Mercor Raises $750 Million at $20 Billion Valuation to Train AI for Professional Work
Read Time:5 Minute, 9 Second
Mercor has raised $750 million in a Series D round at a $20 billion valuation, the AI training company announced on November 10. The San Francisco-based business connects subject experts with companies developing AI. It now plans to put more resources into measuring and improving agents that carry out professional tasks.
Reuters confirmed the funding announcement, reporting that investors included Benchmark, Felicis, Fundomo, Nvidia and funds managed by Apollo affiliates. The round follows a $350 million raise at an approximately $10 billion valuation in October 2025. The new transaction doubles that earlier valuation.
For Indian professionals watching AI-related work, the announcement points to continued spending on specialist knowledge. The funding release gives no India vacancy count, country allocation or new pay schedule. The confirmed development is the completed funding round. Recruitment decisions and project availability will need separate verification.
The funding facts
| Item | Confirmed detail |
|---|---|
| Round | Series D announced on November 10, 2026 |
| Amount raised | $750 million |
| Valuation | $20 billion |
| Headquarters | San Francisco |
| Earlier round | $350 million at approximately $10 billion in October 2025 |
| India hiring numbers | Undisclosed |
The valuation reflects the price attached to the company in the funding transaction. It is separate from the money raised, annual revenue and the budget for any particular project. The announcement also provides no allocation showing how much of the proceeds will go to engineering, research, expert payments or other costs.
What Mercor is building next
In its Series D announcement, Mercor says it will invest in evaluation, post-training and simulated workplaces where agents can practise complex tasks. Post-training means improving an existing model after its initial training. Evaluation tests how reliably it performs against defined requirements.
Mercor also disclosed an employee tender as part of the round, giving employees an opportunity to sell shares. It said it had expanded beyond San Francisco with offices in New York and London. These developments describe the business and its staff. The announcement offers no timetable for a new India office.
Mercor’s expansion poses a practical question for customers. Before giving an agent access to live work, how can they tell whether it will complete the task correctly? A convincing response to a prompt can still leave that question unanswered. Buyers need evidence about the work the system produces, the errors it makes and the time people spend checking it.
Why human judgment is part of AI evaluation
Mercor’s October research agenda describes using domain experts to create tests, study model failures and develop training data. Its APEX work covers management consulting, investment banking, corporate law, accounting and software engineering. The company says it wants to assess models in realistic working conditions, including whether they hide mistakes or exploit flawed incentives.
That focus makes the definition of success important. Consider a proposed test of an agent preparing a report. A reviewer would need to decide which facts must appear, what evidence supports them and which omissions would make the report unusable. A simple score for fluent writing would miss much of the professional judgment involved.
This is an illustration of the evaluation problem, rather than a disclosed Mercor assignment. Actual projects can have different tasks and standards. The relevant skill is being able to explain why an output meets a requirement, document the reasoning and apply the same standard consistently.
How simulated workplaces fit
Mercor announced an agreement to acquire Deeptune in July. That announcement described a combination of software environments, assigned tasks and checks that determine whether an agent succeeded. Deeptune had recreated hundreds of enterprise applications, while Mercor supplied expert input for tasks and scoring.
The July release said Deeptune’s team would join Mercor and expand its New York operation. By November, Mercor’s funding announcement referred to Deeptune as an acquisition it had made earlier in the year. That sequence distinguishes the earlier plan from the status the company now reports.
For a customer, a simulated workplace offers a way to test an agent before letting it operate inside a live system. The value still depends on how closely the test matches the real work. Unusual documents, missing information and conflicting instructions can expose weaknesses that a neat demonstration fails to show.
What Indian applicants should look for
The funding news gives applicants a reason to follow Mercor’s official opportunities. It leaves eligibility and compensation to individual listings. An applicant in Hyderabad or Bengaluru would need a role that accepts their location, matches their expertise and provides workable terms.
A clear listing should let a person establish the task, expected hours, contract type and payment basis before deciding whether to apply. Those details affect the value of an opportunity more directly than the company’s latest valuation. A specialist reviewing technical output may also need to understand the evidence and confidentiality requirements of the assignment.
BoringDude’s recent coverage of Simplify Alpha’s Hyderabad engineering centre illustrates another kind of workforce announcement. That story includes a local staffing target over two years. Mercor’s latest release supplies financial and product plans, with no comparable India staffing target.
What happens next
The next useful evidence will be specific projects, product releases and disclosures showing how customers use the evaluation work. For applicants, current listings will establish which opportunities are open. For enterprise buyers, the important questions include reliability, review effort and the cost of running agents on their own processes.
Mercor has secured more capital to pursue that business. Its next milestones will show how the investment translates into services customers can use and assignments professionals can evaluate on their own terms. The funding announcement establishes the scale of the transaction, while those operating details remain to be disclosed.
Frequently asked questions
How much did Mercor raise in November 2026?
Mercor announced a $750 million Series D round on November 10, 2026, at a $20 billion valuation.
Did Mercor announce new jobs in India?
The funding announcement gives no India vacancy count, country allocation or new pay schedule. Applicants need to check individual official listings.
What will Mercor use the funding for?
Mercor says it will invest in AI evaluation, post-training and environments where agents practise professional workflows.
Is the $20 billion valuation Mercor’s revenue?
No. It is the valuation attached to the funding transaction. The announcement does not report $20 billion in revenue.
Happy
0 %
Sad
0 %
Excited
0 %
Sleepy
0 %
Angry
0 %
Surprise
0 %
0
0
Average Rating