Is Mercor the Future of the (White Collar) Gig Economy?

The gig economy has traditionally meant platforms like Uber, Upwork, or Fiverr connecting people to short-term jobs.

Recently, a new player called Mercor is making waves by applying that concept to high-skilled, knowledge-based work, especially in training artificial intelligence.

Mercor’s rapid rise, innovative model, and hefty valuation have prompted many to ask whether it represents the next evolution of freelance work.

Mercor’s Platform and Business Model: A New Type of Gig Marketplace

Mercor is a labor marketplace at the intersection of AI research and skilled freelance work. Founded in 2023 by Brendan Foody and co-founders, Mercor quickly attracted major venture funding (and attention). In less than three years, Mercor claims to have tens of thousands of contractors on its platform and an annualized recurring revenue of about $500 million, with a recent funding round valuing it at $10 billion.

This explosive growth suggests that Mercor has tapped into a strong demand. But what exactly does Mercor do differently?

At its core, Mercor connects human expertise with AI development needs. The platform pays skilled professionals (often industry experts) up to $200 an hour to perform structured tasks like filling out forms, writing detailed reports, or evaluating AI outputs. These tasks are not the typical “gig” errands or design projects; instead, they are designed to train or improve AI models. For example, a former banking analyst might simulate filling in complex financial forms or a lawyer might draft sample documents, all to provide data that helps an AI learn those workflows.

Mercor acts as the middleman: AI labs and tech companies pay Mercor for access to this tailored data, and Mercor, in turn, pays the freelance “experts.” The CEO, Brendan Foody, describes Mercor’s marketplace as one of the primary channels connecting former employees of industries such as investment banking, consulting, and law with AI labs seeking to automate these sectors.

This model is lucrative because AI developers desperately need high-quality, domain-specific data, which incumbent companies are often unwilling to share. Rather than negotiating expensive data partnerships (or being denied outright), AI labs can use Mercor to hire individuals who possess the domain knowledge gained from working in those companies. In Foody’s words, many big firms “don’t love the idea of having [AI] models that automate their value chain,” so they won’t share data – but former employees can leverage the knowledge in their heads to help train such models. Mercor bets that an employee’s know-how belongs to the employee, not the employer. This “human data” pipeline enables AI labs to bypass corporate red tape and obtain realistic training data, thereby accelerating AI development.

Mercor’s positioning is unique: it has been described as building a “gig economy of expertise”, analogous to how Uber created a gig economy for transportation.

The tasks on Mercor are white-collar, knowledge-intensive gigs, a far cry from the simple data-labeling micro-tasks often outsourced to anonymous click-workers. In fact, Mercor was one of the first data startups to recruit highly skilled U.S. knowledge workers and pay them large sums to train AI models, whereas early AI data firms (like Scale AI) hired crowdsourced labor in developing countries for pennies. This shift has earned Mercor high-profile customers (OpenAI, Anthropic, and Meta are noted clients) and allowed it to pay its contractors $1.5 million per day, all while remaining profitable because the AI labs are willing to pay a premium for the valuable data.

On Mercor’s own site, the range of available roles is telling. Recent listings included “U.S. Physicians & PAs – 30–45 min research survey ($200/hr)”, “Finance Expert ($150/hr)”, “Advanced Git Specialist – Workflow Evaluation ($90–$150/hr)”, and “Biology PhDs (Wet Lab Experience) ($60–$80/hr)”.

Or, “Basketball experts — avid fans, sports journalists, commentators, and former or semi-professional players — to evaluate basketball games”. What’s the task? “You’ll watch basketball games and answer questions in real time, assessing the quality, depth, and accuracy of AI insights, helping us refine our AI’s basketball reasoning, storytelling, and strategic understanding” for $45-$70 per hour.

Many roles involve hiring dozens or even hundreds of contractors, indicating a significant project scale. For instance, 455 finance experts were recently hired at $150/hr for a particular project.

Mercor vs. Traditional Gig Platforms (Upwork, Fiverr)

To understand Mercor’s impact, it is helpful to compare it with the incumbent gig economy leaders, such as Upwork and Fiverr. These platforms are two of the largest freelance marketplaces globally and offer a broad bazaar of services, from graphic design and writing to programming and administrative work – essentially “all skills, all industries”.

