Venture Capital Is Breaking Records. There May Never Have Been Less Reason to Trust the Headline Numbers.

The market is expanding in dollars while contracting in breadth.

On the surface, 2026 looks like one of the strongest years venture capital has ever seen. Funding has surged, valuations have climbed, and exit values have reached levels that would have looked almost impossible a few years ago.

But the deeper you go into the data, the less the headline numbers tell you about the actual market.

In Q2 2026 alone:

• $227.4 billion was invested globally in venture capital.
• Just 10 deals accounted for roughly $105 billion, or about 46% of all capital deployed.
• CB Insights reports that mega-rounds of $100M and above absorbed 81% of global VC funding.
• At the same time, global deal activity remained historically weak, at what CB Insights calls its “lowest point in over a decade.”
• Global VC-backed exit value reached approximately $1.9 trillion in a single quarter, yet the number of exits did not rise with it: 811 in Q1 and 775 in Q2 as first published.

So the market is not simply growing.

Capital is concentrating.

More money is being deployed, but increasingly into fewer companies, larger rounds and later-stage assets. And that creates an uncomfortable question for investors:

Are we looking at a venture recovery, or are a small number of extraordinary companies making the entire market look healthier than it really is?

That distinction matters. Record funding does not automatically mean more investable opportunities. Record exit value does not automatically mean record liquidity. And rising private-market valuations do not automatically mean better future returns.

The numbers are real. The story they appear to tell is not always the real one.

The $1.9 Trillion Exit Record That Wasn’t $1.9 Trillion of Cash

Did liquidity actually return?

Venture capital supposedly produced $1.9 trillion of exits in one quarter. That sounds like liquidity came roaring back.

It did not.

In Q2 2026, KPMG and PitchBook recorded $1.908 trillion of global venture-backed exit value, more than the previous full-year record of roughly $1.5 trillion set in 2021. About $1.8 trillion came from the US, with roughly $1.7 trillion attributed to public listings.

But there is a catch. Exit value is not the same thing as cash returned to investors.

PitchBook values a public-market exit using the company’s valuation at the IPO price. It does not simply count the cash raised in the offering.

SpaceX makes the distinction almost impossible to ignore. KPMG reports that the company raised $75 billion in the offering, rising to $85.7 billion after the underwriters exercised the full over-allotment option. Yet the quarter still produced roughly $1.7 trillion of US public-listing exit value, with SpaceX as the dominant driver.

So one quarter can simultaneously produce $1.7 trillion of reported public-listing exit value and $85.7 billion of actual SpaceX IPO proceeds.

Those are not competing numbers. They measure completely different things.

And that difference matters, because an investor looking at the first number may conclude that liquidity is back. The better conclusion is narrower: public-market valuations are back, and realised liquidity has to be measured separately.

Why this matters for your money

A portfolio can look dramatically healthier because one large company received a public-market valuation. That does not mean LPs received equivalent cash distributions. It does not mean every shareholder sold. And it does not mean the rest of the venture market suddenly became liquid.

So when a manager tells you exit value has exploded, the next question is not how big the exit was. It is how much cash actually came back.

Annual VC-backed exit value, $B. Quarterly labels read in KPMG Venture Pulse Q2 2026, p.14; annual totals are our sums, as the report prints none. 2026 is a half-year and is not comparable with the full years beside it.
Data note: PitchBook changed its exit-value methodology in 2020, expanded IPO classification to include SPACs and reverse mergers in 2021, and introduced a new M&A extrapolation in 2025. Historical comparisons should be interpreted with caution.

The strangest part is that H1 2026 alone is already roughly 1.5 times the previous full-year record. But that does not mean venture investors received $2.3 trillion in cash. It means the valuation methodology produced $2.3 trillion of VC-backed exit value. That is a very different statement.

KPMG says the quarter was “heavily skewed by the merger of SpaceX and xAI, and then the subsequent IPO of the combined company.” SpaceX had acquired xAI in Q1 2026 for $250 billion.

