Part I, VC Funding Is Back. Is Venture Capital?, showed that record headline funding can coexist with scarce capital for the ordinary startup. Its practical measure of market health was the share of companies retaining credible access to another institutional round, with aggregate dollars serving only as background.
Part II asks what drives that divergence and whether it is cyclical or structural.
The evidence points to a regime change—at least while risk-free yields remain attractive and private-market distributions remain weak. Selective liquidity now directs available venture capital to fewer managers and fewer companies.
The post-financial-crisis venture model depended on two conditions: capital had to be cheap, and capital had to circulate.
Both conditions have weakened. Government securities now offer meaningful liquid returns, while weak exits have prevented cash from returning to limited partners. Headline venture funding and reported valuations can still rise because AI mega-rounds dominate funding totals and the financing sample increasingly contains only the strongest companies.
The present market therefore exhibits a widening gap between observed prices and financing access:
The visible price of venture capital can rise even as access to venture capital deteriorates.
Viewing venture capital as a transmission system explains this divergence. The adjustment is operating through three mechanisms: selection, quantity and liquidity.
Investors are selecting fewer managers and fewer companies. Fewer financings are being completed. Secondary transactions are increasingly used to create liquidity and discover prices that primary rounds and portfolio marks have postponed.
Venture capital now operates as a two-speed market: extraordinary companies attract unprecedented sums while much of the remaining ecosystem faces persistent scarcity. In this regime, financing optionality becomes the decisive advantage: companies with time can defer a transaction until terms improve.
LP liquidity sets the first constraint
Venture-market analysis begins with the institutional allocation decisions that sit above funds and startups.
Dry powder, round counts, valuations and new-fund size describe activity inside the industry.
An institutional limited partner sets the upstream constraint by comparing any ten- or twelve-year lockup with available liquid alternatives.
That comparison has become harder to justify.
On 28 August 2026, the US ten-year Treasury yielded 4.73% and the thirty-year yielded 5.22%. [1] An LP could therefore earn a substantial liquid government yield without accepting manager-selection risk, capital calls, uncertain exits, opaque valuations, startup mortality or a holding period that may extend beyond a decade.
The liquid hurdle is no longer trivial
U.S. Treasury par yields on 28 August 2026. Venture must now compensate LPs for illiquidity and risk over a meaningful liquid government return.
Venture capital now has to clear a much higher hurdle, and it has to clear it with a product that has recently been delivering less cash than it promised.
Higher rates weaken venture through both valuation and allocation
The first effect is familiar. Venture-backed companies are long-duration assets: much of their expected value lies many years in the future. A higher risk-free rate reduces the present value of those distant cash flows.
The second effect matters more for the system. Treasuries and investment-grade credit now provide meaningful current returns, so the opportunity cost of illiquidity has risen. Even a modest increase in LP selectivity reduces the range of venture risks that can secure institutional capital. [1]
That produces a recognisable pattern. Fewer managers raise. Established franchises capture a larger share of commitments. Undifferentiated strategies struggle. Funds concentrate deployment in the companies they judge most capable of surviving.
Higher rates therefore do more than reduce theoretical valuations. They narrow the portion of the venture ecosystem that institutions are willing to finance.
Weak distributions turn a higher hurdle into a cash constraint
The first constraint concerns relative attractiveness. The second is mechanical and often binds harder. An LP can retain conviction in venture capital while lacking the cash for another commitment.
Consider a pension fund that committed $100 to a VC fund in 2021. Five years later, the manager may report that its investments are worth $120. With only $10 received in distributions, most of that reported value remains trapped inside private companies.
Now suppose another VC fund asks for a fresh $100 commitment. The reported $120 valuation supplies no cash for the new investment. The pension fund must draw on contributions, sell liquid securities or reduce allocations elsewhere.
This is the essential difference between reported performance and returned capital. TVPI combines realized distributions with the remaining value of unsold investments. IRR annualizes performance and can depend materially on interim marks. DPI measures the cash and securities actually distributed relative to paid-in capital. [2]
For pension funds, endowments and insurance companies, that distinction is fundamental. Pensions, university spending, insurance claims and new fund commitments all require realizable cash. An IPO, acquisition, secondary sale or other liquidity event must convert portfolio value into cash before the capital can be reused.
