- A "barbell" effect has emerged in venture capital. At one end, massive, late-stage rounds concentrated in a few (mostly AI) companies. At the other, early-stage investing thrives on discipline and finding "pre-consensus" bets, with a hollowed-out middle.
- Specialization is the winning strategy for investors and founders – success now requires choosing a clear lane. Investors must decide between a "top-down" (access-driven) or "bottom-up" (conviction-driven) approach.
- Every six months, SVB distills venture capital fundraising, investment and exit dynamics across the venture landscape into its premier State of the Markets report. The report combines SVB’s proprietary data with investor conversations and independent market analysis to equip the innovation economy with timely, practical intelligence.
If 2021 was about velocity and 2022–2023 was about triage, the end of 2025 into 2026 feels surgical: fewer deals, bigger checks and conviction concentrated at the very top.
This tension – abundance at the apex and measured scarcity elsewhere – was a central theme at our State of the Markets H1 2026 launch event earlier last month where we hosted a panel of leading investors to discuss the report’s findings. Together, they dissected what happens as an asset class matures, capital concentrates and AI reshapes both companies and the venture model itself to create new opportunities alongside new challenges. But rather than a story of constraints, the conversation revealed a venture landscape that’s maturing, sharpening and evolving.
Following is a recap of the themes discussed among the panel featuring:
- George Mathew, Managing Director, Insight Partners
- Ben Lerer, Managing Partner, Lerer Hippeau
- Rebecca Kaden, Managing Partner, Union Square Venture
Market analysis: The great bifurcation in venture capital
In 2025, 33% of all US VC dollars went to the top 1% of companies by valuation, up from 12% in 2022. AI valuation premiums versus non-AI business models reached 222% at Series D+ in 2025, with triple-digit premiums even at earlier stages. Meanwhile, just 7% of capital reached the bottom 50%.
Median revenues at raise are higher than 2021 across every stage. Seed companies raising in 2025 showed 322% YoY growth versus 959% in 2021 but off a larger revenue base ($363K vs. $156K).
The translation? Slower growth, more revenue, much higher expectations, and ironically, healthier fundamentals than the frothy days of 2021.
Rebecca Kaden, Managing Partner at Union Square Ventures, posed this question around what she referred to as “the elephant in the room”: “Are investors deploying the capital that companies of today need to scale, or are companies raising massive rounds because venture has so much money to deploy?”
Insight Partners Managing Director George Mathew offered his view of what this capital deployment is building: “This cycle of capital has built out the infrastructure that was necessary. In a few years, with all the scaffolding in place, I expect we will see vertical systems and vertical automations that will look nothing like the applications we’ve known in the past.” In other words, today’s investments are laying the foundation for the next generation of transformative companies.
For perspective, past platform shifts took time to mature. Home internet crossed 50% adoption in 1999; e-commerce followed about eight years later. Platform shifts are lumpy, but history suggests the wait is worth it. Adoption, innovation and monetization rarely move in lockstep but tend to eventually converge.
How capital allocators are adapting (and thriving)
The shifts in company building have also created new opportunities for allocators willing to adapt. Ben Lerer, Managing Partner at Lerer Hippeau, framed the change pragmatically: “There’s just more capital than there are good ideas right now. Even the most celebrated companies’ rounds don’t feel like venture and are unlikely to resemble classic venture returns.”
But for early-stage specialists like Lerer Hippeau, this creates breathing room. “Venture has become obsessed with a small group of really, really, really crazy big companies,” Lerer said, “and we’re not competing in that asset class.”
The implication? Less noise, clearer lanes and better opportunities to build meaningful stakes in exceptional early-stage companies.
Kaden framed today’s venture landscape as two distinct games: “Top-down venture is about access to a finite number of market-winning investments. Bottom-up venture is about constructing a high-conviction, pre-consensus portfolio.” The former relies on scale, access and sustained capital; the latter relies on smaller fund sizes and selective early bets where being right matters more than broad coverage.
The “middle” is marked by growth strategies that once thrived on modest multiple expansion but has largely thinned out. Higher capital costs and ruthless pricing leave little room for alpha. But this clarity is a feature, not a bug. It’s forcing investors to make real strategic choices rather than drifting through the mushy middle.
