US Venture Capital Report 2021 — The Speed Year

2021 was the most active year in the history of US venture capital by nearly every measure. The interesting question is not how much was invested but how the process of investing changed, because those changes explain the vintage's subsequent performance better than any observation about totals.

The central change was a shift in what investors were competing over. In a normal market, capital is scarce relative to opportunities, and the investor's problem is selection — choosing correctly among more candidates than they can fund. Competitive advantage comes from judgement.

In 2021 capital was abundant relative to opportunities. The good companies had multiple term sheets. The investor's problem was no longer which company to back but whether they would be allowed to back it. Competitive advantage shifted from judgement to access.

That shift has a mechanical consequence. The primary way to win access is to be faster and more certain than the alternative. A founder choosing between offers weighs price, but also weighs speed, certainty of close, and how much process they must endure. An investor who can commit in days beats one who needs weeks, even at a similar price.

So diligence compressed. Not because investors stopped believing in diligence, but because the time diligence takes was the exact thing they were competing on. Each individual decision to move faster was rational given what competitors were doing. The aggregate result was a market-wide reduction in the scrutiny applied to capital deployment, at the highest prices in the industry's history.

That combination — maximum price, minimum scrutiny — is what defines the 2021 vintage.

The access problem and why it produces speed

This dynamic is worth setting out carefully because it recurs at every market peak and is consistently misdescribed as investor irrationality.

In a selection market, an investor sees more opportunities than they have capital for. They can be slow, thorough and demanding, because saying no costs nothing — another opportunity follows. Diligence is free.

In an access market, the good opportunities have alternatives. The investor is being evaluated as much as evaluating. Now saying no, or taking too long to say yes, has a real cost: the opportunity goes to a competitor, and the return goes with it.

The founder's decision criteria in an access market are the key to the whole dynamic. Facing several offers at similar prices, they weigh:

• Speed to close — how long until the money is in the bank and they can go back to building.
• Certainty — will this investor complete, or will they find a reason to withdraw at the last moment.
• Process burden — how many meetings, reference calls, data requests.
• Signal value — what backing from this investor communicates to future investors, customers and employees.

Price matters, but at the top of the market, several investors will match on price. Speed and certainty are the differentiators that remain, and both are directly traded against diligence.

Diligence is time. In a market where time is what you are competing on, diligence is what you spend to compete. Nobody decided to stop checking. They decided to be faster, and checking is what got faster.

The individual rationality is what makes this hard to escape. An investor who maintained a full six-week process in 2021 would systematically lose the competitive rounds and be left with the ones nobody else wanted — a form of adverse selection that is arguably worse than paying up for a good company. There was no unilaterally good option.

Several structural adaptations followed, and they are worth noting because they persisted:

• Pre-emptive rounds, in which an investor offers terms before the company has begun a process, removing competition entirely by removing the auction.
• Standing relationships with companies well before they raise, so diligence is continuous rather than compressed into a round.
• Smaller checks at higher prices, accepting worse entry economics in exchange for optionality on a follow-on.

Crossover capital and the imported anchor

The 2019 report describes crossover funds — public market managers taking private positions — as one of the sources of late-stage capital that eroded price discipline. In 2021 their role became more specific and more consequential.

Crossover investors evaluate private companies against public comparables. Their entire analytical apparatus is built for public equities: they model revenue, apply a multiple derived from comparable listed companies, and derive a target price.

Applied to private companies, this imports the public multiple into the private round. If comparable listed software companies traded at a high multiple of revenue, that multiple became the reference for private companies with similar growth.

This had two effects that are easy to state and were widely missed at the time:

• It anchored private valuations to public ones during a period when public multiples were themselves elevated by the low discount rate described in the 2020 report. Private prices inherited the public market's condition.
• It created a transmission channel in both directions. Because the private mark was derived from the public multiple, a fall in the public multiple mechanically implied a fall in the private one — with a lag determined only by when the next round or valuation review occurred.

That second point is the whole story of 2022. The private market did not independently decide to reprice. The anchor it had adopted moved, and the private market was attached to it.

There is a further wrinkle worth noting. Crossover funds report marks to public shareholders on a quarterly cycle and face pressure to mark honestly and promptly. Traditional venture funds report to LPs less frequently and hold more discretion over private marks. So the same underlying position could be marked down quickly on one balance sheet and held at cost on another — which is a substantial part of why 2022's repricing appeared to happen at different speeds to different holders of the same asset.

Why the vintage underperformed

The 2021 vintage's difficulties are frequently attributed to poor company selection. That explanation is mostly wrong and it matters that it is, because it points at the wrong lesson.

Entry price is the dominant variable in venture returns, and it is fixed at the moment of investment. A fund's return is determined by exit value divided by entry value. Everything an investor does after the investment — support, follow-on, board work — operates on the numerator. The denominator is set once and cannot be revised.

