How Many Employees Do Top AI Companies Have in 2026?

How big is OpenAI? Is Cursor really just 50 people? How does Anthropic compare to DeepMind on headcount? These are some of the most-searched questions about AI companies right now — and the answers are surprisingly hard to find in one place.

We compiled employee counts, Glassdoor ratings, open role data, and growth signals for 40 of the most prominent AI and tech companies in our Culture Directory. Whether you're evaluating job offers, benchmarking compensation, or just curious about how the AI industry is scaling, this is the most complete headcount dataset we've published.

The Full Ranking: All 40 Companies

Here is every company in our database sorted by estimated employee count, from largest to smallest. Employee counts are sourced from LinkedIn, Glassdoor, press reports, and company disclosures as of March 2026. Open roles reflect live job postings on each company's ATS.

The range is staggering. Stripe at ~8,000 employees is 160x the size of Cursor at ~50 — yet both are building category-defining products. The median company in our database has roughly 500 employees, but the distribution is heavily right-skewed: 15 companies have fewer than 200 people, while only 4 have more than 5,000.

A few things that jump out from the data: Databricks has the most open roles (807) of any company in our database despite already being 7,000 strong. Vast AI has the highest Glassdoor rating (5.0) at just ~50 employees — small teams often rate higher. And Cohere's 2.9 Glassdoor score stands out as the lowest in the set, a likely consequence of recent restructuring.

The Hypergrowth Club (Growing 50%+ YoY)

Seven companies in our database are scaling at breakneck pace, adding headcount at 50% or more year-over-year. These are the companies where everything — org charts, processes, culture — is being rebuilt in real time.

Hypergrowth sounds exciting from the outside, and it often is. But it creates a very specific set of trade-offs that anyone evaluating a job offer should understand.

The upside of hypergrowth

• Rapid promotions. When a company doubles in size every year, new teams and leadership roles appear constantly. Engineers who joined Anthropic at 400 people in early 2024 may now lead teams of 15. That kind of career acceleration is nearly impossible at a stable 8,000-person company like Stripe.
• Outsized impact. Early employees at these companies shape products used by millions. At OpenAI, a single engineer's decision on how the API handles a given edge case affects every developer building on the platform.
• Financial upside. Equity at a company growing this fast can be transformative. Anthropic's valuation went from $18B to $61.5B in roughly 12 months. Early equity holders have done very well.

The downside of hypergrowth

• Process chaos. When you're hiring hundreds of people per quarter, onboarding strains, documentation falls behind, and decision-making norms haven't had time to solidify. Multiple Glassdoor reviews at both OpenAI and CoreWeave cite "processes catching up to growth" as a key frustration.
• Role ambiguity. Your job description may bear little resemblance to what you're actually doing six months later. Teams get reorganized, priorities shift, and projects get killed as the company figures out what matters at its new scale.
• Culture dilution. When half the company has been there less than a year, the original culture can erode. Mistral AI, with 141 open roles for just ~100 employees, is literally trying to more than double its team — maintaining cultural coherence through that kind of growth is one of the hardest problems in management.

The bottom line: hypergrowth companies reward adaptable, self-directed people who thrive in ambiguity. If you need clear processes, defined career ladders, and stable teams, look at companies in the "stable" or "lean" categories instead.

The Lean Machines (Under 100 Employees)

At the other end of the spectrum, five companies in our database have fewer than 100 employees. These are some of the most interesting workplaces in tech — and the least forgiving.

Consider the contrast: Linear's 60 people build a project management tool used by thousands of engineering teams at companies like Vercel, Ramp, and PostHog. Databricks has 7,000 employees — 117x more people. Cursor's ~50-person team built the AI code editor that has become the default for a massive share of professional developers. These aren't companies that are small because they're early — they're small by design.

What working at a sub-100 company actually feels like

• Extreme ownership. There is no one to delegate to. If something breaks at 2 AM, you're probably the person fixing it. If a feature needs to ship, you're scoping it, building it, and monitoring it in production. This is exhilarating for some people and exhausting for others.
• Zero bureaucracy. No approval chains, no cross-functional alignment meetings, no quarterly planning theater. At Linear, the entire company can align on a direction in a single conversation. At a 7,000-person company, aligning on a direction takes a quarter.
• No career ladder. There is no "Staff Engineer" title to aim for, no promotion committee, no performance review cycle. Your growth comes from the expanding scope of what you build, not from a title change.
• Every hire matters enormously. At 50 people, one bad hire represents 2% of the company. The interview bar is often brutally high as a result. Cursor and Modal are known for particularly rigorous hiring processes.

If you're drawn to this kind of environment, also look at companies in the 100-200 range that share the same ethos: PostHog (~170, intentionally small), incident.io (~140), and Weaviate (~150, just 3 open roles — extremely selective).

The Post-Layoff Rebuilders

Several companies in our database went through significant layoffs in 2022–2024 and are now in various stages of recovery. Their employee counts tell one story; their Glassdoor ratings and open roles tell another.