George added: “I don’t think the same venture capitalists will go after future opportunities, because either you’re going to be great at the early stage and hold on to those companies, or you’ll go in to do large-scale capital deployment that looks different from traditional venture.”
Translation: the best investors may see clearer paths to differentiation than we’ve previously seen.
Kaden agreed, advising that early-stage firms can embrace their distinct game. The opportunity to look a stage earlier than the red-hot center – and even a concentric circle out from where most attention lies – creates significant opportunity.
Founder archetypes: Identifying successful founders in a barbell market
The panel agreed this market barbell in allocation is visible among founders, too, and creating opportunities on both ends.
On one end: experienced operators who know how to recruit, navigate regulation and build durable organizations. George cited infrastructure opportunities and the success of Weights & Biases: “Maturity is necessary when building infrastructure. Lukas Biewald was my first investment at Insight. We exited to CoreWeave last year. I really believe experience framed his impact. Lukas had built CrowdFlower in the past. As a second-time founder, he had the wherewithal to go build Weights & Biases at scale.”
On the other end: young, hungry outsiders. George lit up describing the first healthtech investment in his portfolio, backing a team of consummate outsiders who scaled from a six-figure to eight-figure ARR in 12 months.
The panel agreed that the “middle” is disappearing here too; there are fewer founders who are neither deeply seasoned nor unusually spiky. But here’s the opportunity: for investors who can spot genuine outliers early, the signal-to-noise ratio is improving.
However, graduation rates remain sobering, as only 13% of Series A companies raised a Series B within 24 months. The funnel is tighter, and timelines are longer. But those that do graduate are more resilient and capital-efficient businesses than their 2021 predecessors.
Liquidity outlook: The dam is full
If capital is concentrated at the top, liquidity is the pressure valve at the bottom – and pressure is building in productive ways.
There are now 857 companies with sell-side indications of interest on Forge, a private markets platform, moving in lockstep with the growth in VC-backed unicorns. The last known value of US VC-backed unicorns stands at $4.4T. Half generate more than $800M in revenue, suggesting a deep bench of real businesses preparing for next steps.
M&A dynamics are shifting, too. The share of deals with a VC-backed buyer climbed to 46% in 2025, and sale-price-to-capital-raised multiples have compressed. Strategic buyers are more price-sensitive; financial buyers are increasingly in the driver’s seat.
And yet, sentiment isn’t euphoric, which may be healthy. SVB’s Podcast Sentiment Index showed mixed sentiment across 2025, only modestly above its six-year median, even as ChatGPT’s release in 2023 triggered what one industry voice called “the most important developer platform… maybe ever.” Optimism, but measured; enthusiasm, but earned.
As Lerer pointed out, “It’s amazing if these six companies [SpaceX, Anduril, Stripe, Anthropic, OpenAI, Databricks] go public. But there’s this giant overhang of thousands of SaaS businesses that were really good companies. How does that all work its way through the system?” The answer may take time, but the quality of the backlog suggests the next wave of liquidity could be substantial.
Conclusion: Two games, both winnable
The macro takeaway isn’t that venture is back to 2021 – it has bifurcated. Both paths are viable for those who understand the game they’re playing.
At the top: mega-rounds, AI premiums, index-moving public comps, and $192B raised by the five largest AI companies alone – nearly matching the scale of entire prior eras. Global AI funding has already reached $560B, approaching dot-com totals in real terms. We’re witnessing the infrastructure build-out of a generation.
Below that: slower graduations, longer timelines, tighter check-writing and buyers demanding efficiency. But also: better unit economics, more realistic valuations and opportunities for investors who excel at true company-building.
For founders and VCs, the implication is clear. The market is open for companies that can demonstrate platform-level potential or platform-level performance. And for those focused on the fundamentals rather than the headlines? There’s never been a better time to find overlooked gems, build with discipline and generate outlier returns in the 67% of US VC dollars outside the top 1% of companies that the market isn’t chasing.
The bar is higher. The path is clearer. And for those who adapt, the opportunities are real.
To learn more about these trends and understand what they can mean for your business, read the full H1 2026 State of the Markets report, or contact Ash Bhatia (abhatia@svb.com).