The distribution data shows how serious the circulation problem has become. McKinsey estimates that cash distributions as a share of private-equity assets under management fell to approximately 6% in the twelve months to June 2025, compared with a 16% average between 2015 and 2019. [3] Venture funds show the same constraint: Carta reports that median DPI for the 2019 and 2020 vintages remains barely above zero, and fewer than half of those funds have returned any capital to LPs. [4]
The circulation problem: distributions have collapsed
Cash distributions as a share of private-equity AUM fell sharply versus the pre-2020 norm.
These funds may still contain valuable companies. Some could eventually generate excellent returns. From the LP's perspective, the immediate problem is that the money remains locked inside the system.
That interrupts the normal circulation of venture capital:
The immediate constraint is trapped capital, even among LPs that still expect attractive returns from their existing venture portfolios. Until distributions arrive, each new commitment adds fresh cash to a system still retaining much of the capital already invested.
Higher bond yields constrain willingness; weak distributions constrain capacity.
Public-market shocks can tighten the constraint at the wrong moment
Institutional portfolios introduce a further asymmetry. Private-market allocations are usually measured as a percentage of total assets. If public equities and bonds fall quickly while private marks adjust slowly, private assets become a larger share of the portfolio without the LP making a single new commitment.
The mechanism is well established: when public assets fall faster than private marks, the private allocation ratio rises mechanically and can pressure LP liquidity and commitment pacing. [5] Even after acute stress passes, the effect can persist until public markets recover, private marks reset or distributions arrive.
Strong public markets can therefore restore private-market capacity. A severe correction can remove it precisely when venture-backed companies are themselves becoming more vulnerable. The mechanism is indifferent to timing, which is what makes it dangerous.
The yen carry trade can amplify a global liquidity shock and transmit it into institutional portfolios that ultimately finance venture capital. Federal Reserve research defines the trade as a short yen position funding higher-yielding assets, notes that rapid yen appreciation can produce asymmetric unwinds, and identifies major gaps in the measurement of derivatives exposure. [6]
The macro relevance has increased as Japan has normalised policy. On 31 July 2026, the Bank of Japan set its target for the uncollateralised overnight call rate at around 1.0%. [7] The policy shift increases the global liquidity system's sensitivity to yen appreciation after years of near-zero funding costs.
The transmission risk is specific. If another liquidity shock depresses public portfolios while private markets are already distribution-constrained, the transmission runs:
This is how developments several layers above the startup market determine which companies survive within it.
Scarcity compounds as it moves from LPs to GPs to startups
Scarcer LP capital concentrates at every layer of the venture system.
LPs consolidate commitments among managers with strong relationships, recognised franchises and credible evidence of distributions. Those GPs, facing a less certain fundraising environment, reserve more capital for their strongest existing companies. Those companies then attract a growing share of the capital still available.
The same selection mechanism repeats recursively:
Carta's Q1 2026 fund data shows the same consolidation. Vehicles with at least $100 million of commitments captured 57% of the capital raised by new funds on Carta in 2025, up from 31% eight years earlier. Although 89% of the funds in Carta's sample were smaller than $100 million, 54% of the capital sat in larger vehicles. [4]
LP capital is consolidating into larger funds
Share of capital raised by new funds on Carta captured by vehicles with at least $100m of commitments.
LPs are likely to retain venture exposure while concentrating commitments among fewer managers. This preserves headline participation while removing breadth from the ecosystem.
AI intensifies concentration across the venture market
AI now absorbs enough venture capital to reshape the composition of the market and distort its aggregate statistics.
At the equity level, AI absorbed more than 60% of the venture capital recorded by Carta in Q1 2026; foundation-model companies alone accounted for 14.2%. [8] PitchBook-NVCA likewise reports that the overwhelming majority of first-half 2026 investment flowed to AI companies and rounds of $100 million or more. [9]
This changes composition before it changes price. AI leaders absorb a disproportionate share of available capital, while later-stage medians and funding totals increasingly reflect the companies still able to clear a high selection bar.
AI can therefore lift headline funding while credible access to another institutional round continues to narrow. Its concentration of liquidity strengthens Part I's distinction between aggregate dollars and financing access.
Secondaries manufacture liquidity while operating exits remain scarce
The market has responded to weak distributions by manufacturing liquidity.
Evercore estimates that total private-capital secondary volume reached $226 billion in 2025 and $121 billion in the first half of 2026—the strongest first half on record and 19% above the prior year. GP-led transactions represented $65 billion, or approximately 54% of H1 volume, compared with $56 billion of LP-led activity. [10]
Secondaries are becoming liquidity infrastructure
Private-capital secondary transaction volume. H1 2026 was the strongest first half on record.