Nothing is wrong with the metric. The problem begins when valuation is mistaken for liquidity.

A record exit-value number can tell you that valuations recovered without telling you that liquidity recovered. That is the whole point of this section.

More Dollars. Fewer Doors.

Is venture really booming, or just concentrating?

If the $1.9 trillion exit record made venture look more liquid than it really is, the funding numbers create a second illusion: more money does not necessarily mean more opportunity.

Global venture funding reached $560.3 billion in the first half of 2026, according to KPMG and PitchBook. That is more capital than venture received during the entire years of 2022, 2023, 2024 or 2025. Only the 2021 boom was larger. And 2026 had only used six months.

Global VC funding, $B. 2026 represents H1 only. Source: KPMG Venture Pulse Q2 2026, p.7. Annual totals are calculated from the quarterly figures reported by KPMG and PitchBook.

At first glance, that looks like a historic recovery. Then you look at who is actually receiving the money.

CB Insights, using an independent venture database, describes Q2 2026 as the quarter when global deal count fell to its “lowest point in over a decade.”

So we have two things happening at the same time: record-level capital deployment, and historically weak deal activity.

That is not a contradiction. That is concentration.

And the concentration becomes even harder to ignore when you look inside the headline number. In Q2 2026, KPMG and PitchBook recorded $227.4 billion of global venture investment. The ten largest deals alone absorbed approximately $105 billion. That means roughly 46% of all venture dollars deployed globally in the quarter went into just ten companies.

CB Insights sees the same phenomenon through a different dataset: $100M+ mega-rounds captured 81% of all global venture funding in Q2.

This changes the meaning of the boom. The market is not necessarily producing dramatically more investable companies. It is producing dramatically larger cheques for a much smaller group of companies.

A rising market can expand your opportunity set. A concentrating market can do the opposite.

You may see record amounts of capital entering venture while finding that the companies responsible for those records are already held by top-tier funds, priced at premium valuations, heavily oversubscribed, or simply unavailable to you.

So the real question is not how much money is going into venture. It is: how much of that market can you actually invest in? That is a very different question.

The number we refuse to pretend is precise

There is another statistic that appears to strengthen the same argument. KPMG reports 811 VC-backed exits in Q1 2026 and 775 in Q2, apparently a 4.4% decline. It would be convenient to write that exit activity is falling while exit values explode.

But we cannot defend that conclusion yet.

Private-market databases continue discovering and confirming transactions months after a quarter closes, and the revisions can be enormous. KPMG’s initially published Q1 2026 global deal count was 8,464. In the next data vintage, the same quarter appeared as 10,277 deals. That is not a rounding difference. It is more than 1,800 additional transactions appearing after first publication. Q2 2025 was similarly revised from 7,356 to 8,860, while Q1 2026 exit value moved from $413.5 billion to $434.7 billion.

So a reported 4.4% change in exit count is simply too small to treat as a confirmed trend when the underlying database has recently experienced double-digit revisions.

The $1.9 trillion exit-value spike is different. A move from $434.7 billion to $1.908 trillion is far too large to disappear through normal database revisions. One conclusion is robust. The other is not. And knowing the difference is part of investing.

Private access is not disappearing. It is splitting in two.

How do I get access to the growth?

There is a popular version of this argument that says venture funds are dying and money is fleeing the asset class. The data does not support it.

NVCA, reporting on the same quarter, says fundraising “rebounded sharply, with venture firms raising nearly as much capital through June as during all of 2025” and that exit activity improved, offering “encouraging signs that liquidity conditions may finally be strengthening.”

But read the qualifier, because that is where the story is: commitments “remained concentrated among a small group of established managers,” and “strong headline numbers continue to mask significant concentration across investment, fundraising, and exits.”

Meanwhile the structural routes into private companies are expanding, not contracting. The two tables below set the sides of that split next to each other, each with its own scope, because they come from different providers measuring different markets.

