Mercor’s last confirmed financing valued it at $10 billion after a $350 million Series C announced on October 27, 2025. In July 2026, the company was reportedly discussing a possible financing at a $20 billion valuation, but those talks were described as early-stage—not a completed round. The more important shift is in the business itself: Mercor has grown from an AI-assisted recruiting startup into a platform supplying expert labor, training data, evaluation, and related tools to AI companies and enterprises.
What Mercor does now
Mercor connects organizations developing or deploying AI with people who have specialized professional knowledge. Its work can include generating training data, reviewing model answers, ranking outputs, diagnosing errors, testing AI agents, and demonstrating professional workflows. The company is not an AI model developer; it supplies human expertise and related infrastructure.
Mercor’s own product categories—Work, Build, Hire, and Evaluate—reflect a broader scope than conventional recruiting. Its mission and product description also present enterprise AI-agent development: helping organizations encode internal knowledge, workflows, and standards into custom agents. These descriptions establish Mercor’s intended offering, not independent evidence of adoption or performance.
How a recruiting startup became an AI-work platform
Founded in 2023, Mercor initially focused on AI-assisted hiring. Its early product automated resume review, candidate matching, AI interviews, and payroll workflows, first targeting software engineers and other technology workers. Reporting on its February 2025 Series B described a $100 million round at a $2 billion valuation and the company’s recruiting product. TechCrunch’s account of that round also discussed the platform’s hiring features.
Demand from AI labs expanded the use case. Instead of only helping employers fill conventional roles, Mercor began connecting labs with specialists—including doctors, lawyers, bankers, scientists, and engineers—to help train and evaluate models. The shift matters: the investment thesis is now less about automating recruitment and more about organizing skilled human input for AI development.
How Mercor’s valuation climbed
| Date | Financing or reported valuation | What it indicates |
|---|---|---|
| 2023 | $3.6 million seed round | General Catalyst-led seed financing, according to TechCrunch’s company history. |
| 2024 | $32 million Series A at a $250 million valuation | Benchmark-backed expansion, reported by TechCrunch. |
| February 20, 2025 | $100 million Series B at a $2 billion valuation | An eightfold increase from the reported Series A valuation, per TechCrunch. |
| October 27, 2025 | $350 million Series C at a $10 billion valuation | Five times the Series B valuation. Mercor announced the round, led by Felicis with Benchmark, General Catalyst, and Robinhood Ventures participating; Mercor’s announcement and TechCrunch’s report confirm it. |
| July 9, 2026 | Possible $20 billion valuation discussed | Early-stage financing talks, not a completed financing, according to TechCrunch. |
A financing valuation is the price implied by a private funding round, not a continuously quoted public-market value. As of August 18, 2026, $10 billion is the last confirmed financing valuation in the cited reporting; the possible $20 billion figure remains a reported discussion.
Why AI companies need specialized human judgment
As AI systems move beyond generic text generation, improving them can require more than basic labeling. A model may need expert feedback on whether an explanation is sound, a professional workflow is complete, a tool was used appropriately, or an answer handles a consequential scenario responsibly. Specialists can supply domain-specific reasoning, preference rankings, error diagnosis, and examples of how work is actually performed.
That creates a potential outsourcing opportunity. An AI lab can build models internally while relying on an outside operator to find and screen people, verify credentials, set up contracts and payments, manage projects, and review work quality. Mercor’s bet is that this operational layer can make scarce expertise easier to procure at scale.
Mercor also argues that automation may increase the value of people who provide judgment, oversight, and domain knowledge. That is the company’s strategic thesis, not a settled labor-market outcome: AI could increase demand for expert evaluation in some areas while reducing the need for human input in others.
What the reported numbers show—and do not show
In October 2025, TechCrunch reported company-provided figures that Mercor paid contractors more than $1.5 million per day and had more than 30,000 experts earning an average of over $85 per hour. Those figures are not audited financial disclosures. Mercor’s newsroom now claims $4 million in daily payments to its expert network, more than 400 employees, and more than 5 million domain experts. The newer claims are company-reported, and the expert-count definition may differ substantially from the earlier roster figure; the two sets of numbers should not be treated as directly comparable. October 2025 reporting · Mercor newsroom
In July 2026, TechCrunch reported that CEO Brendan Foody said Mercor’s annualized revenue run rate had crossed $2 billion and had reportedly doubled in four months. The report does not establish whether that figure means net revenue retained by Mercor, gross customer spend, bookings, or another measure, nor does it present it as audited revenue. Earlier reporting described a model that charged clients hourly finder’s fees and matching fees. July 2026 report · February 2025 report
This distinction is central to valuation. If a reported run rate includes money paid through to contractors, it is not economically equivalent to the same amount of recurring software revenue. Marketplace volume can be large while the platform retains only a portion, and it also incurs costs to recruit, verify, support, pay, and quality-check workers. Without a clear revenue definition and margin data, a valuation-to-revenue multiple would be misleading.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errors- What portion of customer spending does Mercor retain as net revenue?
- What are gross margin and contribution margin after contractor payments, quality control, support, processing, and compliance?
