US Venture Capital Report 2019 — The IPO Reckoning

Between roughly 2010 and 2019, US venture capital underwent a structural change that is easy to describe and was hard to see while it happened: companies stopped going public.

Or more precisely, they went public much later. The median age at listing extended substantially. Companies that in an earlier era would have listed at a few hundred million dollars of value instead raised private rounds at several billion and stayed private for years longer.

This had a consequence that was not much discussed at the time. Public markets are the mechanism by which venture valuations are validated. A private round sets a price between a company and a handful of investors. A public listing sets a price among thousands of participants who can all say no. When companies stay private for a decade, an entire cohort of valuations accumulates without ever being tested.

2019 was the year a substantial part of that cohort came to market at once. Several of the largest and most-discussed private companies of the decade listed within months of each other.

The outcomes divided along a line that, in retrospect, is obvious: companies whose unit economics survived disclosure did acceptably; companies whose economics depended on assumptions that disclosure undermined did not. In one high-profile instance, the listing was withdrawn entirely — the disclosure process alone was sufficient to destroy the valuation, without a single share trading.

The reckoning that began was real, and it began to propagate backward into private markets in late 2019. It did not finish. The 2020 policy response reopened funding on more generous terms than before, and the correction was deferred to 2022.

How the late stage lost its discipline

The mechanism is worth setting out because it explains the whole cohort rather than any individual company.

Venture capital had traditionally been supply-constrained on the capital side. There were a limited number of firms, they raised funds of limited size, and the discipline that produced came from scarcity: an investor who overpaid had less money for the next deal.

Through the 2010s, capital supply into late-stage private companies expanded from sources that did not operate under that constraint:

• Crossover funds — public equity managers taking private positions. Their comparison set was public companies, and their fund sizes dwarfed traditional venture funds. A position that would be a fund-defining bet for a venture firm was a small allocation for them.
• Sovereign wealth and large strategic pools, deploying at a scale that made price a secondary consideration relative to access.
• Corporate venture arms, some of which had strategic objectives that made financial return a partial rather than complete decision criterion.
• Dedicated growth funds raised specifically for late-stage deployment, which created an obligation to deploy at that stage regardless of whether attractive opportunities existed there.

That last point is the important one and it generalises. A fund raised for a specific stage must deploy at that stage. If the supply of good companies at that stage does not grow proportionally with the capital raised for it, the surplus capital does not sit idle — it bids up the companies that exist. Price discipline erodes not through anyone's poor judgement but through the structure of the commitment.

Capital raised for a stage must be deployed at that stage. When the money grows faster than the companies, the difference shows up in price.

The arithmetic of an overstated valuation

One of the most useful things 2019 clarified is that a significant portion of headline private valuations were overstated by a mechanism that is arithmetic rather than debatable.

The standard practice. A company raises at a price per preferred share. The headline valuation is calculated by multiplying that price by all shares outstanding — preferred and common alike.

Why that overstates. The preferred shares being sold are not the same security as the common shares held by founders and employees. Preferred typically carries a liquidation preference: in a sale, preferred holders receive their money back (sometimes a multiple of it) before common holders receive anything.

That preference has value. The investor is buying a security with downside protection that common shareholders do not have. A rational investor pays more for the protected security than they would for unprotected common stock.

The consequence. Multiplying the protected price across the unprotected shares assumes all shares are worth the same. They are not. The headline number therefore exceeds the true value of the whole company, and the size of the gap grows with the aggressiveness of the terms.

A worked illustration, using round numbers to show the mechanism rather than to describe any specific company:

A company raises $100m at a $1bn "post-money valuation" — 10% of the company, priced at

$10 per preferred share, with a 1× liquidation preference. If the company later sells for

$200m, the preferred holders take their $100m back first. The remaining $100m is split

across the other 90% of shares. The preferred position returned 1.0×. The common position

returned roughly 11 cents on each dollar of the implied $10 share price.

*Derived: illustrative arithmetic, round numbers chosen to expose the mechanism. Not a

description of any actual transaction.*

Additional structure widens the gap further. Participating preferred takes the preference and shares in the remainder. Ratchets issue additional shares to the investor if a later round or listing prices below a threshold, protecting them at the direct expense of common holders. Each of these makes the headline valuation a worse estimate of enterprise value.

Listing collapses all of this. In a public offering, preferred typically converts to common. Every share becomes the same security, the protections disappear, and the market prices the whole company as one thing. The overstatement does not decline gradually — it is removed in a single step.

This is why several 2019 listings priced below their last private round without anything having changed about the business. The business did not get worse. The measurement got honest.

What disclosure revealed

Public listing in the US requires a registration statement containing audited financials, a management discussion of results, related-party transactions, governance arrangements, and risk factors written under legal liability for their completeness.

For most companies this is a procedural exercise. For the 2019 cohort, in several cases, it was the first time certain information had been assembled for outside scrutiny — and the assembly itself changed how the businesses were understood.

