When I see a $20 billion valuation on a private AI data company with no public revenue figures, my stomach knots. That’s not fear—it’s pattern recognition. I’ve watched too many crypto projects pump their token prices on zero fundamentals only to collapse when the hype ran dry. Now, the same script is playing out in the AI training data space, and this time the numbers are even bigger.
Let’s talk about Mercor. The company is reportedly discussing a $20 billion valuation as AI training demand supercharges its growth. If true, that would make Mercor the most valuable pure-play data labeling firm on the planet—surpassing even Scale AI, which sat at $13.8 billion post-funding in 2024. But here’s the thing: the article raising this number also flags two major concerns—safety and revenue sustainability. And in my nine years of watching market narratives form and break, that combination of awe and worry is the loudest alarm bell.
Context: This Isn’t Just Another Data Labeler
Mercor sits at the bottleneck of the AI boom. Every frontier model—from GPT-5 to Claude 4—needs millions of human feedback signals to align its outputs. RLHF (Reinforcement Learning from Human Feedback) isn’t a nice-to-have; it’s the difference between a chatbot that rambles and one that follows instructions. Mercor provides the workforce and the platform to deliver that data. Think of it as the backstage crew for the AI rockstars.
The industry is hot. Scale AI, Appen, Labelbox—they’re all racing to lock down contracts with OpenAI, Anthropic, Google DeepMind. But Scale AI’s $13.8 billion valuation came with estimated annual revenue of $200–300 million. That’s a price-to-sales multiple in the 50–70x range. For Mercor to justify $20 billion, its revenue would need to be proportionally higher—think $400 million to $1 billion—or grow at an absolutely blistering pace. Can it? Possibly. But the article itself signals doubt by naming “revenue sustainability” as an unresolved worry.
Core: The Order Flow Behind the Hype
Let’s break down the order flow here. VCs and institutional investors are pouring billions into AI infrastructure—chips, datacenters, training runs. But they’ve also realized that without high-quality data, the best GPUs are useless. So the capital naturally cascades down to data providers like Mercor. The valuation discussion is part of that flow—a signal that investors are willing to pay a premium for exposure to the data layer.
But here’s where my battle-tested instincts kick in. I’ve manually tracked vesting schedules and token distribution in DeFi projects. I’ve seen how easy it is to inflate TVL with incentives, and how quickly those users vanish when the subsidies stop. Mercor’s business model is different—it’s real B2B revenue, not token farming. Yet the same risk persists: customer concentration. If Mercor’s top three clients account for 70% of revenue, a single contract loss could crater the company. And we don’t know who those clients are. The article offers no contract names, no revenue breakdown. That’s a red flag in any market.
Based on my experience auditing data-driven startups during the 2020 DeFi Summer, the companies that survive downturns are the ones with diversified revenue streams. Uniswap didn’t rely on one liquidity pair. Compound didn’t depend on one asset. Mercor, by contrast, is sitting in a niche where a handful of AI labs hold all the power. If OpenAI decides to build its own labeling platform—or acquire a competitor—Mercor’s moat shrinks overnight.
Contrarian: The Smart Money Is Hedging
The mainstream narrative is “AI is the future, so data is gold.” The contrarian angle is that the gold rush has already peaked. Look at Appen, once a darling of the data labeling space, now trading at a fraction of its 2021 highs. The problem is commoditization: basic data labeling can be done by cheaper labor overseas or by self-supervised AI. The high-margin portion—expert annotation (lawyers, doctors)—is small and slow to scale. Mercor’s $20 billion valuation assumes it will capture nearly all of that premium segment, but that’s a tall order when Scale AI has the same ambition.
Retail investors—and even some smaller funds—see the $20 billion sticker and think “buy the dip before IPO.” But the smart money is already hedging. The article’s mention of “safety” concerns is code for regulatory risk. The EU AI Act, China’s data governance rules, and even US potential oversight on training data provenance could impose costly compliance burdens. Mercor might need to invest millions in privacy tech and auditing just to keep existing clients happy. That eats into margins.
I remember the Terra collapse. Everyone thought the algorithmic stablecoin was a miracle until the collateral failed. Mercor’s valuation is not a stablecoin, but it rests on a similar assumption: that AI demand will grow exponentially forever. If the next GPT release underwhelms—or if a recession cuts corporate AI budgets—Mercor’s revenue growth could stall. And a stall at a 50x P/S ratio is a 90% drawdown.
Takeaway: What to Watch
Don’t chase the headline. The $20 billion number is a negotiation starting point, not a fair value. Here’s what matters:
- CLIENT DISCLOSURE: If Mercor reveals contracts with multiple top-tier AI labs and a customer concentration below 30%, the valuation becomes more credible. Until then, treat it as hype.
- REVENUE GROWTH: Look for leaks or filings showing year-over-year revenue. Anything below 100% annual growth at this stage is a sell signal.
- SAFETY AUDIT: Mercor needs a SOC 2 Type II certification and public commitment to data provenance. Without it, regulatory risk is a ticking bomb.
I’ve been in this game long enough to know that the most dangerous moment is when everyone agrees on a narrative. Right now, the AI data narrative is too cozy. Trust the hands, not just the charts. We need to see the actual transactions behind the valuation before we call it a buy.
Community first, coins second. Always. And in this case, the coin is trust in a company’s fundamentals. Mercor hasn’t earned that trust yet.
Follow the people, follow the profit. Right now, the people are employees and early investors looking for an exit. The profit is in the private markets, not in public excitement. Stay cautious, stay grounded, and let others be the exit liquidity.