Two Reports, One Market, Different Stories: Analyzing the Venture Capital Divergence of 2025

J.P. Morgan's Innovation Economy Update (H2 2025) and SVB's State of the Markets (H1 2026) both aim to tell the story of venture capital in 2025. The emerging theme is: the big are getting bigger and secondary markets are quietly, but powerfully, reshaping the rules of the game.

The Headline Numbers: Aligned on Capital, Split on Sentiment

On the surface, both reports agree on the big picture. J.P. Morgan reports approximately $337B in U.S. venture investment in 2025 across roughly 14,775 deals. SVB arrives at a nearly identical figure of ~$340B. Both acknowledge this puts 2025 on par with the 2021 peak in dollar terms.

But the framing could not be more different.

J.P. Morgan leads with cautious optimism: “Despite a tumultuous macro environment, the venture ecosystem was active, driven by insatiable demand for transformative technologies.” The report acknowledges concentration but treats it as one of several themes alongside AI adoption, macro trends, and a recovering exit market.

SVB is blunter. Its authors open by calling out the fundamental contradiction: “Overall, it was a great year in venture, even as some participants were left clapping on the sidelines.” The report explicitly compares the venture ecosystem to a “How do you do, fellow kids?” meme, arguing that a $10B Series J deal sitting next to a group of seed deals cannot meaningfully be called the same asset class. SVB names what J.P. Morgan implies: venture capital is bifurcating into two distinct markets, and the traditional one is shrinking.

The Big Are Getting Bigger: Concentration by the Numbers

Both reports devote significant space to capital concentration, but they measure it differently, and the contrast is illuminating.

J.P. Morgan frames concentration through deal tiers. The top 5 deals in 2025 captured 29% of all deployed capital, up from just 4% in 2021. The top 10 deals captured 34%, and the top 100 deals captured 53%. In 2025, the top 10% of deals accounted for 80% of deployed capital, compared to 67% in 2022.

SVB approaches concentration from the company-valuation perspective. The top 1% of companies by valuation absorbed 33% of all VC investment in 2025, up from just 12% in 2022. Meanwhile, the bottom 50% of companies by valuation accounted for only 7% of all investment. Nearly half of all capital went to “just a few dozen deals over $500M.”

The concentration isn't limited to the companies receiving capital — it extends to the funds deploying it. SVB's data reveals a stark divide on the manager side: just 594 investors participated in $500M+ deals in 2025, while 11,200 investors participated in sub-$100M deals. Yet that small cohort of large-check writers deployed nearly half of all venture dollars. The composition of these investors is telling — only 55% of participants in $500M+ deals are traditional VC firms, with the remainder comprising PE/growth equity (14%), crossover/hybrid funds (11%), and other non-traditional investors. By contrast, sub-$100M deals remain dominated by conventional VC firms (46%) and family offices/angels (27%).

On the fundraising side, the picture is equally lopsided. $1B+ mega-funds have captured a growing share of total LP commitments, even as overall U.S. VC fundraising fell nearly 20% YoY to its lowest level since 2019. Platform firms like a16z, which raised $15B across multiple strategies and stages, are operating at a scale that reshapes the market around them. SVB notes that since the pandemic, “deal sizes have skyrocketed, and fund sizes have grown in lockstep in an attempt to maintain ownership percentages and participate in the best deals.” The result is a fundraising environment where a handful of mega-managers raise the majority of LP capital, while smaller and first-time managers — who accounted for roughly one-fifth of all U.S. VC fundraising in 2025 — fight over the remainder.

Both reports arrive at the same conclusion from different angles: the power law distribution in venture has developed an extraordinarily heavy head — not just at the company level, but at the fund level too. In 2025, 24 companies received billion-dollar VC rounds, compared to 17 in 2021. But the total number of deals has fallen 25% since 2021 — the same capital is flowing into far fewer companies, deployed by far fewer managers writing far larger checks.

The Fundraising Gap: Who's Actually Raising?

Here is where the two reports diverge most sharply.

J.P. Morgan focuses on valuation growth. Early-stage median valuations reached an all-time high of $64M, and later-stage medians hit $92M. Series A valuations jumped 37% since 2021. Particularly in San Francisco, startups command a “considerable valuation premium,” with later-stage SF companies hitting median valuations of $361M versus $92M nationally. The picture is one of an expensive but rewarding market.

SVB tells the other side of the story. VC deal count under $100M — the measure most relevant to “normal” startups — is at a decade low. U.S. VC fundraising dollars fell almost 20% YoY to the lowest level since 2019. Fund managers are spending nearly two years in market, double the time it took during the 2021 boom. Graduation rates from one round to the next have been cut in half compared to pre-pandemic levels.

J.P. Morgan does acknowledge a troubling trend in follow-on rates: 82% of companies that raised early-stage rounds in H2 2023 had not raised a follow-on within 24 months, compared to 63% for the H2 2020 cohort. But this is presented as one data point among many, rather than the central story. SVB makes it the throughline: the median revenue required to raise a Series A increased 35% YoY, and only about 3% of seed companies graduate to Series A within 12 months.

The Role of Secondaries: Lubricant or Enabler of Concentration?

Both reports cover the growing secondary market, but their interpretations point to different conclusions about what secondaries actually do to the venture ecosystem.