The 2021 vintage entered at the highest prices in the industry's history. That alone, independent of any judgement about which companies were backed, compressed the achievable return.

Three compounding factors made it worse:

• Follow-on rounds priced against a changed market. A company that raised in 2021 at a high valuation and needed capital in 2023 faced a market that would not support that price. The result was a down round, a structured round, or a sale — all of which impair the 2021 position specifically.
• Structure moved against earlier investors. When later rounds carry preferences and ratchets, they sit ahead of the earlier round in the payout stack. The 2021 investor's position is subordinated by events after their investment.
• The exit environment closed. The path to realising value narrowed at precisely the moment the vintage needed it.

The instructive point is that many 2021-vintage companies were and are genuinely good businesses. They grew, built real products, and acquired real customers. The vintage's problem was not that the companies were bad. It was that the price paid for them assumed a set of conditions that did not persist.

Venture returns are made at entry and realised at exit. The middle is execution, and execution cannot fix a denominator.

This is also the strongest available argument for vintage diversification — deploying consistently across years rather than concentrating in whichever year feels most opportune. The years that feel best to deploy in are, mechanically, the years with the highest prices.

What compressed diligence actually skips

The claim that diligence compressed in 2021 is usually stated generally. It is more useful to specify what gets dropped when a process moves from six weeks to six days, because the dropped items are not random — they are the ones that take calendar time regardless of effort.

What survives compression, because it can be done quickly by adding people:

• Financial review. Reading the accounts, building a model, checking the metrics. This is labour-intensive and parallelisable.
• Market sizing. Desk research, which can be done in days.
• Product assessment. Using the product, reviewing the technology, talking to the team.

What does not survive compression, because it requires elapsed time:

• Reference calls beyond the founder's list. The founder supplies references immediately. Finding people not on that list — former employees, customers who churned, investors who passed — requires reaching people who have no reason to respond quickly. This is where the most informative diligence usually lives and it is the first casualty.
• Customer diligence at depth. Speaking to enough customers to distinguish enthusiasm from a representative sample takes weeks, because customers respond on their own schedule.
• Cohort analysis over time. Watching retention behave over several months is not something more analysts can accelerate.
• Observing the team under conditions. A relationship built over months reveals things a compressed process does not.
• Legal and technical review of anything unusual. Standard documents review quickly; unusual ones do not.

The pattern is that compression preserves the diligence that can be performed on supplied information and eliminates the diligence that requires independent sourcing. That is exactly backwards from an information standpoint, because supplied information is the least likely to contain a surprise.

Adding analysts speeds up the work that uses what the company gave you. It does not speed up finding out what the company did not.

This is why the structural adaptations that emerged — pre-emption and continuous relationships — were rational responses rather than mere aggression. An investor who has known a company for two years has done the elapsed-time diligence in advance. When the round arrives they can move in days without having skipped anything. That is the only way to be fast and thorough simultaneously, and it is why the practice persisted after 2021.

What an allocator could act on

Size to entry price, not to conviction. Entry price is the dominant determinant of venture returns and it is fixed at the decision point. Everything afterwards operates on the numerator. A vintage entering at the highest prices in the industry's history was structurally disadvantaged before any company underperformed.

Deploy at a constant rate. Vintage diversification is the only reliable defence against a coordination problem an investor cannot exit. The years that feel best to deploy in are, mechanically, the years with the highest prices — which means accelerating into a strong market is systematically buying at the worst points.

Prefer clean terms to a better headline price. Structure that protects a later investor subordinates the earlier one. A 2021 position was frequently impaired not by the company's performance but by the terms of rounds that came after it.

Ask what a private mark is anchored to. Where crossover capital set the price against public comparables, the private valuation is a function of the public multiple — which means it moves when the public multiple moves, on a lag set only by when the next round occurs. That transmission channel is identifiable at the time and it made 2022 mechanical rather than surprising.

Note that the same asset can be marked differently by different holders. A crossover fund reporting quarterly to public shareholders and a venture fund reporting to LPs hold different valuation obligations. Divergent marks on the same company are not evidence of disagreement about value — they are evidence of different reporting regimes, and the more frequently-marked holder is usually the more current one.

What 2021 established for US venture

• The selection-to-access shift was demonstrated at scale, along with its mechanical consequence for diligence.
• Individually rational speed produced a collectively poor outcome — a coordination problem rather than a judgement failure.
• Crossover capital anchored private marks to public multiples, creating the transmission channel that made 2022 inevitable once public multiples fell.
• Entry price was reconfirmed as the dominant return variable, and the case for vintage diversification was made more forcefully than any argument could.
• Pre-emption and continuous diligence entered the standard playbook and did not leave when conditions changed.