Benchmarking fast-growing AI-native startups
What the data reveals about 17 of the fastest growing AI-native companies
👋 Welcome to AI-native GTM!
AI-native companies are re-writing the GTM playbook. Every other week, I will highlight the stories, frameworks and patterns behind some of today’s fastest growing startups. You can expect deep dives, analysis and insights to inspire the next generation of AI-native founders and operators.
Top AI-native startups are reaching significant revenue milestones in record time. In the past two weeks alone Gamma and Sierra both reached $100M ARR and - not to be outdone - Lovable reached $200M ARR, doubling its revenue in just 4 months!
I analyzed 17 of the fastest growing AI-native startups across key performance metrics to see what makes them different from the rest. The analysis includes revenue and funding milestones, employee count, growth velocity, funding efficiency, and operational efficiency.
These companies—15 venture-backed and two bootstrapped—have all been founded in the last five years and crossed the $50 million ARR threshold, making them among the most successful in the current AI wave.
They are the following: ElevenLabs, Lovable, Gamma, Mercor, Harvey, Sierra, Genspark, Anysphere, Midjourney, Bolt, Perplexity, Manus, Heygen, Higgsfield AI, Cognition, Surge AI and Together AI. Here is the link to the dataset.
Let’s dive in 👇
Revenue Milestones
Let’s start with the most recent publicly announced annual revenue numbers achieved by these companies.
Top Tier ($500M+):
Surge AI leads with $1.4 billion in revenue.
Anysphere, Midjourney, and Mercor each achieved $500 million in revenue, forming a second tier of high performers.
Mid Tier ($100M-$200M):
ElevenLabs and Lovable at $200 million.
Perplexity at $150 million.
Heygen, Gamma, and Harvey at $100 million each.
Emerging Tier ($40M-$90M):
Companies like Manus ($90M), Cognition ($80M), and Bolt ($40M) represent rapidly growing but earlier-stage players.
This is clearly not an apples to apples comparison: announcement dates vary, revenue numbers are not all current and we are not yet factoring time to revenue or company stage. More to be revealed below.
Revenue Velocity
Let’s even the playing field and look at Revenue Velocity which calculates the average revenue added per month to reach the announced milestones. For lack of more accurate data, we are calculating this starting from their founding dates. This will provide a better insight into the pace of revenue generation.
Velocity Leaders:
Surge AI - $22.2M/month (exceptional outlier)
Mercor - $15.6M/month
Anysphere - $14.7M/month
Midjourney - $11.1M/month
High Velocity (>$4M/month):
Lovable - $8.3M/month
Bolt - $8.1M/month
ElevenLabs - $5.0M/month
Perplexity - $4.4M/month
The mean velocity is $6.6M/month, while the median is $3.7M/month, again showing significant variance in growth rates. Surge AI’s velocity is more than 3x high velocity performers, demonstrating an exceptional pace of revenue generation.
Funding Milestones
Funding levels vary dramatically, from bootstrapped to heavily funded.
Mega-Funded (>$1B):
Anysphere - $3.3B
Perplexity - $1.22B
Harvey - $1.01B
Very Well-Funded ($300M-$700M):
Cognition - $696M
Sierra - $635M
Together AI - $533M
Mercor - $486M
Genspark - $360M
Well-Funded (<$300M):
ElevenLabs - $291M
Lovable - $224M
Bootstrapped:
Surge AI - $0
Midjourney - $0
Capital Efficiency
Let’s look at revenue per dollar raised to see how these companies are tracking towards return on investment. Surge AI and Midjourney are excluded since they are bootstrapped. Keep in mind that the data doesn’t factor in the timing of funding vs. when the revenue milestones were actually reached. That said, it still provides an interesting picture.
Very Capital Efficient (>1.0x):
Heygen - 1.52x
Gamma - 1.15x
Manus - 1.06x
Mercor - 1.03x
Higgsfield AI - 1.00x
These companies have generated more revenue than they’ve raised, demonstrating great underlying unit economics.