Historically, capital was expected to circulate through a simple chain:
Increasingly, the chain includes an alternative route:
This is a genuine financial innovation. It relieves LP pressure and keeps the system moving when IPO and M&A markets remain weak.
A secondary sale transfers ownership and creates liquidity while the underlying company remains private. Its transaction price can test a previous valuation; operating validation comes from an IPO or strategic acquisition. This distinction matters when interpreting strong valuation data.
Rising valuations increasingly reflect a selected financing sample
Rising venture valuations coexist with tight LP liquidity because the financing sample has become narrower and more concentrated.
Carta recorded $30.4 billion of startup funding in Q1 2026, with more than 60% going to AI. The down-round rate fell to 11.4%, while Series B and Series C pre-money valuations rose 17.2% and 12.5% year on year even as early-stage primary valuations softened. [8] PitchBook-NVCA then reported more than $400 billion of first-half investment, while stressing that the recovery remained concentrated among AI companies, mega-rounds and established managers. [9]
Reported valuations remain firm because tighter conditions change the composition and size of the financing sample before they change every completed round's price.
Investor selection and falling financing volume therefore carry much of the adjustment that a broad price decline would otherwise reveal.
AI has changed the population inside stage-level statistics
Stage-level statistics now require separate AI and non-AI cohorts because the valuation gap within each stage can exceed the gap between stages.
At Series A, Carta estimated a median valuation of approximately $300 million for foundation-model AI companies, compared with $55 million for non-AI businesses—a premium of roughly 5.5x. [8] The same stage label therefore spans distinct capital markets.
“Series A” now describes two different capital markets
Carta median Series A valuation: foundation-model AI companies versus non-AI companies.
The aggregate therefore says less about broad financing conditions than it once did. A record funding total can be produced by a narrow set of exceptional companies while the ordinary startup faces a harder path to its next round.
A median reduces the influence of outliers while remaining highly sensitive to changes in the population that completes a round. As fewer ordinary companies raise and unusually valuable AI businesses form a larger share of completed rounds, the observed sample strengthens while the broader startup population remains weak.
Every stage median therefore requires a composition check. Carta's simultaneous early-stage softness, later-stage gains and 5.5x Series A AI premium describe divergent financing cohorts, making a uniform repricing interpretation untenable. [8]
The Series A label now combines companies with radically different capital intensity, strategic scarcity and financing access.
Early-stage softness confirms that rates still matter. At later stages, the extraordinary AI capital cycle and a narrower priced-round sample obscure their effect.
The valuation signal is therefore split: early stages still reflect tighter money, while later stages increasingly reflect composition and selection.
Financing contraction appears first in quantity
A struggling startup can avoid a down round by postponing fundraising, cutting costs, accepting an insider bridge, issuing a SAFE or convertible, selling secondary shares, pursuing an acquihire, selling itself — or failing.
The weakest companies disappear from the priced-round dataset entirely.
A market that once financed 100 companies can produce 40 rounds at strong prices while many of the remaining 60 never complete another priced financing.
Selection then produces a higher observed median from a weaker underlying market:
The median improves partly because weaker companies are excluded from the sample.
That is survivorship bias inside venture-market statistics, and it is structurally identical to the distortion in aggregate funding. Mega-rounds make headline dollars describe the winners, while selective financing makes median valuations describe the survivors. Both statistics omit the condition of the broader ecosystem from which those companies were selected.
Secondary prices reveal the correction that primary rounds postpone
Private-market prices update episodically.
A listed company receives a new price whenever markets move. A private company may retain its previous valuation for two or three years simply because no transaction forces the old mark to change.
The system therefore contains three different concepts of value:
- Primary-market value: the price at which a selected company can raise fresh capital.
- Portfolio NAV: the value at which an existing investor continues to carry the asset.
- Clearing value: the price at which another investor or strategic buyer will actually transact.
These sit close together when liquidity is abundant. In a selective market, they diverge sharply.
Brex provides a visible example of the gap between the last primary mark and an eventual clearing value. The company announced a $12.3 billion valuation in 2022; Capital One announced a $5.15 billion acquisition in 2026, approximately 58% below that peak. [11][12] As a single case, Brex illustrates the mechanism and scale of forced price discovery across one asset; it offers no market-wide estimate for 2021–22 valuations.
Clearing value can be far below the last primary mark
Brex illustrates how delayed private-market price discovery can end when a strategic transaction forces a clearing price.