And the companies themselves stay private longer. The aggregate value of US unicorns is $4.3 trillion. PitchBook’s own comment on that figure is the clearest statement of this shift we have found:

The private market is becoming harder to access conventionally and easier to access structurally. Those are not the same thing.

Side one: conventional access is concentrating

One firm took more than eighteen cents of every new dollar committed to US venture funds over a twelve-month period. That is the concentration story on the fund side, and it is the reason a rebound in the aggregate can coexist with a first-time manager failing to close.

Side two: structural access is expanding

GP-led volume more than tripled in five years. That is not a workaround that appeared during a bad patch. That is infrastructure being built.

The last row is the one to keep in view. A wider set of doors is not the same as a better set of doors, and the people already walking through them are not uniformly happy about what is on the other side.

Put plainly: traditional access is concentrating in established funds, while secondaries, continuation vehicles, direct deals and SPVs are becoming permanent infrastructure rather than workarounds. The IPO is increasingly a liquidity event after years of private value creation, rather than the starting line for public investors. By the time public markets gain access, a materially larger share of a company’s growth may already have happened.

What investors are paying today

Am I overpaying?

Global medians for 2026 to date. This is the table to price against.

Logarithmic scale: each gridline is ten times the one below. The range runs from $0.2M to $1,697.5M and a linear axis would flatten every early stage to nothing. Source: KPMG Venture Pulse Q2 2026, pp. 9 and 10.

A median seed round is $2.7 million. A median Series D+ company is priced at $1.7 billion before the round opens. Those two companies share a market only in the sense that they appear in the same report.

How to read a manager in a mark-heavy market

Which metrics matter now?

In a quarter where the largest liquidity headline in history was mostly a valuation, the natural reaction is to throw out valuation-based metrics. That is the wrong correction, and the institutions did not make it.

ILPA standardises the reporting of IRR and TVPI or MOIC alongside cash-flow data rather than in place of it. And in McKinsey’s survey of 300 large LPs, DPI is now tied with MOIC as the second most important metric shaping allocation decisions, with IRR still the leading focus.

The distinction that matters is not which metric to use. It is which part of the return each one describes. TVPI includes unrealised NAV. Interim IRR also depends on unrealised valuations. DPI excludes unrealised holdings entirely, which is exactly why it behaves differently in a year like this one.

In a mark-heavy market the question is not “what is your IRR?” It is “how much of that IRR has actually been realised?”

Read the column on the right next to a quarter in which one listing produced $1.7 trillion of exit value. Three of these four measures can improve without a single dollar reaching an investor.

The context for that question: McKinsey reports DPI as a share of total private equity assets under management at about 6% for the twelve months to June 2025, against a 2015 to 2019 average of 16%, with five-year rolling DPI at its lowest recorded level. The counterweight, and it is a real one, is that buyout fund distributions exceeded capital calls in 2025 for the second year running.

What the data cannot tell you

How do I choose one deal?

Everything above describes the weather. It tells you the market is concentrated, that a record exit figure is mostly a valuation, and that access is bifurcating rather than closing.

It does not tell you whether one specific company survives that. That comes down to whether the business reaches its next milestone without a new round, whether the plan was built from measured demand or from what the market might give, and whether the people running it have done this before when it was difficult.

This is why I would rather work inside a company for a year than read its deck for an hour. In a market this concentrated, most of what separates a good company from a good investment sits in things no dataset records.

What this adds up to

The $1.9 trillion record tells us something important about venture capital. Just not what the headline suggests.

Capital is available. Liquidity is less certain. Valuations are high. Access is concentrating conventionally while widening structurally. And a handful of companies can now move an entire asset class’s statistics.

That makes aggregate numbers less useful, not more. For an investor the questions become narrower:

• How much of the return is realised?
• How dependent is it on a mark?
• Why am I being offered this particular deal?
• What am I paying relative to its stage and fundamentals?
• Who still owns it after the liquidity event?
• And what has to happen operationally for my return to exist?

A record market can still be a bad place to buy. A concentrated market can still contain exceptional investments.

The difference is underwriting.