Valuation has surged 80x in two years, with ARR exceeding $2 billion: Why has Mercor become one of the most profitable platforms in the AI era?

Over the past two years, the hottest topic in the AI industry has always revolved around models. OpenAI, Anthropic, xAI, and Google DeepMind have continuously pushed the ceiling of model capabilities; GPUs, computing power, and data centers have become new hotspots pursued by capital.

However, many people have overlooked a fact: what determines the capabilities of the next-generation model is not just computing power, but data.

Today, the most valuable data is no longer internet text, but high-quality reasoning data created by professionals such as doctors, lawyers, programmers, and researchers (Reasoning Data).

Whoever can organize these experts will hold the most critical means of production in the next stage of AI.

It is precisely for this reason that a company that barely builds models or sells GPUs has grown from a recruitment startup to an AI unicorn with an expected valuation of 20 billion US dollars in just two years.

It is Mercor.

01

Valuation Soared 80 Times in Two Years: Why Did Mercor Suddenly Explode?

Mercor was founded in 2023 by three young entrepreneurs who were all recipients of the Thiel Fellowship.

The company's latest financing round is underway. Market sources show that this round plans to raise approximately 500 million US dollars, targeting a valuation of 20 billion US dollars. If the deal is completed, it means Mercor's valuation will double again in less than a year. Previously, the company completed a 3.5 billion US dollars Series C financing in 2025 with a valuation of 10 billion US dollars; even earlier, its valuation was only around 250 million US dollars. In just two years, the valuation has increased by nearly 80 times.

What is even more astonishing is its business growth rate.

In June 2026, the company disclosed that its annualized revenue had exceeded 2 billion US dollars, doubling again compared to four months prior.

Today, Mercor has more than 30,000 professional experts, paying over 4 million US dollars in compensation to experts every day, making it one of the world's largest AI expert data platforms.

Its backing investors are also extremely prestigious, including top Silicon Valley funds such as Benchmark, General Catalyst, Felicis Ventures, Robinhood Ventures. As the company enters a new round of financing, Mercor has become one of the most high-profile AI infrastructure projects in the primary market.

02

What Mercor Does Is Not Recruitment

Many people who see Mercor for the first time mistakenly think it is an AI recruitment company.

In fact, that was only its initial entry point.

What Mercor truly does is to build a two-sided market that connects the world's top professional talents with AI labs.

As large language models enter the post-training era, relying solely on internet data can no longer continuously improve model capabilities. More and more models are beginning to rely on professionals such as doctors, lawyers, financial analysts, and software engineers to evaluate, correct, score, and provide reasoning guidance for model outputs.

These tasks are not traditional data labeling, but core training data that directly determines the professional capabilities of the model.

What Mercor provides is exactly such a high-quality expert supply system.

AI companies put forward their requirements, Mercor uses its AI interview and screening system to quickly match global experts, who then continuously participate in model training, evaluation, and reinforcement learning.

In other words, Mercor is not a recruitment platform, but building the talent infrastructure for the AI world.

03

How Does Mercor Make Money?

Mercor's business model is very simple, yet highly scalable.

AI labs purchase professional talent services from the platform on an hourly basis, and Mercor charges a platform service fee for each transaction. Current industry estimates generally put its platform commission rate at around 30%—35%.

For example, if an AI company wishes to hire a medical expert to participate in model training and pays 150 US dollars per hour, Mercor will pay approximately 95 US dollars of that to the expert, with the rest going to platform revenue.

Therefore, the publicly announced 2 billion US dollars ARR of Mercor is essentially the platform's annualized gross payment volume (GPV), not revenue in the traditional SaaS sense.

Calculated based on a platform commission rate of approximately 30%, the company's actual annualized revenue has reached around 600 million US dollars, which is already a staggering figure for a startup that was founded just two years ago. Compared to traditional headhunting firms that require extensive manual operations, Mercor uses AI to automatically complete expert screening, capability assessment, and project matching, significantly reducing delivery costs and enabling rapid platform expansion.

This model is closer to two-sided platforms such as Uber and Airbnb, rather than traditional software companies.

