AI Jobs Are Paying More Than Ever as Startups Compete for Top Engineers

A wave of competition among artificial intelligence startups is changing how much tech professionals earn, with some roles now paying well into the upper six figures. Federal disclosure documents, required for companies hiring foreign workers under the H-1B visa program, reveal how aggressively AI companies are offering high base salaries to attract skilled employees.

These records show only base compensation and leave out other incentives such as stock options, signing bonuses, or performance awards, which often make total pay significantly higher. Still, the raw numbers offer a rare look into how much top startups are paying to stay ahead in the race for AI talent.

OpenAI Offers Broad Salary Bands for Core Staff

OpenAI, the San Francisco-based research lab behind ChatGPT, reported some of the highest base salaries across its technical and operations teams. The role of software engineer appears in several listings, with salaries ranging from $200,000 to $440,000, depending on specialization. Research-focused roles, such as research engineers and research scientists, were reported at up to $440,000.

Hardware engineering roles also made the list, with pay reaching $360,000. Other technical jobs, including security engineers and intelligence investigators, were listed between $310,000 and $382,500. Non-technical positions such as program staff, finance, and community support carried base salaries between $220,000 and $270,000. In some cases, data scientists and design professionals were paid up to $385,000.

These salaries do not reflect the company’s broader compensation package, which includes equity and variable bonuses. Some recent listings on OpenAI’s website suggest software engineers can earn as much as $590,000 annually when all components are included.

Anthropic Lists the Highest Research Engineering Compensation

Anthropic, another high-profile AI startup based in the U.S., disclosed particularly high pay for its research roles. A research engineer at the company can earn between $340,000 and $690,000. Some listings for members of the technical staff also reached $405,000.

The highest-paid technical staff manager received $690,000. Roles in finance and strategy were also well-compensated, typically falling between $230,000 and $285,000. Anthropic’s hiring includes other departments as well, with recruiters earning $170,000 and operations staff around $230,000. A regulatory lead position was listed at $210,000, while account executives in sales were paid about $126,000.

Cohere, Glean, and Abridge Are Paying Above Traditional Tech Levels

Other AI startups are competing with similar offers, even if their operations are smaller. Cohere, which builds large language models for business use, offered $240,000 to members of its technical staff.

Glean, a company focused on enterprise search tools, paid software engineers between $190,000 and $260,000. Its machine learning engineering roles came with base salaries of $210,000, while software engineering managers reached $250,000. A data science team lead at Glean was listed with a base salary of $230,000.

Abridge, a health tech company using AI to automate clinical documentation, reported software engineers earning between $235,000 and $240,000. These are senior roles with experience requirements, placing them near the upper salary range for most health-related AI positions.

Grammarly Offers Strong Compensation Across Teams

Grammarly, which integrates AI into writing and grammar tools, has expanded its technical hiring with salaries that reflect the sector’s competition. Machine learning software engineers earned up to $318,000, while engineers working on data infrastructure could receive $315,000.

AI engineering positions were listed at $223,000. Other technical roles such as analytics engineers and platform developers were typically paid between $250,000 and $280,000. Even non-engineering roles in user research, marketing, and product design received high salaries. Product designers, for instance, were paid up to $230,000, while user researchers earned more than $221,000.

Mistral AI and Safe Superintelligence Use Wide Salary Ranges

Mistral AI, based in Europe but hiring in the U.S., listed salary bands for AI scientists between $280,000 and $350,000. A spokesperson later clarified that salaries in the U.S. could fall anywhere between $150,000 and $450,000.

Safe Superintelligence Inc., a new company co-founded by former OpenAI researchers, showed a wide range for its technical staff as well. Listed salaries ranged from $150,000 to $500,000, which reflects a diverse hiring strategy across seniority levels.

Smaller Startups Show Upward Trends in Design and Strategy Roles

Anysphere, the team behind Cursor (an AI code assistant), offered $250,000 to a product designer. Synthesia, which creates AI avatars for video generation, listed a customer success manager role at $120,000. Despite being smaller firms, these companies appear to be aligning with the pay expectations set by larger AI labs.

Poolside, another startup with a focus on software and AI integration, listed a performance director at $230,000.

Thinking Machines Lab Matches Top-Level Compensation

Thinking Machines Lab, led by Mira Murati, has entered the high-salary tier with listings of $450,000 to $500,000 for core technical roles. The company also disclosed that a co-founder and machine learning scientist earned within the same range. These figures suggest a tight alignment with compensation levels offered by longer-established AI firms.

The Competitive Landscape Is Pushing Salaries Higher

These figures come from mandatory government filings, which companies submit when sponsoring employees through the H-1B visa system. While they exclude other elements of compensation, they still provide a clear benchmark.

Companies working in AI are consistently raising base pay to compete for a narrow slice of talent capable of building the next generation of machine learning models. From product design to platform engineering, salaries reflect a landscape where specialized knowledge continues to carry a premium.

Methodology

To analyze compensation trends across top artificial intelligence startups, Insider examined publicly available salary disclosures submitted to the U.S. Department of Labor. These filings are part of the H-1B visa application process, which requires employers to report salary information when seeking to hire foreign workers for specialized roles. Only base annual salaries were considered for this analysis. Supplemental compensation such as equity grants, signing bonuses, or performance-based incentives was not included, as these details are not required in the federal disclosure process.

The data collection focused exclusively on AI-focused startups with private market valuations exceeding $2 billion, excluding companies that specialize in robotics, autonomous vehicles, or military-related technologies. Salary listings were drawn from the most recent filings available in 2024 and early 2025.

Where multiple salary entries existed for the same job title within a company, reported figures were expressed as ranges to reflect variation by experience, location, or specialization. In cases where company representatives provided updated or clarifying salary bands, those figures were included to provide more accurate context.

This approach offers a rare look into base compensation practices within leading AI startups, given that most companies in the sector do not publicly share pay structures.

Read next: Trust Gaps Remain: Everyday Scenarios Where AI Chatbots Struggle to Deliver