NVIDIA's silicon runs the training clusters of every frontier lab on this index. Its stock returned 340 percent in the twelve months through August 2026, making RSU vesting outcomes highly variable by grant date. By market capitalization, it is the most valuable AI infrastructure company on the planet.
It scored 53 out of 100 on the ENTRA Pay Transparency Index Q2 2026, placing 12th among 20 scored AI employers, in the BB tier, 32 points below index leader Hugging Face (85, AA) and one point above Amazon AWS AI (52). The score does not reflect what NVIDIA pays. Senior Deep Learning Engineers clear $280,000 to $420,000 in total compensation (Levels.fyi Q2 2026, n=64 offers). Principal Research Scientists run $480,000 to $720,000 (Levels.fyi Q2 2026, n=28 offers). The score reflects what NVIDIA publishes, and those are not the same thing.
What NVIDIA Discloses and What It Withholds
The clearest picture of NVIDIA's disclosure posture is in its job postings. A Principal AI Research Scientist role at Santa Clara headquarters, captured in the ENTRA Q2 2026 audit, carries a posted band of $228,000 to $399,250 (CA SB 1162 public filings). That range exists because California SB 1162 requires it. The question is what happens outside the law's reach.
In most cases: nothing. Non-California roles at NVIDIA, including US roles in Texas and Washington and roles at international sites, routinely carry no posted salary range. A researcher considering a NVIDIA role in Austin or Zurich sees no band before applying. A researcher in Santa Clara sees one.
The score breakdown illustrates the pattern (ENTRA Pay Transparency Index Q2 2026): Salary Band Disclosure 11 of 30, because CA-only compliance earns partial credit and voluntary global disclosure earns the rest; Equity Compensation Transparency 16 of 25, with RSU details disclosed to offer recipients but not publicly available pre-offer; Internal Pay Equity Data Published 8 of 20, with the NVIDIA Diversity Report 2025 providing pay equity signals but no AI-role-specific analysis; Compensation Philosophy Public 12 of 15, the strongest single-dimension result; Market-Rate Commitment 6 of 10, with no explicit percentile target disclosed.
The Compensation Philosophy score is the number worth reading carefully. NVIDIA has articulated a pay philosophy in public-facing materials, and the index reflects it. The deficit sits on the dimensions where articulation gives way to data: the posted bands, the equity disclosure, the internal equity numbers.
The RSU Opacity Problem: Variable Equity, No Benchmark
NVIDIA grants equity as RSUs that vest quarterly. In a year when NVDA returned 340 percent, quarterly vesting produced substantially different outcomes depending on grant date. An RSU awarded in August 2025 vested at prices that did not exist when the grant was issued. A candidate modeling an August 2026 grant is working with a different starting price and different forward assumptions.
That dynamic is not specific to NVIDIA. The transparency problem specific to NVIDIA is the absence of public grant-size data at the role-band level. Candidates cannot benchmark the equity component of a NVIDIA offer before receiving one, because there is no published reference point for what comparable candidates in the same role family receive at grant.
Levels.fyi aggregates total compensation for NVIDIA AI roles (n=64 for Senior Deep Learning Engineer, Q2 2026), providing a total-comp reference for candidates who know to look. The RSU grant size as a standalone figure by role band is not systematically published in a form that enables pre-offer benchmarking. The Equity Compensation Transparency score of 16 of 25 reflects this partial state: credit for disclosing mechanics to offer recipients, deficit for what candidates cannot see before the conversation begins.
NVIDIA's 340 percent stock return is a compelling data point. The grant inputs a candidate would need to model their personal exposure to that return are not public.
What Reaching A Tier Would Require
The A tier starts at 70. NVIDIA is at 53. The 17-point gap is specific.
Salary Band Disclosure earns 11 of 30. Extending disclosure to non-California US roles alone could add 8 to 12 points on the index's most heavily weighted dimension. NVIDIA's Compensation Philosophy score of 12 of 15 indicates the internal band architecture exists. Publishing those bands in job postings is operational, not structural.
Internal Pay Equity Data Published scores 8 of 20. Adding AI-role-specific compensation data to the existing NVIDIA Diversity Report framework, median total compensation by function and seniority band, could move this score to 12 or 14. The Diversity Report 2025 is the foundation; the gap is specificity.
Market-Rate Commitment scores 6 of 10. A public statement naming a compensation percentile target closes most of the remaining gap on this dimension.
A realistic path to A tier: Salary Band Disclosure to 20, Internal Pay Equity Data to 13, Market-Rate Commitment to 9. Composite: 70. None of those changes require restructuring compensation. They require publishing information that governs decisions already being made internally.
The talent pressure to move in that direction will come from the frontier lab competitive set. Candidates NVIDIA recruits overlap with those Anthropic (62, BBB), OpenAI (65, BBB), and Google DeepMind (60, BBB) recruit directly. All three score above NVIDIA on pay transparency. None reach AA. The field is converging, and the gap at the lower end of the cluster is the most likely to close first as competitors extend their California disclosure posture to global postings.
This is the closing Sunday spotlight of ENTRA's Salary Transparency Month, August 2026. Anthropic at 62, BBB. Hugging Face at 85, AA. NVIDIA at 53, BB. All three pay at or above market for AI talent. The variable, as it has been all month, is how much of that story a candidate can read before the first recruiter call. NVIDIA's 340 percent stock return makes the company's case on its own terms. The outstanding question for Q3 2026 is whether candidates can benchmark what they would receive against that return before saying yes.
ENTRA Pay Transparency Index Q2 2026 methodology: five weighted dimensions, Salary Band Disclosure (30 pts), Equity Compensation Transparency (25 pts), Internal Pay Equity Data Published (20 pts), Compensation Philosophy Public (15 pts), Market-Rate Commitment (10 pts); 20 AI employers scored; data window January to June 2026; posting audit July 2026; BB tier: 50 to 59; A tier: 70 to 79; AA tier: 80 to 89. NVIDIA score breakdown: Salary Band Disclosure 11/30, Equity Compensation Transparency 16/25, Internal Pay Equity Data Published 8/20, Compensation Philosophy Public 12/15, Market-Rate Commitment 6/10; total 53/100. Salary data: Senior Deep Learning Engineer $280K to $420K TC and Principal Research Scientist $480K to $720K TC per Levels.fyi Q2 2026, n=64 and n=28 respectively. NVIDIA headcount: approximately 36,000 globally as of Q2 FY2027, per NVIDIA earnings reports; AI-focused engineering and research estimated 4,000 to 6,000. Principal AI Research Scientist posted band $228,000 to $399,250 per ENTRA Q2 2026 posting audit, CA SB 1162 public filings, Santa Clara, CA. NVDA 12-month return of +340% through August 2026 per public market data. NVIDIA Diversity Report 2025: published annually; source NVIDIA corporate site. Individual named companies (Anthropic, OpenAI, Google DeepMind, Hugging Face) did not confirm ENTRA Pay Transparency Index scores.
Global Career PlatformFind AI talent. Find your next role.
Booking is hotels. · Airbnb is apartments. · ENTRA is global careers.