The key difference is that these platforms are largely self-service: clients scout talent and manage projects, while the platform provides tools and takes a fee.

Mercor, by contrast, is far more specialized and hands-on. It does not present a smorgasbord of categories or a project catalog for anything under the sun. Instead, Mercor concentrates on roles related to AI development and data. Rather than letting freelancers and clients simply find each other, Mercor’s model is to actively source and vet experts, then match them to gigs defined by Mercor’s client companies. This makes Mercor feel more like an elite recruitment or outsourcing firm than an open marketplace. In other words, Upwork excels in breadth and flexibility, whereas Mercor narrows in on depth and fit.

Another difference is talent vetting and quality control. On Upwork or Fiverr, virtually anyone can sign up and start offering services (with some basic identity verification on Upwork). Quality is enforced indirectly through client reviews and platform rules. Mercor, conversely, selects for high-skill talent upfront. The contractors on Mercor often hold advanced degrees or industry pedigrees (e.g., former Goldman Sachs bankers, former McKinsey consultants, PhDs, etc.).

Mercor’s screening (including tests or trials) is more akin to a job application process than simply setting up a profile. In fact, Mercor often invites applicants to a paid trial project as part of vetting – an uncommon practice in freelancing. One freelancer’s account describes being “invited to complete a paid work trial, involving evaluating AI model answers against rubrics” and getting $300 for the onboarding task once it was approved.

Pay structure at Mercor is also quite different. Upwork and Fiverr offer a vast range of pay rates, often influenced by global market dynamics. Many gigs on those sites can pay relatively low (e.g., single-digit dollars per hour for commoditized tasks), especially due to competition from lower-cost regions. There are also high-paying opportunities, but one common complaint on traditional platforms is the tendency for prices to drop to the lowest common denominator. Mercor flips this script by advertising high fixed rates for each role, typically ranging from $60 to $200 per hour, as seen in their listings. These rates are set by the platform/client, not negotiated by the freelancer in a bidding war.

However, Mercor’s model also restricts participation more than Upwork or Fiverr do. Many Mercor projects are only open to certain geographies or backgrounds. Some freelancers have voiced frustration that Mercor’s “remote” jobs often require you to be based in the US, Canada, the UK, or similar, excluding many willing workers elsewhere. Upwork and Fiverr are globally inclusive (aside from sanctioned regions), so opportunities are more evenly distributed worldwide on those platforms (which, conversely, contributes to the lower wage competition). Mercor’s selective approach can leave qualified individuals in countries like Spain or India waiting on the sidelines for a chance.

This highlights an issue of scalability and inclusivity: Mercor may need to expand its geographic reach as it grows, but doing so could necessitate additional controls to manage quality and compliance.

Platform Transparency, Pay Structure, and User Experience

A crucial aspect of any gig platform’s success is how it treats its users (both the clients and the workers). Mercor earns praise in some areas here, but also faces scrutiny:

Transparency and Trust: Mercor has positioned itself as a worker-friendly and transparent platform – for instance, by not charging freelancers to join, paying for trial work, and posting clear pay rates per role.

However, there have been questions about Mercor’s honesty in its recruitment process. In mid-2024, some job seekers complained that Mercor (and a similar startup, micro1) acted like “ghost recruiters” – stringing candidates along without ever making a real offer. A viral LinkedIn post accused Mercor of collecting personal data from interviews and forms just to train their AI models, with no intention of hiring the candidates. The author observed that Mercor would have multiple recruiters interview someone, tell them they were in a talent pool, and even market the person’s profile to clients, but never actually offer them a job or payment. All the while, Mercor would allegedly record the interviews and harvest the data. This severe allegation essentially portrayed Mercor as running a scam, exploiting eager job seekers by providing them with free data.

There’s also the issue of corporate transparency and ethics. There are also real concerns around transparency and ethics. Mercor operates on a thin line between monetizing individual expertise and risking the leakage of proprietary information. While the company says it instructs contractors not to share confidential materials, some job postings have pushed uncomfortable boundaries, such as seeking access to live production codebases for AI training. Clients whose former employees participate often have no visibility into this work, raising concerns about consent and the protection of trade secrets. Although no major legal cases have surfaced yet, Mercor’s stance that “knowledge belongs to the individual” may face pushback as the platform scales, making careful governance essential to its long-term credibility.

