Mercor Stock and Shares: The $10 Billion AI Training Platform You Have Not Heard Of Yet

The AI infrastructure conversation is dominated by chips, data centres, and model releases. Nvidia's valuation, hyperscaler capex, and the race to train the next frontier model capture most of the attention and most of the capital allocation discussion.

Underneath that layer sits a less visible but increasingly critical problem. Frontier AI models do not improve through compute alone. They require high-quality human feedback from domain specialists -- engineers, researchers, lawyers, doctors, financial analysts -- who can evaluate model outputs with the kind of nuanced judgement that generic annotation cannot provide. As models become more capable and are deployed into more complex real-world tasks, the demand for that expertise grows rather than shrinks.

Mercor is the platform built to supply it. The company connects AI labs and enterprise customers with a managed network of domain experts who generate the training signals, validation outputs, and reinforcement learning feedback that modern AI development depends on. It went from founding to a reported $10 billion valuation in under three years. Mercor stock is not yet publicly available, and Mercor shares are not listed on any exchange.

This is the investment case.

What Mercor Is and How It Works

Mercor was founded in San Francisco in 2023, initially as an AI-powered recruiting platform. The original product used AI to automate resume screening, interviewing, and candidate matching for employers. That product was useful. It was not the opportunity.

The pivot came from recognising a structural problem in AI development. The biggest AI labs -- OpenAI, Anthropic, Google DeepMind, and others -- were building models that required increasingly sophisticated human input to improve. Basic annotation work, the kind that could be crowdsourced at low cost, was insufficient for training models on complex reasoning, legal analysis, scientific research, or financial modelling. They needed experts. Finding, vetting, and managing those experts at scale was operationally expensive and fragmented.

Mercor rebuilt its platform around that problem. Today it operates a managed marketplace of more than 30,000 vetted domain experts spanning engineering, science, law, medicine, finance, and journalism. AI labs and enterprise customers access this network to generate expert-level training data, validate model outputs, and build the reinforcement learning feedback loops that improve model performance on specialist tasks.

The platform combines automated matching -- using Mercor's own AI to vet and place experts efficiently -- with managed workflow infrastructure that integrates expert human feedback directly into model development pipelines. Mercor does not just connect buyers and sellers. It manages the quality, throughput, and integration of the work.

Mercor Stock Valuation: From Zero to $10 Billion in Three Years

The funding trajectory tells the story of a company that found genuine product-market fit quickly and scaled it.

Mercor raised its Series A in 2024 at a reported valuation of $250 million, led by Benchmark. By 2025, it raised a Series B of $100 million at a reported $2 billion valuation, led by Felicis. In October 2025, Mercor closed a $350 million Series C at a reported $10 billion valuation.

That is a 40x increase in valuation in roughly 18 months.

To contextualise the pace: a $10 billion valuation for a company founded in 2023 puts Mercor among the fastest private market value creation stories in recent AI infrastructure history. The reported average contractor earnings of approximately $85 per hour indicate a marketplace with genuine liquidity and engagement, not a nominal network with low utilisation. The addition of Sundeep Jain as President in 2025 -- a former Uber Chief Product Officer -- signals a move toward the operational scale required for a potential public listing.

Mercor shares are not publicly traded. The $10 billion figure reflects the October 2025 Series C pricing. Private valuations are point-in-time estimates and do not update continuously, but the trajectory from $250 million to $10 billion in under two years reflects consistent institutional confidence in the business. Investors asking about Mercor stock price have no live reference point -- there is no ticker, no exchange listing, and no continuous price discovery until a public listing occurs.

Why the AI Training Market Is Structural, Not Cyclical

The most important thing to understand about Mercor's market position is why demand for its services increases as AI improves rather than declining.

Early AI models were trained primarily on large datasets of publicly available text and structured data. Human annotation at this stage was relatively generic -- labelling images, classifying text, rating outputs on basic quality dimensions. This work could be sourced cheaply and at scale from general labour pools.

Frontier models operating at the capability levels that OpenAI, Anthropic, and Google are targeting require something different. They are being trained to reason through complex legal documents, generate accurate medical analysis, produce research-grade scientific outputs, and make financial recommendations that can be evaluated against real-world outcomes. The humans needed to generate training signal for these tasks are not generic annotators. They are domain specialists with graduate degrees, professional credentials, and years of applied expertise.

As models become more capable and are deployed into more specialised enterprise workflows, the bar for the human feedback that improves them rises. That means Mercor's market does not shrink as AI advances. It expands.

This is the dynamic that makes Mercor less of a staffing marketplace and more of a critical layer in the AI value chain. The company is not competing with AI. It is the infrastructure that AI development currently cannot function without at the frontier.