The categories that mattered most:

• Unit economics at scale. A business can look attractive at aggregate revenue growth and unattractive at the level of the individual customer or transaction. Disclosure requires enough detail to compute the second. Several businesses growing rapidly turned out to be growing rapidly because each unit of activity was subsidised.
• The composition of growth. Growth funded by discounting or incentives is different from growth from underlying demand. Disclosure of marketing spend against cohort retention makes the distinction visible in a way private reporting often does not.
• Governance and related-party arrangements. Dual-class structures, transactions between the company and its founders, and unusual control provisions were disclosed in detail. Public investors buying a minority stake price governance differently from private investors with board seats and information rights.
• The gap between adjusted and reported figures. Private reporting frequently emphasises adjusted metrics that exclude categories of cost. Public filings require reconciliation to standard accounting, and the size of the reconciliation was itself informative.

A private valuation can rest on what has not been asked. A public listing asks everything, in writing, under liability. That is not a stricter version of the same test — it is a different test.

The behavioural shift

The consequence for the venture market was rapid and specific: "growth at any cost" stopped being fundable, essentially within a year.

For most of the 2010s the dominant advice to founders in large addressable markets was to prioritise growth over profitability, on the reasoning that market position was the scarce asset and could be monetised later. That reasoning was not irrational — in several categories it produced enormously valuable businesses.

But it rested on a condition: that capital would remain available to fund losses until the position was established. 2019 demonstrated that the eventual buyer of the company — the public market — would apply a different standard, and that the standard would be applied at the least convenient moment.

The observable changes through late 2019 and into 2020:

• "Path to profitability" entered standard diligence as a required section rather than a later-stage concern.
• Burn multiple and capital efficiency became metrics founders were expected to know and defend.
• Late-stage rounds slowed as investors began underwriting to a public comparable rather than to the next private round.
• The advice given to founders changed, from optimising growth rate toward optimising the growth-to-burn relationship.

This shift was genuine and would have propagated fully through the market over the following two years. The 2020 policy response reversed it almost entirely, and it had to happen again — with far more force and far more damage — in 2022.

What "growth at any cost" was actually claiming

The strategy that 2019 discredited is usually caricatured, and the caricature makes it harder to see why it was adopted by serious people or why its failure was specific rather than total.

The claim was not that profitability is unimportant. It was that in certain markets, market position is the scarce asset, and a company that establishes position can monetise it later at a margin unavailable to a competitor who prioritised near-term profit.

The claim rests on identifiable conditions, and where those conditions hold it is correct:

• Winner-take-most dynamics. The market must have network effects, switching costs, or scale economies strong enough that the leading position is durable rather than merely current.
• A monetisation route that opens later. The position must be convertible into profit — through pricing power, through adjacent products, or through a cost structure that improves with scale.
• Capital availability until it does. The strategy requires funding losses through the period, which means it depends on capital conditions the company does not control.
• Competitors who cannot outspend you. If a better-capitalised competitor pursues the same strategy, position is not established — it is bid for, and both parties lose.

Where these held, the strategy produced enormously valuable businesses, and several of the largest technology companies were built this way. That is why it was adopted.

Where they did not hold, it failed for specific reasons:

• Markets with weak network effects produced positions that were not durable. Customers acquired by subsidy left when a competitor subsidised more.
• Businesses where the unit economics never improve with scale could not monetise the position, because the losses were structural rather than investment.
• Capital availability proved conditional, as 2019 and then 2022 demonstrated, which broke the third condition for everyone simultaneously.

The 2019 lesson is therefore narrower and more useful than "growth at any cost was wrong." It is that the strategy has preconditions, that those preconditions are assessable in advance, and that the strategy was applied in many markets where they did not hold. A strategy with conditions applied without checking the conditions is not a strategy — it is a habit.

What an allocator could act on

Compute the implied common value from the capitalisation table. Every S-1 contains it. The gap between the headline private valuation and the implied common-share value is arithmetic, not judgement, and it was frequently large. The same computation is available from the term sheets of any private holding.

Read the risk factors as diligence, not boilerplate. A registration statement's risk factors are written under legal liability for their completeness. For companies moving from private to public, they frequently contain the clearest available statement of what the business's vulnerabilities are — assembled by lawyers whose interest is in omitting nothing.

Check the reconciliation between adjusted and reported figures. Private reporting emphasises adjusted metrics that exclude categories of cost. Public filings require reconciliation to standard accounting. The size of that reconciliation is itself informative and is not available before a company files.

Ask which of the four preconditions the growth strategy depends on. Winner-take-most dynamics, a later monetisation route, capital availability, and competitors who cannot outspend. Naming which are assumed makes the thesis falsifiable and identifies which one, if it fails, ends the strategy.

Watch for stage-committed capital as a leading indicator of price discipline. A fund raised for a specific stage must deploy at that stage. When capital raised for a stage grows faster than the supply of companies at that stage, the surplus bids up the companies that exist. This is observable from fundraising data and it predicts price erosion before the prices erode.

What 2019 established for US venture

• Public listing is a disclosure test, not merely a liquidity event, and it is the only forcing mechanism the private market has.
• Headline private valuations are mechanically overstated by preference structure, and the overstatement is computable from filings rather than estimated.
• Stage-committed capital erodes price discipline when it grows faster than the company supply at that stage.
• The check on late-stage pricing had been absent for a decade, which is why the audit produced such a wide spread of outcomes.
• The correction was interrupted, not completed, making 2019 the direct ancestor of 2022.