AI and ML engineers are among the most in-demand H-1B candidates in 2026. This guide covers the companies sponsoring the most AI visas, what the roles pay, and how to position yourself for a sponsored offer in the most competitive segment of the tech job market.
Artificial intelligence and machine learning roles have become the fastest-growing segment of H-1B sponsorship over the past three years. According to DOL LCA disclosure data, ML Engineer, Data Scientist, AI Research Scientist, and related AI-adjacent roles (MLOps Engineer, AI Infrastructure Engineer) have grown as a share of all H-1B filings from under 3% in 2020 to over 8% in 2025.
This growth reflects a fundamental shift in the tech hiring market. While traditional software engineering hiring has cooled since the 2021–2022 boom, AI and ML hiring has remained strong through 2025–2026 even as overall tech employment contracted. Companies building LLM products, AI infrastructure, and enterprise AI applications are competing for a small pool of qualified candidates globally — and H-1B sponsorship is often a prerequisite for hiring from that pool.
The defining feature of AI H-1B hiring in 2026 is the bifurcation between two tracks: research-oriented roles (AI Research Scientist, Applied Research Scientist) concentrated at major labs (Google DeepMind, Meta FAIR, Microsoft Research, OpenAI, Anthropic, xAI, Apple AI/ML), and product/engineering roles (ML Engineer, MLOps, AI Platform Engineer) that are distributed across the entire tech sector from startups to Fortune 500.
Each track has different H-1B dynamics. Research roles often attract candidates with exceptional publication records who may also qualify for O-1A or EB-1A directly. Product/engineering roles are largely standard H-1B cap-subject petitions and subject to the lottery. Understanding which track you are on matters for both job search strategy and immigration planning.
Based on DOL LCA data and H-1B disclosure records through 2025, these are the highest-volume H-1B sponsors for AI and ML positions:
Google / Alphabet: The largest single H-1B sponsor for AI roles. Google DeepMind, Google Brain (now merged), and Google Research collectively file hundreds of H-1B petitions annually for ML researchers and engineers. Product teams (Search, Ads, YouTube) also file significant volumes for applied ML roles. Google's H-1B approval rate historically exceeds 95% for AI/ML roles due to the specificity and strength of their job descriptions.
Meta (Facebook): Meta AI Research (FAIR) and Meta's product AI teams are among the top 10 H-1B sponsors overall. Meta actively recruits from top CS PhD programs globally and regularly sponsors both new graduates and experienced researchers. Meta's compensation packages for AI roles (base + RSUs) frequently exceed $400K total for senior researchers.
Microsoft / LinkedIn / GitHub: Microsoft's AI investment (including the OpenAI partnership) has driven significant H-1B volume for AI roles across Azure AI, Copilot, Bing, and research teams. Microsoft sponsors across all experience levels, from new MS/PhD graduates to senior researchers.
Amazon / AWS: Amazon sponsors heavily for applied ML roles across Alexa, AWS AI services (SageMaker, Bedrock), and Amazon's logistics and recommendation systems. AWS's growth in enterprise AI has created significant demand for ML engineers at all levels.
OpenAI, Anthropic, xAI: These newer AI labs have become significant H-1B sponsors despite their smaller size relative to Big Tech. OpenAI reportedly filed 300+ H-1B petitions in 2024–2025. Anthropic and xAI are smaller but growing. These companies also lead in O-1A and EB-1A sponsorship for exceptional researchers.
Nvidia: Nvidia's GPU dominance makes it a key employer for AI infrastructure, CUDA optimization, and hardware-software co-design roles that require specialized AI knowledge. Nvidia's H-1B volume has grown significantly since 2023 as demand for GPU engineering expertise expanded.
DOL prevailing wage determinations set the floor for H-1B compensation. AI/ML roles primarily fall under two SOC codes with different wage levels:
SOC 15-2051 (Data Scientists): This code covers most ML Engineer and Data Scientist roles. 2026 prevailing wages at Level III (experienced) in major metros: San Francisco-Oakland-Hayward $205,000; Seattle-Tacoma-Bellevue $195,000; New York-Newark-Jersey City $188,000; San Jose-Sunnyvale-Santa Clara $210,000; Boston-Cambridge-Nashua $182,000; Austin-Round Rock $168,000; Chicago-Naperville $162,000.
