The TAM Scam: Why the next economy won’t need your pitch deck’s market-size slide

Private equity has a problem.

More than 13,500 U.S. companies are currently sitting in private-equity portfolios waiting for an exit. Nearly 4,000 have been held for six years or longer, and analysts estimate the backlog could take roughly nine years to clear at the current pace.

The usual explanation is straightforward: interest rates rose, buyers became disciplined, and companies purchased at exuberant 2021 valuations are exceedingly difficult to sell. But there is another possibility buried inside those numbers.

It isn’t merely that private equity cannot sell companies. It is that PE cannot sell enough companies at prices that validate the marks upon which portions of its reported returns have been based. That turns an exit problem into a measurement problem.

That measurement problem is the other side of the machine-commerce story: as technology makes economic activity more observable, the proxies we built for an earlier era become harder to defend.

When the proxy becomes the product

Internal Rate of Return (IRR) has long been one of private equity’s defining performance metrics. But IRR is not synonymous with operating value creation. Leverage can amplify it, capital call timing can skew it, and early distributions can artificially boost it.

The Institutional Limited Partners Association (ILPA) documented how financing mechanics alter these numbers: in one simplified scenario, delaying investor capital through subscription lines increased reported IRR from 6.62% to 7.98% — while the total-value multiple actually dropped from 1.45x to 1.35x because financing costs consumed real value. The return metric improved while investors made less money.

Venture capital suffers from its own structural distortion: Total Addressable Market (TAM).

Every founder knows the slide: Healthcare is a $4 trillion industry; AI will be worth trillions by 2035; capture just one percent, and we’ve created a billion-dollar company.

This is the TAM Scam. Not because market sizing has zero utility, but because the scam begins when we quietly replace evidence of actual demand with evidence that a large category exists. Declaring that a company addresses a $70 billion market tells us nothing about how many buyers need the product, what they will pay, or the cost of acquiring them.

TAM is possibility dressed up as measurement.

We built an economy of proxies

For most of modern business history, directly observing economic activity was slow, expensive, or impossible. We developed proxies to compress uncertainty into numbers that executives, lenders, and investors could use.

These abstractions were useful measurement technologies. But treating the abstraction as reality becomes unreasonable once new tools make the underlying economic event directly measurable.

We automated uncertainty before we automated truth

The great irony of the AI boom is that we have become extraordinarily good at generating probabilistic answers while much of the economy still runs on approximations.

AI can estimate demand, predict churn, score creditworthiness, and infer a diagnosis. But probability is not provenance. A language model can predict what probably happened, but it cannot, by itself, establish an authoritative chain of custody showing what actually happened.

Intelligence can be probabilistic; economic custody cannot be.

This requires a different class of infrastructure: cryptographic identity, signed events, verifiable credentials, governed data provenance, and machine-readable rights. Blockchain is not a magical truth machine — bad information can still be entered into any database. But a properly governed deterministic ledger can preserve something generative AI cannot establish on its own: a known history of state.

As AI agents begin executing transactions autonomously, probabilistic machines should not also be responsible for deciding what the underlying truth was.

From proxy economics to observed economics

Once machine commerce allows software to purchase resources directly — three database records, one inference, 400 milliseconds of compute — economic activity becomes granular enough to transact and granular enough to measure.

Imagine an investor presentation that no longer cites a $70 billion TAM slide, but instead demonstrates:

“This asset was requested 41,217 times, fulfilled 40,988 times, produced $184,611 of instant settlement, generated a measurable downstream outcome, and distributed proceeds according to programmable rules.” 

That is not forecasting a market; that is observing one. When every economic event can carry its own identity, price, permissions, fulfillment, and settlement history, the transaction stops being merely evidence of the market — the transaction becomes the market.

This is also where industrial tokenization separates from early crypto speculation. The least interesting version of tokenization is putting a speculative wrapper around an asset. The consequential version makes economic objects machine-readable, permissionable, composable, and measurable.

A megawatt of energy, a unit of shipping capacity, a verified clinical dataset, or 400 milliseconds of compute derives its economic value from utility and fulfillment — not speculative market sentiment. Rights travel with the asset itself: who accessed it, at what price, whether delivery occurred, what outcome followed, and where the proceeds went.

The result is not merely programmable money; it is a programmable audit trail of economic behavior.

Reality is becoming queryable

The private-equity exit backlog looks like a finance story, but it is an early warning about the limits of proxy economics.

When inexpensive leverage and expanding multiples disappear, what replaces them is the deterministic answer: What did the business actually produce?

Not what multiple was assigned. Not what TAM was cited. Not what mark sits inside a portfolio. What was purchased? What was delivered? What cash actually moved?

We once had to estimate markets because we could not instrument them. Increasingly, we can. And as the economy itself becomes queryable, the defining question in business stops being “How big could this market be?” and becomes:

“Show me what actually happened.”

Explore MOMENT’s Digital Commerce Readiness Standards (DCRS).

Selected Sources

• The Guardian. Private equity faces existential crisis in US as unsold companies pile up
• McKinsey & Company. Global Private Equity Report 2026
• Institutional Limited Partners Association (ILPA). Enhancing Transparency Around Subscription Lines of Credit
• Institutional Limited Partners Association (ILPA). Subscription Lines of Credit and Alignment of Interests
• x402. Official Protocol Documentation
• ERC-8004. Trustless Agents — Open Standards for the Agentic Web

The first three are especially important. The Guardian supports the 13,500-company PE backlog framing, McKinsey directly supports the 2% older-vintage versus roughly 15% newer-vintage IRR comparison and notes that much of the newer performance remains unrealized, and ILPA provides the exact example where use of a subscription line raises IRR from 6.62% to 7.98% while TVPI falls from 1.45x to 1.35x.