Pay Structure and Fairness: As noted, Mercor offers excellent pay on paper for many roles. That said, not everyone on Mercor is making $150 an hour. The flashy rates apply to certain contracts, but Mercor also handles large-volume data labeling projects that pay more modestly. A recent revelation provided a stark example: Mercor had a contract with Meta (Facebook), codenamed “Musen,” involving approximately 5,000 data labelers at $21/hour. In late 2025, Mercor abruptly shut down the project and laid off the contractors without warning, despite the project’s expected duration. To add insult to injury, Mercor offered the workers a chance to return to a new project (“Nova”) that was almost identical to the old one, but at a pay rate of only $16/hour (approximately 24% less). Workers described the move as “demoralizing,” and many felt they had no choice but to accept the lower pay because they needed the income to support themselves and their families. Mercor justified the pay cut by saying the new project had steadier work and higher hour caps (letting people work more hours), but effectively this was a cost-saving maneuver that shifted risk onto the labor force.

This episode shows a more troubling side of Mercor’s gig model. Despite its utopian branding of “creating abundance for everyone” in the AI economy, Mercor is still a business that may exploit workers when convenient. The dynamic is reminiscent of other gig economy practices (e.g., rideshare companies changing pay algorithms or delivery apps cutting bonuses), which have often left workers feeling expendable. In Mercor’s case, highly skilled contractors may enjoy premium rates. Still, those performing more routine labeling or evaluation work – effectively the “ghost workers” behind AI – can face instability and downward pressure on their pay.

It’s yet another example that gig work in any form can lead to precariousness. As one analysis in Futurism put it, “the number of ‘freelance workers’ is soaring [in AI]… jobs are being degraded… rising tides on Wall Street aren’t raising all boats”.

If Mercor is part of the future of work, it might carry forward the same old struggles between labor and cost-cutting, unless checks and balances emerge.

Market Impact and Mercor’s Long-Term Role

Mercor’s rise has already had tangible impacts on the market for talent and for AI development:

It has unlocked a new labor pool for AI. Many professionals who would never have considered doing gig work, such as finance MBAs, PhD researchers, or seasoned attorneys, are now dabbling in contract work via Mercor. Some are moonlighting while still at their day jobs, earning extra on the side. Others are between jobs or retired, and Mercor gives them a way to monetize their hard-earned expertise for a few hours a week. The concept of a “white-collar gig worker” is being normalized. This trend could expand beyond AI: if Mercor (or copycats) move into other verticals, we might see gig marketplaces for all sorts of specialized knowledge tasks.

Mercor and its peers are also accelerating AI research and development. By making industry-specific data more accessible, they’re potentially shortening the time it takes to train AI models and bring AI applications to market. OpenAI, Anthropic, and others can iterate faster on models for finance or law because they can quickly obtain realistic data from Mercor’s contractors, rather than negotiating data sharing with entities like Goldman Sachs or a large law firm.

This shifts some power away from traditional incumbents (whose “data moats” become less impregnable) and into the hands of AI startups and labs. It’s a form of disruption: knowledge that was siloed in corporate walls is being productized through Mercor.

Incumbent companies may resent this (their trade secrets and proprietary processes are essentially being used to automate parts of their business). Still, as Foody argues, those who embrace this shift will be on the right side of history.

In fact, Mercor envisions working more directly with industries such as law, finance, and medicine, partnering with companies to help them leverage their own data for AI, rather than relying solely on external AI labs. If that happens, Mercor’s role could evolve from “outsider” data liberator to an accepted consultant that helps established firms modernize. That would certainly cement its place in the future of work as a mediator between legacy industries and AI technology.

Competitive dynamics: Mercor’s presence is forcing other platforms and companies to adapt. We’ve mentioned how Scale AI and Surge AI adjusted their strategies. Even big tech firms might consider building their own expert networks rather than relying on third parties. For example, OpenAI could hypothetically internalize some of these processes (though doing it in-house at scale would reduce flexibility).