Layoffs leave scars on company culture that can persist for years. Here's what the data shows about how these companies are recovering — and what it means for prospective employees.

The trust deficit

Cohere's 2.9 Glassdoor rating is the starkest example. When a company goes through restructuring, the employees who remain often report lower morale, increased workloads (doing the work of departed colleagues), and anxiety about future cuts. Cohere's 124 open roles suggest the company is trying to rebuild, but the low rating indicates trust hasn't fully recovered.

Stripe tells a more nuanced story. CEO Patrick Collison's public apology letter after the 2022 layoffs was widely praised for its honesty. The company has since rebuilt to ~8,000 employees with 527 open roles. Its 4.0 Glassdoor rating, while not stellar, is respectable — though notably lower than peers like Plaid (4.6) or Ramp (4.2) in the fintech space.

Airtable went from ~1,200 to ~900 after a 27% reduction. The company has stabilized with just 52 open roles — a sign that it's operating lean rather than aggressively rebuilding. Its 4.1 Glassdoor rating suggests the remaining team has found a sustainable groove.

What this means for job seekers

• Ask directly about the layoff. In interviews, it's completely fair to ask: "How has the team changed since the layoff? What's morale like now?" The quality of the answer tells you a lot about the company's self-awareness.
• Check the Glassdoor trajectory. A company with a 4.0 that used to be 4.5 before layoffs is in a different situation than one that's always been at 4.0. Look at review dates relative to the layoff date.
• Hiring intensity matters. Brex with 253 open roles for ~1,600 people (16% ratio) is hiring aggressively, which could mean either genuine recovery or churn-driven backfilling. Airtable with 52 roles for ~900 people (6% ratio) looks more stable.

What Team Size Means for Your Experience

Company size is one of the strongest predictors of your day-to-day work experience — often more than industry, mission, or even compensation. Here's what each tier actually feels like, based on employee reviews across our database.

One pattern worth noting: Glassdoor ratings tend to be higher at smaller companies. The average rating for sub-200 companies in our database is 4.2, compared to 4.0 for companies over 2,000 employees. This doesn't necessarily mean smaller companies are "better" — it may reflect self-selection (people who join tiny startups tend to be more aligned with the culture) and survivorship bias (unhappy employees at small companies leave rather than review).

The most important question isn't "what size is best?" but "what size is best for me, right now?" If you're early in your career and want mentorship and structure, a 2,000+ company with established engineering ladders might be ideal. If you're a senior engineer who wants to build something from scratch, a sub-100 team could be the most rewarding thing you ever do.

Open Roles vs Team Size: The Hiring Intensity Ratio

One of the most revealing metrics in this dataset isn't employee count or Glassdoor rating — it's the ratio of open roles to current headcount. We call this the "hiring intensity ratio," and it tells you how aggressively a company is trying to grow relative to its current size.

Mistral AI tops the list at a staggering 141% — the company is trying to hire more people than it currently employs. This is the kind of growth that completely transforms a company. If you join Mistral today, more than half your coworkers in six months will be people who aren't there yet. That's both an incredible opportunity (you'll be "senior" by tenure almost immediately) and a real risk (the company you join won't be the company you're working at a year from now).

Cursor at 116% is similarly aggressive. The company has found explosive product-market fit with its AI code editor and is racing to scale the team to match demand. Compare this to Stripe at 7% — a mature company doing measured, strategic hiring to backfill and incrementally grow. Both are valid strategies, but they produce radically different employee experiences.

The middle tier (15–30%) represents what might be called "healthy growth" — companies like Anthropic (30%), OpenAI (18%), and Databricks (12%) that are expanding meaningfully but not at a pace that completely disrupts existing culture. If you want growth-stage energy without startup-stage chaos, this is the sweet spot.

How to Use This Data

Employee count is one data point. To make a genuinely informed decision about where to work, combine it with culture signals, compensation data, and your own priorities. Here's how to go deeper with the tools we've built.

• Culture Directory — Full culture profiles for all 40 companies, including Glassdoor breakdowns, employee reviews, culture values, and open roles. Start here to understand any company's DNA.
• Compare Tool — Put any two companies side by side on culture, ratings, size, and values. Try Anthropic vs OpenAI or Cursor vs Linear to see how similar-looking companies actually differ.
• All 5,668 Jobs — Browse every open role across all 40 companies, with filters for culture values, role type, location, and seniority. If team size matters to you, filter by company to see only roles at companies of the size you prefer.

If you're evaluating multiple offers, team size should be a first-order consideration — not an afterthought. The difference between joining a 50-person company and a 5,000-person company is often bigger than the difference between industries, tech stacks, or even compensation packages. It shapes who you work with, how decisions get made, how fast you grow, and what your day actually looks like.

The data in this article will change quickly. AI companies are growing (and occasionally shrinking) at unprecedented rates. We refresh all job and company data daily and will update this article quarterly. If you spot an inaccuracy, let us know on Twitter.