Over the past 25 years, the time required for startups to reach a $1 billion valuation (“Unicorn”) has compressed dramatically. Companies that once took two decades to become unicorns now do so within months, or in some cases are created as one.
this represents a key shift in venture capital. It changes how companies are financed, how capital concentrates, and how returns are (or aren’t) generated.
Unicorn status is no longer a milestone. It is an entry price.
And if it becomes the starting line rather than the finish line, we need to rethink how venture capital works.
Compression of Time
It now takes effectively zero years to reach unicorn status.
For the latest cohort of unicorns, the median time to reach a $1 billion valuation is zero years. Companies are being valued at $1 billion in their very first institutional round.
This does not mean the average startup is born a unicorn. It means that, within the newest cohort of companies that become unicorns, the median is now priced like one from day one. That distinction matters, but the shift is still clear.
To understand how we got here, we analysed more than 6,000 VC-backed companies founded over the past 25 years, tracking how long it took each to reach unicorn status or go public via IPO. The pattern is remarkably consistent.
Companies founded in 2000 took a median of 21 years to reach $1 billion. By 2010, that number had fallen to 11 years. By 2020, the fastest companies reached unicorn status at inception for the first time ever, while the cohort median had fallen to 4.5 years. By 2024, the median reached unicorn status within roughly one year.
As recent cohorts mature, the final median may adjust upward, but the compression trend remains clear.
At first glance, the pattern appears anchored to 2021, the year that saw record capital deployment and exits. One could argue that the compression is simply a by-product of that liquidity spike, rather than a lasting shift.
Yet the data does not stop in 2021. Time-to-unicorn continued to decline even as markets slowed, suggesting the trend is structural rather than cyclical.
The decline has been surprisingly consistent. Roughly one year has been shaved off for every year of founding, almost mechanically, for a quarter of a century.
But what we don’t realize is that while going from 21 to 20 years is barely noticeable, going from 2 years to 1 cuts the timeline in half. Today, we have effectively reached the floor: we cannot get to unicorn status faster than instantly.
Faster Fundamentals, Higher Valuations
AI companies are generating revenues at a speed and scale never seen before.
The compression of time-to-unicorn we mentioned earlier is not only visible in valuations, but increasingly in fundamentals. Revenue acceleration that once took a decade now unfolds within a few years – or even months.
Take Anthropic as an example. The company reached unicorn status within a year of its founding in 2022, generating effectively no revenue at the time. Valuation came first, built on expectations of rapid revenue expansion and a massive addressable market.
Four years later, Anthropic is reportedly operating at a $14 billion revenue run rate, implying roughly 10x year-over-year growth since inception. Few companies in history have scaled revenue at that pace. It is therefore not surprising that venture managers are now chasing “centicorns”.
This dynamic has reshaped pricing across private markets. Indexed to 2015, median pre-money valuations have risen at every stage. Series D+ valuations are up more than 5x. Series C more than 4x. Earlier stages between 3 and 4x. Even after the 2021 correction, valuations across most stages rebounded quickly toward new highs by 2025.
Importantly, this trend did not begin with AI.
Software companies in the 2010s were already showing exponentially faster growth. Software could be deployed instantly worldwide, with minimal cost, high gross margins, and recurring subscription revenue. Growth curves steepened materially and multiples followed.
Valuations also reflected larger outcome expectations, as software began displacing traditional industries (such as hotels with Airbnb or taxis with Uber), offering significant upside to category leaders.
As the graph above shows, median valuations increased by ~2-6x between 2015 and their 2021 peak. After the reset in 2022–2023, valuations still rebounded quickly. By 2025, most stages had surpassed prior highs.
The driver, however, has now evolved.
If software-as-a-service (“SaaS”) accelerated revenue through efficiencies and network effects, AI accelerates it through direct task completion and labour displacement. In some cases, AI products claim to remove entire layers of human work (developers, support teams, analysts), allowing revenue to scale at a pace previously unseen.
The difference is increasingly visible in operating metrics. Leading AI companies are reaching revenue milestones in a fraction of the time compared with early SaaS leaders. Higher valuations are often a function of stronger early fundamentals, not purely speculative expansion – though caution should be assigned to long-term revenue durability.