Recent companies financed under today's stricter selection may therefore be reasonably marked. The overhang sits inside the 2021–22 vintages, where companies raised unusually large rounds at unusually high prices and then avoided new transactions while consuming the capital.
Eventually something forces discovery: the company needs another round, an investor seeks liquidity, employees exercise options, a strategic buyer makes an offer, or the company runs out of cash.
Until that moment, the correction occurs through channels that headline valuations barely capture:
- time, as companies wait longer to raise;
- quantity, as fewer companies secure financing;
- structure, as bridges and convertibles postpone repricing;
- secondaries, as shareholders accept discounts;
- M&A, as buyers establish a clearing value;
- failure, as some assets ultimately clear at zero.
Reported primary valuations are lagging indicators of stress. Down rounds often record the final stage of a correction already visible in financing delays, structure and secondary discounts.
The more useful forward indicator is the gap between the latest stated valuation and the price at which existing investors can actually obtain liquidity. A wide gap signals stale marks. A narrowing gap signals that price discovery is progressing.
Secondaries now perform two functions: they release liquidity and establish clearing prices for an asset class that historically avoided continuous repricing.
The market is clearing through selection, quantity and liquidity
The transmission mechanism runs from sovereign yields to company-level clearing values.
Higher sovereign yields increase the opportunity cost of illiquidity. Weak exits suppress distributions. Weak DPI reduces LP commitment capacity and increases manager selectivity. Capital concentrates among fewer GPs, which concentrate reserves among fewer companies. AI absorbs an exceptional share of venture equity. Companies excluded from priced rounds delay, restructure, sell or fail. Secondaries then provide liquidity and expose the gap between portfolio marks and clearing values.
This produces a market in which headline funding and valuations rise while deal counts fall, stage conversion deteriorates, intervals between rounds lengthen, older vintages face wider gaps between stated and clearing values, GP fundraising concentrates and startup mortality increases.
These conditions describe a market clearing through selection, quantity and liquidity, with completed-round prices recording only part of the adjustment.
The most likely outcome is a prolonged two-speed ecosystem
Three paths remain plausible.
A liquidity shock could move the system abruptly from delayed to forced price discovery. The severity of the outcome would depend on how many companies lose the ability to choose when they transact.
Capital productivity and financing optionality will define the next cycle
Headline funding is now too compositionally distorted to serve as a measure of venture health. A more useful dashboard would monitor the transmission chain itself.
At the macro level, the relevant indicators are sovereign yields, Japanese policy rates, USD/JPY and credit spreads. At the LP level: DPI, secondary volume, discounts to NAV, and private-market allocations relative to target. At the GP level: fund closings, median fund size, the interval between funds, and emerging-manager fundraising. At the startup level: deal count, stage conversion, time between rounds, bridge and insider-round frequency, headcount contraction, cash runway, distressed M&A and shutdowns. For price discovery: the gap between primary valuation, portfolio NAV and secondary or M&A clearing value.
Together, these indicators measure the breadth of capital circulation and the degree to which liquidity is accumulating around the strongest visible assets.
For fifteen years, cheap and abundant capital made growth the dominant strategic objective. A selective-liquidity regime rewards a different variable: capital productivity.
How much revenue can a company create per dollar raised? How quickly can it reduce dependence on external financing? Can it survive a thirty-month fundraising interval? Can a VC fund convert higher NAV into distributed cash? Can an LP continue allocating without depending on continually rising public markets?
These questions point to the same source of bargaining power: the capacity to defer a transaction.
A startup with runway or positive cash flow can choose when to raise. A fund producing distributions can choose which LP capital to accept. An LP with liquidity can invest when other allocators are constrained. A cash-rich acquirer can dictate terms when sellers require exits. A secondary investor with committed capital gains leverage when shareholders need liquidity.
The common advantage is not possessing capital. It is possessing time.
The market has reopened before it has healed
Venture capital now occupies a transitional regime between the initial contraction and a broad new expansion.
The market has moved beyond the initial interest-rate shock, the first valuation reset and the worst period of financing paralysis. Investment, fundraising, valuations and exits are recovering.
LP-level liquidity remains impaired.
LP distributions remain weak. Fundraising remains concentrated among established managers. Record secondary activity shows that liquidity still has to be manufactured. Many companies financed during 2021–22 have yet to encounter a genuine clearing price. SVB estimates that 2,345 VC-backed companies are on pace to fail in 2026—the highest level in recent history—with more than one-third of the failures founded during the zero-rate era. [13]
The venture market is therefore best understood as being in a phase of:
Selective reopening over an unresolved liquidity overhang.