- How much demand comes from each customer, and how much revenue is recurring rather than project-based?
- How often do customers return, and how much work is contracted or committed rather than extrapolated from recent activity?
APEX and the possibility of a software layer
Mercor presents APEX as a system for assessing whether AI can perform economically valuable work. Its research-product listing includes APEX Benchmarks, APEX-Agents, APEX-Accounting, and APEX-SWE.
Evaluation could be more durable than one-off data projects if companies repeatedly use a trusted measure across model releases or require evidence before deploying agents. A benchmark might also give developers and buyers a shared way to discuss performance. But Mercor’s product description alone does not establish that APEX is an industry standard, independently authoritative, or widely adopted.
The moat question: network, workflow, or labor brokerage?
A large expert pool could help Mercor find appropriate candidates faster, improve screening and matching, and fulfill projects more reliably. Historical performance data might make those matches better over time, while enterprise integrations could make the service harder to replace. Those are plausible mechanisms, not proof of durable network effects.
The critical test is whether customers rely on proprietary data, workflow integration, and demonstrated quality—or mainly buy access to contractors they could source elsewhere. Recruiting, interviewing, payment, and task-management tools can be copied. A durable advantage would need to combine verified expertise, trustworthy performance histories, reliable quality control, and customer dependence.
Mercor’s competitive set therefore spans more than recruiting apps. Scale AI, Surge AI, and Turing compete in overlapping areas of data, model training, evaluation, or technical talent; traditional staffing firms, expert networks, freelance marketplaces, professional-services firms, and in-house AI-lab teams can also address parts of the same need. TechCrunch’s September 2025 report identified Scale AI, Surge AI, and Turing as relevant competitors. Comparative market shares and margins are not established by that reporting.
The business could ultimately be valued as a labor marketplace, a provider of completed data and evaluations, a software workflow platform, or an enterprise agent partner. Those models have different economics. The more revenue depends on labor fulfillment and contractor pass-throughs, the harder it is to justify software-like margins; software and repeatable evaluation products could change that mix if they become meaningful and defensible.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Risks that can weaken the investment case
Revenue quality and customer concentration
High gross payment volume does not reveal how much value Mercor retains. A concentrated customer base would add another vulnerability: a large AI lab could reduce projects, change vendors, build its own expert network, or renegotiate terms. Publicly available figures cited here do not establish customer concentration, retention, or take rates.
Automation and disintermediation
AI labs may recruit experts directly or build internal evaluation teams. Better models might also generate synthetic training data or evaluate some outputs themselves, reducing demand for certain tasks. On the other hand, more capable models could increase demand for difficult expert-led tests. The direction and scale of either effect are uncertain.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Quality, fraud, and trust
A global expert marketplace must ensure that the person doing the work is qualified and is the same person who passed screening. It also has to guard against credential fraud, plagiarism, AI-generated submissions, inconsistent grading, conflicts of interest, and feedback optimized for task completion rather than accuracy. For high-stakes work, confidentiality and legal rights to use resulting data matter as much as matching speed.
Security, labor, and compliance
Mercor’s newsroom lists a June 25, 2026 update concerning a security incident, and July 2026 reporting also refers to an earlier data breach. The available cited material does not establish the incident’s technical details or full impact. Mercor’s newsroom · TechCrunch, July 9, 2026
Using contractors across jurisdictions also creates potential worker-classification, wage, tax, benefits, intellectual-property, confidentiality, licensing, export-control, and sensitive-information obligations. TechCrunch’s July 2026 report says several contract workers filed lawsuits, but the cited account does not provide enough detail to assess the claims or their status. It would be inappropriate to draw a legal conclusion from that report alone.
What enterprise buyers should verify
Mercor’s enterprise offering appears sales-led rather than a transparent, standardized self-serve product. Its enterprise contact page, company site, expert work portal, and product information describe routes into the service, but public standardized enterprise pricing was not established in the cited sources.
For an organization considering Mercor for expert work, AI evaluation, or agent development, diligence should cover:
- Whether pricing is based on gross project spend, hourly rates, or a separate platform fee, and what portion goes to experts.
- Minimum project size, quality guarantees, rework or replacement terms, and service-level commitments.
- How experts are verified, how work is reviewed, and how the company handles fraud or inconsistent results.
- Ownership and permitted use of prompts, task outputs, evaluation data, and other project materials.
- Data retention and deletion, confidentiality protections, security documentation, and safeguards for sensitive professional information.
- Worker classification, tax, licensing, and cross-border compliance relevant to the project’s jurisdictions.
- References from customers using the service for the same professional domain and use case.
Does Mercor signal a new era in talent acquisition?
Mercor’s trajectory points to a real change in how talent platforms may serve AI companies: matching experts to short-cycle training, evaluation, and workflow tasks rather than only placing people in conventional jobs. But the $10 billion valuation is a financing milestone, not proof that this model has software economics or a durable moat. The decisive issue is whether Mercor can turn access to scarce human expertise into repeatable, trusted infrastructure—or remains an intermediary whose growth depends on contractor-heavy projects and a small set of large buyers.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