Taken together, the data suggests that valuation cycles are increasingly front-loaded. As revenue scales faster, markets price future outcomes earlier in a company’s life. What used to be a milestone reached after years of operating history is now often embedded in the first financing rounds. The sequence has shifted. Scale is increasingly expected to follow valuation, rather than justify it.
Kingmaking: The Cause or the Effect of Compression?
A rational reaction to compression, but one that creates its own risks.
In 2025, roughly 40% of US venture dollars went into the top 10 deals, and close to 50% of LP capital flowed into the top 10 funds. Venture has increasingly become a market of haves and have-nots.
The natural question is whether kingmaking is the cause of this compression, or its consequence.
Our data suggests it is primarily a reaction.
Time compression began long before the current wave of capital concentration. As companies started reaching $1 billion valuations in years rather than decades, the window to enter early and build meaningful ownership narrowed dramatically.
If a company can plausibly reach tens or even hundreds of billions within a few years, concentrating capital into perceived category leaders becomes mathematically rational.
When entry valuations start high, return multiples compress unless outcomes get significantly larger. A $1 billion entry price requires a $10 billion or $100 billion outcome to generate traditional venture returns. The incentive is therefore to own as much as possible of the few companies capable of clearing that bar.
But once kingmaking takes hold, it creates a feedback loop.
The anointed company attracts more capital, more talent, and more strategic partnerships, reinforcing its lead and making markets increasingly winner-take-most. At that point, capital does not merely fund outcomes – it helps shape them.
This logic also reshapes fund structures. Larger vehicles can deploy capital at the speed and scale required to participate meaningfully in concentrated outcomes. As a result, kingmaking is reflected not only at the company level, but also in the justification of increased fund sizes.
Since the beginning of 2025, the ten largest US venture funds have accounted for roughly half of all capital raised.
The logic holds on paper.
It becomes harder to defend when private valuations move too far away from public market comparables. When private companies are capitalized at levels rivaling the world’s largest public firms, yet generate only a fraction of their revenue, kingmaking relies increasingly on future assumptions rather than present fundamentals.
When valuation embeds years of anticipated scale at inception, the margin for execution error narrows materially. Any slowdown in growth can compress both multiples and liquidity simultaneously, particularly if public markets are unwilling to validate those expectations. What appears rational in a regime of persistent hypergrowth can therefore become fragile when assumptions are tested.
It is rational, but fragile.
Investor Implications
Time compression changes the math of venture capital.
When companies begin their journey at $1 billion, the historical finish line becomes the starting line.
To generate traditional venture multiples, investors now require $10 billion or $100 billion outcomes. Such outcomes remain statistically rare. Only about a dozen VC-backed companies created since 2010 have crossed $100 billion in valuation, five of which did so in 2025.
For larger VC funds, it becomes a game of identifying true outliers. The margin for error narrows materially. Ownership matters more than ever.
Consider a recent AI round raising $200 million at a $1.8 billion post-money valuation. Even if a single investor secured half the round, it would acquire only 5.5% of the company for $100 million.
At a $20 billion outcome (historically considered a stellar exit) that 5.5% stake would return $1.1 billion. For a multi-billion-dollar vehicle, that may not even return the fund 1x.
In other words, even “winning” the allocation may not be enough.
The issue is no longer access to unicorns. It is whether the eventual outcome justifies the entry price.
Where Do We Go From Here
Future outcomes need to be larger than ever before.
The timeline to reach a unicorn valuation is compressing. The number of unicorns is increasing. Software capabilities, market scale, and AI-driven productivity gains are accelerating fundamentals.
But underwriting the next phase of outcomes increasingly requires a stronger assumption: that AI will meaningfully replace portions of the workforce and permanently reshape cost structures across industries.
Venture capital is pricing expectations years in advance. That is neither inherently right nor wrong.
The more relevant question is what outcomes can reasonably be underwritten, and whether those outcomes will still find a buyer when the time comes.
If $1 billion is now the starting line, the real bet is no longer on becoming a unicorn.
It is on becoming something far larger.