Different parts of the system are operating at different points in the cycle.
At the top of the market, leading AI companies have already entered another expansionary phase. They raise unprecedented amounts of capital at rising valuations.
Across the broader startup population, the market remains in late contraction or early repair, with financing governed by much harsher investor selection.
At the LP and fund level, the system remains inside a liquidity-repair cycle. Improving portfolio marks have yet to produce sufficient cash distributions.
The market's apparently conflicting signals follow directly from these different positions in the cycle.
Headline funding can reach record levels while an ordinary startup's probability of securing another institutional round deteriorates. Median valuations can rise while weaker companies disappear from the financing sample. Fund NAVs can recover while LPs remain unable or unwilling to recommit. Secondary-market volume can reach records precisely because traditional exits remain insufficient.
The market has reopened before it has healed.
The next phase depends on whether improving valuations and exit activity can restore the circulation of capital before another macroeconomic shock forces the remaining correction.
What would constitute a broad recovery
A broad recovery would require more than a handful of mega-IPOs, AI financings or large fund closes. It would require:
- rising DPI over several consecutive quarters;
- broader exit participation;
- improving stage-conversion rates;
- stronger non-AI deal counts;
- narrowing secondary discounts while volume remains healthy; and
- renewed fundraising access for managers outside the established franchises.
Until those indicators improve together, the recovery remains narrow. A durable turn requires improvement across several of these indicators at the same time.
The alternative is forced price discovery
If sovereign yields remain elevated, Japanese monetary normalisation produces global deleveraging, public markets correct or LP liquidity deteriorates further, a large population of 2021–22 companies may be forced to seek financing or liquidity simultaneously. The correction would then move from delayed repricing through bridges, extensions and stale marks to explicit repricing through down rounds, secondaries, distressed M&A and failure.
The evidence suggests that transaction activity is past its trough while the liquidity overhang remains. [9]
The present market sits in the middle of a transition from the broad liquidity of 2021 to a more selective capital regime.
The strongest companies and managers have already entered a new capital cycle. Much of the remaining ecosystem is still clearing the excesses of the previous one.
Financing optionality now creates the decisive divide across companies, sectors and venture firms.
Participants able to defer financing hold leverage over those dependent on continuous liquidity.
LPs with available cash can dictate terms to GPs. GPs producing distributions can command LP commitments. Funds with reserves can defend their strongest portfolio companies. Startups with runway or positive cash flow can choose when to raise. Those that require another financing round merely to survive remain exposed to every tightening further up the capital chain.
A genuine turn in the venture cycle requires cash distributions and broad, credible access to another institutional round to recover together.
Until then, strategic advantage comes from extending runway, improving capital productivity and retaining control over financing timing.
Survival converts fundraising from an immediate requirement into a discretionary choice, while credible access to another institutional round preserves the value of exercising that choice.
References
- [1]U.S. Department of the Treasury, “Daily Treasury Par Yield Curve Rates,” 28 Aug 2026. Source
- [2]Institutional Limited Partners Association, “Private Equity Glossary” (DPI and TVPI definitions). Source
- [3]McKinsey & Company, “Global Private Equity Report 2026,” 10 Feb 2026. Source
- [4]Carta, “VC Fund Performance: Q1 2026,” 4 Jun 2026. Source
- [5]Institutional Limited Partners Association, “LP Perspectives on the Impact of COVID-19,” 8 Apr 2020 (denominator-effect mechanism). Source
- [6]Federal Reserve Board, Gagnon and Chaboud, “What Can the Data Tell Us About Carry Trades in Japanese Yen?” IFDP No. 899, Jul 2007. Source
- [7]Bank of Japan, “Statement on Monetary Policy,” 31 Jul 2026. Source
- [8]Carta, “State of Private Markets: Q1 2026,” 29 May 2026. Source
- [9]PitchBook-NVCA, “Q2 2026 Venture Monitor,” Jul 2026. Source
- [10]Evercore Private Capital Advisory, “H1 2026 Secondary Market Review,” Jul 2026. Source
- [11]Brex, “Welcoming Karan and Our Series D-2 Round,” 11 Jan 2022. Source
- [12]Capital One, “Capital One to Acquire Brex,” 22 Jan 2026. Source
- [13]Silicon Valley Bank, “State of the Markets Report H2 2026.” Source