04

Why Are AI Companies Increasingly Dependent on Mercor?

In the past, large language models mainly relied on public internet text for pre-training, but as high-quality internet data is gradually exhausted, the industry has entered the post-training era.

What can truly continue to improve model capabilities is no longer more web pages, but expert knowledge.

Doctors can judge whether a medical diagnosis is correct, lawyers can evaluate whether legal analysis is rigorous, senior programmers can assess code quality, and researchers can help models understand complex scientific problems. These professional judgments are things that models cannot learn solely from online text.

As a result, various AI labs have begun to purchase professional talents on a large scale.

Mercor is right at the center of this trend. It does not help companies recruit employees, but helps companies like OpenAI and Anthropic continuously acquire the human knowledge needed to train the next generation of models. As model capabilities continue to deepen into professional fields, the expert network provided by Mercor will also continue to expand, which gives the platform itself an increasingly strong network effect.

05

Why Was Mercor Able to Quickly Build Moats?

Theoretically, anyone can build an expert database; what is truly difficult is how to efficiently verify expert capabilities and continuously complete large-scale global supply-demand matching.

From the early days of its startup, Mercor has integrated AI into the entire screening process: the platform uses AI to automatically complete expert interviews, resume analysis, skill assessment, and project matching, greatly shortening the traditional recruitment cycle.

At the same time, as more and more AI companies publish projects through Mercor, the platform has accumulated a large amount of data on different experts' capabilities, project performance, and training effects.

These data continuously optimize the platform's matching efficiency, making it increasingly difficult for latecomers to replicate.

For AI companies, what matters even more is speed. A model company may need to find hundreds of people with professional backgrounds to participate in training within a few days, a task that traditional headhunters can barely accomplish.

What Mercor has built is not just a list of experts, but an infrastructure that can continuously, quickly, and high-quality output expert capabilities.

06

Why Are Investors Continuously Increasing Their Stakes?

Mercor's growth rate essentially reflects the changes in the demand structure of the AI industry.

In the past few years, capital has mainly focused on the models themselves. Today, more and more investment institutions are beginning to realize that the real bottleneck limiting the improvement of AI capabilities is no longer the model, but high-quality expert data.

Mercor happens to occupy this critical node. Compared with a single AI application, it serves almost all cutting-edge AI labs. As companies such as OpenAI, Anthropic, Google DeepMind, and xAI continue to increase their post-training investments, the demand for expert data will also continue to expand.

At the same time, Mercor has gradually expanded from a pure talent platform to the entire AI data production infrastructure. Recently, the company announced the acquisition of Deeptune, an AI Agent training company, to further improve its AI data production capabilities, demonstrating its long-term strategy of building a complete AI data supply chain.

For investment institutions, this means that Mercor is not a company that benefits short-term from the AI boom, but has the potential to grow into a new infrastructure platform in the AI era.

07

The Truly Noteworthy Aspects of Mercor

Many people look at Mercor and think it has simply moved recruitment into the AI era, but what truly attracts capital is not the recruitment business itself, but its position at the core of the AI industry chain.

In the next few years, the focus of competition in the AI industry will likely gradually shift from model capabilities to high-quality data, professional knowledge, and continuous optimization capabilities, and Mercor is becoming the most important bridge connecting global experts and AI models.

It neither trains models nor manufactures chips, but holds an increasingly important resource in the future development of AI — human professional knowledge. As AI continues to enter professional fields such as healthcare, law, finance, and scientific research, the demand for high-quality expert data will only grow larger.

Conclusion

The Next Competition in AI Is Not Just About Models, But About Knowledge

In the next few years, the more important question may become: Who can continuously access the highest-quality human knowledge.

The value of Mercor lies not in its development of new models, but in its ongoing construction of a new type of infrastructure connecting global professional talents and AI.

If GPUs are the computing power of AI and data centers are the factories of AI, then companies like Mercor are becoming the knowledge supply chain of the AI world.

This is why more and more top investment institutions are willing to continuously bet on this company.

This article is from the WeChat Official Account "Silicon Rabbit", author: Silicon Rabbit, published with authorization from 36Kr.