J.P. Morgan reports that secondaries deal dollar volume continued to increase in 2025 as “founders, employees and early investors sought liquidity in the absence of a more active exit environment.” However, the number of secondary transactions continued to decline — down to 176 from a peak of 774 in 2022 — constrained by “persistent gaps between buyers' and sellers' price expectations.” J.P. Morgan describes secondaries as “an important mechanism, but not a primary driver of liquidity.”

SVB takes a more provocative position. The report shows secondary sell indications of interest (Sell IOIs) growing in lockstep with the number of U.S. VC-backed unicorns, which reached 857 with a combined valuation of $4.4T. Pricing has improved since 2023, and larger companies have disproportionately higher demand on secondary platforms. SVB quotes Eric Thomassian, Head of Private Company Relations: “The secondary market is not a substitute for IPOs. It's a liquidity mechanism that extends runway, relieves pressure of companies going public too early, rewards employees, and dampens volatility along the exit path — but it doesn't eliminate the need to go public.”

What neither report fully explores is the feedback loop that secondaries create in reinforcing the “big getting bigger” dynamic. Consider the mechanism:

The secondary market, paradoxically, may be contributing to the concentration it was designed to alleviate. By giving the largest companies the ability to stay private and continue raising private capital, secondaries extend the period during which these companies dominate the capital allocation landscape. The liquidity they provide is real but heavily skewed: larger companies with more trading demand get better pricing, while smaller companies on secondary platforms face thin markets and wider bid-ask spreads.

AI: Accelerant of Concentration

Both reports agree that AI is the defining force of the current venture cycle, but they quantify it differently. J.P. Morgan reports AI's share of all U.S. venture investment reached a high of 71% in Q1 2025, with over $400B raised by U.S. AI startups since the emergence of ChatGPT. SVB frames it as $560B in total AI investment since 2022, “on par with total deployment during the dot-com era.”

J.P. Morgan emphasizes that AI is driving legitimate enterprise adoption, citing J.P. Morgan's own 450+ AI/ML use cases in production and growing productivity gains, particularly among mid-size businesses. SVB takes a more cautious stance, acknowledging that “five AI companies have achieved values three times higher than all dot-com-era IPOs” and noting “clear signs of overexuberance in AI: high burn multiples, low revenue per employee and high valuation premiums for companies with less revenue than their peers.”

The AI dimension amplifies the concentration story. AI companies require enormous capital for compute and talent — SVB notes that the tech industry is transforming into a “capital-heavy industry.” This inherently favors well-funded incumbents. Of the 10 largest pure-AI rounds in 2025, J.P. Morgan notes that seven were for companies based in San Francisco, with only one outside the Bay Area. The geographic concentration layers on top of the capital concentration, creating a compounding advantage for the companies already at the top.

The Macro Divergence: Tariffs vs. Consumer Fragility

The reports also differ in what macro risks they choose to emphasize. J.P. Morgan devotes attention to tariffs, noting the average effective tariff rate rose from 2.4% to 16.8% by December 2025, making it one of the top five challenges for innovation economy leaders. They also engage in the dot-com comparison directly, arguing today's market is more robust.

SVB focuses on a different fragility: consumer inequality and the economy's dependence on AI spending. Their analysis shows that AI spending categories (hardware, R&D, software, data centers) are under 4% of GDP yet contributed 10-30% of quarterly GDP growth in 2025. Without AI spending, the economy would have experienced a deep contraction in Q1 2025. Meanwhile, the disparity between high and low earners is widening — from 2020-2023, the top 20% of wage earners added $30K more to their annual income than the bottom 20%, creating concentrated risk in consumer spending.

What This Means for the Innovation Economy

Reading these reports together reveals a venture market that is simultaneously booming and struggling, depending on where you sit. If you are one of the few dozen companies raising $500M+ rounds, 2025 was extraordinary. If you are among the thousands of startups competing for the remaining capital, it was one of the hardest fundraising environments in a decade.

The secondary market is playing an increasingly important but underappreciated role in this dynamic. By providing selective liquidity to the largest and most-traded private companies, secondaries reduce the urgency for IPOs and allow mega-companies to continue absorbing private capital. The effect, whether intentional or not, is to reinforce the very concentration that makes secondary liquidity necessary in the first place.

For LPs, the implication is that the traditional VC capital cycle — invest, exit, redistribute — is being elongated and partially short-circuited. For fund managers, the market is bifurcating between mega-platforms that deploy large checks and operate more like asset managers, and smaller, earlier-stage managers who offer the differentiated returns LPs are increasingly seeking. For founders outside the AI mega-round orbit, the path forward likely involves demonstrating capital efficiency, durable revenue growth, and realistic expectations about timelines.

Both J.P. Morgan and SVB are optimistic about 2026. J.P. Morgan hopes the return of exit-driven capital will “spark a strong 2026.” SVB expects “moderate deal growth driven by the AI platform shift and growing investor confidence.” But neither can escape the central tension: the innovation economy is thriving at the top and thinning everywhere else. Until capital finds its way back to a broader base of companies, venture capital in its traditional form will continue to exist in the shadow of something much larger — and fundamentally different — that happens to share its name.