Mark Yusko Puts Bitcoin's Fair Value at $105,000. The Market Says $75,701

There's a $29,299 gap between what bitcoin's network math says it's worth and what the market is actually paying for it. Mark Yusko thinks that gap is the trade.

The Morgan Creek Capital CEO went on air Wednesday, September 16, and put a hard number on it. Using Metcalfe's Law, the observation that a network's value scales with the square of its user count, he pegged bitcoin's fair value at $105,000. Spot at the time was $75,701. That's about 40% below the $126,080 high bitcoin touched back in October 2025.

"So the fair value of bitcoin today, based on Metcalf's law, it's about $105,000, but it's $75,000," Yusko said. "Okay, so it's on sale. You should accumulate things that are on sale."

He gives credit to Tim Peterson, who runs the model tracking it. The logic is simple enough. If a network is worth what its users make it worth, and user growth hasn't broken, then a 28% discount between model and market is a gift, not a warning sign.

Simple logic, though, runs into a complicated tape.

Why the discount exists

Bitcoin started climbing in August after the U.S. Treasury said it would at least double the size of its liquidity-support buyback operations. That put pressure on the dollar and handed a tailwind to assets that don't pay a yield. The debasement trade, the idea that you hold something hard because the currency you'd otherwise sit in is losing value, came back into fashion fast.

But that trade was already hot through 2025, and it drove bitcoin to $126,080. Then October arrived and the money went somewhere else. Specifically, it went to artificial intelligence equities. Traders chasing chip makers decided revenue guidance from a semiconductor company was a better story than digital scarcity. Bitcoin lost momentum for reasons that had almost nothing to do with bitcoin.

So what closes a gap like that? Time and patience, mostly. Metcalfe's Law is a solid sanity check on network value. It's a lousy timing tool, and it always has been. It tells you what a network might be worth if adoption keeps compounding. It says nothing about when the market decides to agree.

Yusko's broader argument is about what actually survives a few decades. Over a 30-year stretch, he points out, 85% of companies disappear. That's a brutal base rate if you're picking individual stocks for a retirement account you won't touch for years. Gold has held value for 5,000 years. Now, in his framing, there's gold and there's bitcoin.

Is that a fair comparison? Partly. Bitcoin's 17-year track record is a rounding error next to gold's, and the drawdowns are twice as violent. But the case he's making isn't about volatility. It's about persistence, and about the fact that equities carry single-company failure risk that a decentralized network structurally doesn't.

The part the valuation debate misses

Here's where I'd push back on the whole exercise. Arguing about whether bitcoin is worth $75,000 or $105,000 is an argument about the reserve asset. It's the least interesting question in crypto right now.

Tokenization isn't a narrative. It's a rails upgrade. And the rails are where the actual revenue is getting built. Tokenized treasuries, private credit, commodities, real estate, they're all being turned into yield-bearing instruments that settle on-chain, some of them with permissioned access and real collateral sitting behind them. That's the stablecoin moment for treasuries, and it's already underway.

If Yusko is right that bitcoin is the monetary hedge, the corollary is that everything built on top of its rails becomes the productive layer. That part of the market doesn't depend on the debasement trade staying hot. It depends on settlement getting cheaper and collateral getting more programmable. Those are engineering problems, not sentiment problems, and engineering problems get solved on a schedule.

Who wins in that world? Custodians with bank charters. Asset managers who figure out distribution before their competitors do. And the infrastructure players tokenizing things nobody thought could be tokenized two years ago. The real world is coming on-chain, one asset class at a time, and the firms handling that transition are collecting fees whether bitcoin trades at $75,000 or $150,000.

Who loses? Anyone still selling a token with no underlying cash flow and calling it a store of value. The market's gotten picky, and that's healthy. DePIN makes more sense when you ignore the name, because the networks doing actual physical work, compute, bandwidth, energy, have something you can underwrite. The ones with a whitepaper and a Discord don't.

The takeaway

Yusko's call is direct. A $29,299 discount to your own model is either a reason to buy or it's nothing, and he's clearly in the first camp. The catch is that this trade only works if you can sit through the stretches where the market ignores your math. 2026 has been one long stretch of exactly that.

The better version of the trade isn't just buying the discount. It's paying attention to what's being built on the network while everyone else argues about its price. Bitcoin might be on sale. The rails, meanwhile, are getting more valuable every quarter, and those don't need Metcalfe's Law to make their case.