Rapidly Approaching Efficiency (0.5x-1.0x):
Lovable - 0.89x
ElevenLabs - 0.69x
Relatively Capital Intensive (<0.5x):
Bolt - 0.37x
Together AI - 0.19x
Sierra - 0.16x
Harvey - 0.10x (lowest efficiency)
Keep it mind that it generally took years for SaaS companies to hit these types of numbers. Numbers <1.0x are typical for high-growth technology companies prioritizing market capture over immediate returns.
Employee Count
Team sizes reflect different scaling strategies. Note that there is a margin or error here as these numbers were gleaned from a combination of public announcements and LinkedIn sleuthing.
Larger Teams (200+ employees):
ElevenLabs - 400
Harvey - 350
Anysphere - 300
Heygen - 260
Together AI - 250
Medium Teams (100-200 employees):
Sierra and Cognition - 200 each
Midjourney - 170
Surge AI - 130
Manus - 120
Lean Teams (<100 employees):
Lovable - 45
Bolt - 24 (smallest team)
Genspark - 30
Revenue per Employee
This metric reveals dramatic differences in workforce productivity. It should be noted that a number of these companies (e.g Mercor) have a large number of contractors that work for them so these numbers are mostly reflective of core full-time teams.
Ultra-Efficient (>$4M per employee):
Mercor - $16.7M per employee (exceptional efficiency)
Surge AI - $10.8M per employee
Lovable - $4.4M per employee
Highly Efficient ($2M-$3M per employee):
Midjourney - $2.9M per employee
Perplexity - $2.5M per employee
Gamma - $2.0M per employee
Very Efficient ($500K-$2M per employee):
Anysphere - $1.7M per employee
Bolt - $1.7M per employee
Higgsfield AI - $829K per employee
The mean revenue per employee is $2.8M, with a median of $1.7M. Mercor’s efficiency is nearly 6x the median, suggesting unique business model advantages. Typical revenue per employee benchmarks for SaaS companies were generally $100,000–$150,000 for early-stage companies, rising to $200,000–$300,000 for scaling and mature companies, and $250,000–$400,000+ for top performers going public. AI-native companies are built different.
Synthesis: Comparative ranking by profile
The dataset shows remarkable diversity in growth strategies and efficiency metrics. That said, using the metrics above, these companies naturally cluster into a few profiles.
Hyper-scaled, highly efficient leaders
Surge AI – Top in absolute revenue ($1.4B), top velocity, top-tier revenue per employee, all bootstrapped.
Midjourney – $500M revenue, strong velocity and revenue per employee, also bootstrapped.
Elite high-velocity, lean teams
Mercor – $500M in 32 months, very high velocity (~$15.6M/month), highest revenue per employee, and strong revenue per $ raised.
Lovable – Fast to $200M (24 months), high velocity and solid efficiency both per employee and per dollar.
More capital-intensive scale plays
Anysphere, Perplexity, Harvey, Together AI, Sierra, Cognition – Large funding rounds and solid revenue, but lower revenue per $ raised, consistent with aggressive investment in product, infra, or market capture.
Early fast movers
Bolt, Genspark, Higgsfield AI, Manus – Small to mid-sized milestones but very quick time-to-milestone and good capital efficiency, particularly Bolt (5 months to $40M) and Higgsfield AI/Manus (≈$1+ revenue per $ raised).
Efficiently funded mid-scale
Gamma, Heygen, Manus, Mercor, Higgsfield AI – All above $1 revenue per $ raised, mixing reasonable funding with strong monetization.
We are still in the early innings of the AI-native race and awesome companies are being built every day and everywhere. While limited and certainly somewhat flawed, the dataset suggests that execution speed and operational efficiency may be more important than absolute funding levels for achieving revenue milestones. Founders and investors beware.
This analysis is based on public information, interviews, and company materials as of November 2025. Some details may have changed since publication.
I wonder about other AI companies that has managed to become an integral part of the AI ecosystem and yet have not reached this list - cursor or glean
The revenue per employe numbers are wild, especially Midjorney's bootstraped approach hitting $2.9M per. Shows how much AI-native companys can achive with lean teams.