles across 2026–2027 — is the variable most likely to produce the largest regional surprise: UAE AI graduate medians could close 10–15 percentage points of the current gap with EU figures within 18 months if the hiring programs attached to those investments materialise as structured.
Want this data in your hiring planning cycle? The ENTRA Talent Index panel is updated quarterly. Compensation bands, equity benchmarks, and role-specific offer data are available to verified HR and talent acquisition professionals. Contact intelligence@entracareers.com with your company domain for panel access.
How we ranked
The AI Graduate Salary Index 2026 is scored across the following dimensions:
- Compensation Competitiveness — Median total compensation (base + bonus + equity, annualised) for roles with 0–2 years of experience (Source: Levels.fyi Jan–Apr 2026, n=4,200+; Mercor contractor rate surveys n=1,800; LinkedIn Salary Insights Apr 2026)
- Regional Conversion — All non-USD figures converted at Apr 2026 spot rates: GBP/USD 1.28, EUR/USD 1.09, AED/USD 0.27 (Source: Federal Reserve H.10 release, Apr 2026)
- YoY Delta — Computed against Q1 2024 baseline from prior ENTRA Talent Index panel and Levels.fyi historical salary data
Data window: January 2024 — April 2026 Sample size: 4,200+ Levels.fyi self-reported data points; 1,800 Mercor survey responses; 620-company ENTRA Talent Index proprietary panel Year-over-year delta: Computed against Q1 2024 baseline (first prior-edition anchor for this index)
Limitations:
- Private companies not represented on Levels.fyi may be under-represented in compensation data, particularly for EU and UAE markets where public compensation disclosure norms are weaker than in the US
- Contractor annualisation (RLHF Specialist and adjacent roles) assumes median billed hours; high-output contractors at the p90 level may earn materially above tabulated figures
Inquiries about methodology: methodology@entracareers.com