SOC 15-1252 (Software Developers, Applications): Many ML Engineer roles are filed under this code rather than 15-2051. Level III wages are generally 10–15% lower than 15-2051, making this a less favorable code for salary purposes but more straightforward for specialty occupation qualification.
Actual market compensation at top AI companies significantly exceeds prevailing wages. Total compensation (base + equity) for senior AI researchers and engineers at Google, Meta, OpenAI, and Anthropic routinely ranges from $400K to $1M+ annually. The H-1B prevailing wage is the floor, not the ceiling — the issue is companies must pay at least the prevailing wage.
For PhD candidates entering through OPT/STEM OPT before H-1B cap: this pathway is increasingly popular in AI because it allows 3 years of work authorization post-graduation without lottery exposure, during which the employer sponsors H-1B. Many AI companies prefer this path because it avoids losing a new hire to an unlucky lottery draw.
The O-1A visa (Extraordinary Ability in Sciences, Education, Business, or Athletics) has no annual cap and no lottery. For AI researchers with strong publication records, it is a viable — and often superior — alternative to the H-1B lottery. O-1A approval requires demonstrating extraordinary ability through at least 3 of 8 evidentiary criteria.
For AI researchers, the strongest criteria are: publications in peer-reviewed journals or conferences of major significance (NeurIPS, ICML, ICLR, CVPR, ACL, EMNLP are all recognized); high citation counts (Google Scholar h-index matters here); judging or reviewing for recognized journals or conferences; invited talks or workshops at major AI venues; press coverage in outlets covering AI research; original contributions of major significance (key algorithm innovations, widely-adopted open-source frameworks).
Many O-1A petitions for AI researchers are approvable with 3–5 years of research output from a strong PhD program. First-author papers at top conferences, 500+ total citations, and 2–3 invited talks is a commonly achievable bar for PhDs from strong programs. Work with an immigration attorney experienced in O-1A to assess your record before applying.
O-1A as an H-1B backup: some AI companies now offer "O-1A if lottery miss" as part of their offer. This means: file for H-1B in the lottery, and if not selected, the company files O-1A as a backup. This strategy keeps high-demand AI talent working without losing them to the lottery randomness. Not all companies offer this — ask during negotiations if this matters to you.
For international students and professionals targeting AI roles with H-1B sponsorship, the following factors most strongly influence both the likelihood of receiving an offer and the quality of the employer's sponsorship program:
Degree pedigree and research output: Top AI labs (OpenAI, Anthropic, DeepMind, FAIR) recruit almost exclusively from PhD programs at Stanford, MIT, Carnegie Mellon, Berkeley, UW, Cornell, and similar programs. Publication record at NeurIPS/ICML/ICLR is effectively required for research scientist roles. For ML engineer roles, strong MS degrees from these programs with relevant research experience are competitive.
Demonstrated specialization: "Generalist data scientist" is a commoditized role with high H-1B RFE risk. Specialization — vision transformers, RL from human feedback, diffusion models, LLM fine-tuning, graph neural networks — makes the specialty occupation case easier and makes you more competitive for the roles where sponsorship is most readily available.
Open source contributions: Major contributions to PyTorch, TensorFlow, Hugging Face, JAX, or popular ML libraries are visible signals of expertise that carry weight both in hiring decisions and O-1A applications. A library with 500+ GitHub stars attributed to your work is objectively strong evidence of contribution.
Target cap-exempt employers strategically: If lottery anxiety is a concern, target universities, university-affiliated research labs (Stanford HAI, MIT CSAIL, CMU ML Department), nonprofit research organizations (Allen Institute for AI), and government labs (NIST, NIH) for initial positions. These are cap-exempt and allow you to build experience and publications before transferring to industry — which can then be done without cap exposure on cap-exempt portability.
Aisha Patel is a technical recruiter specializing in AI/ML talent acquisition at major tech companies. She has placed hundreds of international researchers and engineers into sponsored positions and advises on H-1B and O-1A immigration strategy.