Business Model and Unit Economics

Mercor operates a managed marketplace model. Revenue comes from finder's fees, matching fees, and usage-based charges that AI labs and enterprise customers pay for expert contributions. Contractors are engaged on hourly or project-based terms, with Mercor capturing a margin on facilitated work.

The model has natural operating leverage. As the expert network grows and the matching infrastructure becomes more sophisticated, the marginal cost of placing an additional expert into a workflow decreases while the quality and speed of matching improves. A larger network also becomes more defensible: AI labs that have integrated Mercor's workflow infrastructure into their training pipelines face real switching costs.

Potential expansion paths include subscription or enterprise SaaS offerings tied to matching, orchestration, and reinforcement learning infrastructure. As Mercor's involvement moves deeper into the AI development pipeline -- from providing experts to managing entire feedback loops -- the revenue per customer relationship expands.

The reported $85 per hour average contractor earnings is a useful proxy for the quality tier Mercor is operating in. This is not a low-cost annotation platform. It is a specialist marketplace, and the pricing reflects that positioning.

Competitive Position and Moat

Mercor's defensibility rests on three compounding advantages.

The first is network scale. A network of 30,000 vetted domain experts across multiple high-value disciplines took time, capital, and operational investment to build. Replicating it requires the same. More importantly, the vetting process -- Mercor uses its own AI to interview and evaluate experts -- creates a proprietary quality signal that improves as the dataset of expert performance grows. That feedback loop is not easily replicated by a new entrant starting from scratch.

The second is workflow integration. AI labs that have built Mercor's expert feedback directly into their training pipelines have created operational dependency. Switching to a different provider requires rebuilding those integrations, retraining the workflow, and accepting a quality transition period. That is a real cost that increases the stickiness of existing customer relationships.

The third is the data and IP position. Mercor owns proprietary algorithms for expert vetting, interviewing, and talent matching, as well as reinforcement learning and human-feedback orchestration tools. The accumulated interaction data from expert contributions to AI training creates a dataset that improves the matching and orchestration infrastructure over time. The more the platform is used, the better it gets. The better it gets, the more it is used.

What Investors Are Underwriting

Buying Mercor shares at this stage -- through a structured pre-IPO vehicle -- means taking a view on several distinct questions simultaneously.

Is expert human feedback a permanent structural requirement for frontier AI, or a transitional bottleneck? If models eventually become self-improving without human feedback, the market Mercor serves contracts. The current evidence from leading AI labs suggests that human-in-the-loop processes remain essential at the frontier, and that the bar for the quality of that feedback rises as model capabilities increase. But this is the core uncertainty that any honest analysis of Mercor must acknowledge.

Can Mercor maintain its market position as the category grows? Larger talent platforms, staffing companies, and AI-specific competitors are aware of the market Mercor is building. The network and workflow moat is real, but it is not impregnable. Execution matters.

Does the growth rate justify a $10 billion valuation? A 40x valuation increase in 18 months reflects significant institutional confidence. Whether that confidence proves well-founded depends on the revenue trajectory -- which is not publicly disclosed -- and whether the business can sustain the growth rate that the Series C pricing implies.

The Risks That Belong in This Analysis

Revenue transparency. Mercor is a private company. Its revenue figures are not publicly disclosed. The valuation is based on what institutional investors with access to the financials were willing to pay. Outside investors working from public information are making a judgement with less data than those who set the Series C price.

Model self-sufficiency risk. The long-term scenario where AI models reduce their dependency on high-quality human feedback is the primary existential risk for Mercor's business model. The company's response -- building workflow infrastructure and data assets that extend beyond pure annotation -- is the right strategic direction, but execution risk remains.

No confirmed IPO timeline. Mercor has not announced a public listing. Mercor stock and shares are not available on public markets and there is no confirmed date for when that will change. Capital committed to a pre-IPO position should be treated as long-term and illiquid until a defined exit event occurs.

Concentration risk. If a significant portion of Mercor's revenue comes from a small number of AI lab customers, changes in those relationships would have an outsized effect on the business. Customer concentration details are not publicly available.

How to Buy Mercor Shares Before an IPO

Mercor stock is not available through standard brokerage accounts, and Mercor shares cannot be purchased on any public exchange. For most investors, structured pre-IPO vehicles are the only available path to economic exposure before a listing occurs.

WLTH currently offers tokenised economic rights to Mercor exposure through its pre-IPO access platform. Investors can explore the Mercor opportunity directly on WLTH, or browse other available private market positions on the WLTH marketplace. This is economic exposure to private market performance, not direct equity or shareholder rights in Mercor or any underlying company.

The AI training infrastructure layer is still being built. The companies that own the critical nodes in that infrastructure are still private. Mercor is one of them.

WLTH provides tokenised economic rights to private market exposure. This does not constitute financial advice. Capital is at risk.