Traditional freelance platforms like Upwork have also taken notice of the AI gold rush – Upwork’s CEO in 2023 highlighted a surge in AI-related job postings and even launched an “AI Services” hub. However, those were more about hiring AI experts for projects, rather than hiring people to train AI. If Mercor’s model proves sustainable, larger freelance platforms might try to replicate aspects of it (perhaps by creating specialized vetted talent pools or integrating with companies like Mercor).

Conversely, we might see acquisitions: a bigger company could acquire Mercor or a competitor to instantly gain the capability. Given Mercor’s valuation, any acquirer would need deep pockets – but it’s not unimaginable if the likes of Microsoft, Google, or Meta decided that owning part of the AI gig workforce is a strategic move.

From a critical lens, one must ask if this is truly the future of work or a transitional phase. Mercor’s long-term vision is that AI will become incredibly advanced – “better than the best consulting firm, investment bank, and law firm” in the not-so-distant future.

If that’s the case, what happens to all the human experts who helped train those AIs?

Mercor argues that this transformation will be “broadly positive” and create abundance, not scarcity. It’s the classic techno-optimist view that new tech will free humans for higher pursuits. Yet, skeptics worry that these freelancers are effectively training their replacements. The very title of that Bloomberg piece, “Training AI to Take Over,” has a double meaning. Indeed, some of Mercor’s contractors likely realize that if ChatGPT becomes “better than the best lawyer,” it could reduce the need for human lawyers. Mercor itself could face a dwindling market in certain tasks once AI models get sufficiently good.

On the flip side, AI is far from autonomous; it constantly needs updating, tuning, and correcting as it enters new domains. We might simply see the nature of tasks shift: today it’s writing sample memos, tomorrow it might be auditing AI decisions or curating AI outputs – roles that still require human judgment. Mercor, with its pool of adaptable experts, can pivot to meet whatever new needs arise, keeping AI systems aligned and helpful.

Scalability and sustainability: Mercor’s ability to scale will depend on managing quality while adding more people and projects. A “shadow market” is already emerging, where Mercor (and similar platforms) accounts are bought and sold illicitly by those trying to get in on the action. This indicates huge interest – people are even risking scams to obtain a “verified” contractor account for AI training gigs. It also indicates Mercor’s vetting is strict enough that some resort to cheating. The company will need to continually invest in fraud prevention, identity verification, and platform security to ensure that the experts are who they claim to be. (Internal documents from Scale AI showed how rampant duplicate accounts and VPN misuse became when gig work demand spiked, and Mercor likely faces similar issues.) If quality control slips, Mercor could quickly lose the trust of its clients.

Another factor is competition from AI itself. There is a paradox in what Mercor does: it provides humans to improve AI, but many AI researchers are concurrently trying to reduce reliance on human labelers by developing self-training methods. For example, techniques such as reinforcement learning from AI feedback (where AIs evaluate each other) or synthetic data generation could reduce the need for human contractors over time. Mercor appears to be aware of this, but believes that the complexity of real-world tasks will keep humans in the loop for the foreseeable future.

In fact, Mercor and others are also exploring “training environments” and other advanced methods to enhance AI agents, which could lead to the development of new service lines. Nonetheless, the volume of simple labeling work may eventually decline as AI gets better at automating its own training data. Mercor’s focus on high-end, complex data is a hedge against that; they’re in the hardest-to-automate segment of data work.

Is Mercor “the future” of the gig economy?

It certainly represents one compelling vision of it: a future where platforms harness technology to intelligently connect skilled humans with highly specialized, well-paid gigs. In this vision, transparency is higher (clear pay rates, no unnecessary fees), location is irrelevant (work from anywhere, provided you are qualified), and the work itself pushes the frontier of innovation (collaborating with AI).

Mercor has demonstrated that this model can operate at scale and generate substantial economic activity, a notable achievement.

Mercor is indeed shaping the “gig economy of expertise.”

But as its workers accelerate the world’s most valuable technology, the platform is proving that even at $200 an hour, the future of work remains a high-stakes, precarious business built on training your own replacement.