In the first half of 2026, 263 mega-funding rounds exceeding $100 million captured 81% of global venture capital. The remaining 19% (approximately $51 billion in the United States) was distributed among a limited number of deals, with global deal count reaching a ten-year low according to CB Insights, rather than across tens of thousands of companies as one might expect.

This figure, drawn from CB Insights’ State of Venture Q2 2026 report, says something precise about the state of innovation capitalism: capital no longer diffuses, it accumulates. The global entrepreneurial ecosystem now operates at two speeds, with on one side a very closed club of mega-beneficiaries and on the other a silent mass of startups awaiting funding that never comes.

The Essential

AI is absorbing virtually all global venture capital. In the first half of 2026, 263 mega-rounds exceeding $100 million concentrated 81% of total investments, according to CB Insights. This concentration—unprecedented in twenty years—stems from the size of the bets necessary to compete with OpenAI, Anthropic, or xAI: foundation models require billions before generating a dollar in revenue. Early-stage and early-growth startups are bearing the brunt: their fundraising is becoming rarer in volume. In the long term, if capital finances only the existing giants, the creative destruction that regenerates the economy weakens, and ecosystems outside Silicon Valley risk lasting decline.

The Race for AI Has Changed the Geometry of Financing

Venture capital has always known cycles of concentration. The internet bubble of 1999-2000 also drew capital toward a few promising sectors. But the current dynamic differs on one decisive point: the gap between what a bet on AI costs and what a bet on the web cost is orders of magnitude different.

OpenAI raised $40 billion in the first quarter of 2025, then achieved an even more substantial round in the first quarter of 2026, estimated at $110 to $122 billion, the largest in venture capital history. xAI, Elon Musk’s company, closed a $6 billion Series C round in December 2024. Anthropic raised several billion more. These figures differ profoundly from what venture capital financed in the 2010s: they resemble funding for industrial infrastructure more than bets on startups. Training a competitive foundation model requires thousands of GPUs for months, electricity that states struggle to provide, teams of engineers whose salaries exceed those of most industrial sectors. A $100 million round no longer suffices to enter the race.

This shift has a mechanical consequence: venture capital funds themselves have repositioned. The large vehicles—Andreessen Horowitz, Sequoia, General Catalyst—have raised increasingly large funds and directed a significant portion of their commitments toward these mega-rounds. The logic is rational for managers: a position in OpenAI or Anthropic offers potential liquidity far superior to ten bets on Series A startups. The problem is that this individual rationality produces collective imbalance.

19% for Everything Else: What the Numbers Don’t Say

81% of capital for 263 rounds. The remaining 19% distributes among a restricted number of deals, with CB Insights signaling a deal count at a ten-year low. Of this total, the vast majority receives funding below $5 million, often insufficient to cross the product development threshold and reach initial customers.

Concentration itself is not the only problem. Venture capital has always operated with massive outcome inequalities: that’s its selection logic. What changes is the level at which selection operates. In previous cycles, mega-rounds coexisted with robust financing of seed and Series A stages, the pool from which tomorrow’s mega-rounds recruit. Today, this pool is contracting.

According to PitchBook data, the number of Series A rounds closed in the United States in the first half of 2026 shows a notable decline compared to the first half of 2024. In Europe, the trend is similar. Simultaneously, median valuations in seed rounds have actually progressed substantially over 2023-2025, reaching historical records—the median U.S. post-money valuation stood around $20 to $24 million in 2025, compared to $12 to $14 million in 2023 according to Carta and CB Insights. These valuation increases at the seed stage contrast, however, with the scarcity of financing volume at subsequent stages, revealing growing disconnection between ecosystem phases.

The disconnection is striking: while a handful of companies absorb tens of billions, founders attempting to raise $2 million to test a product face more selective investors, longer processes, and less favorable term sheets than in 2021 or 2022. The money is there. It’s elsewhere.

The Problem of Missed Creative Destruction

Economist Joseph Schumpeter theorized the mechanism by which capitalism regenerates itself: new companies emerge, shake up established players, and force the entire system to adapt. This creative destruction assumes capital is available for challengers, not just for champions already in place.

The current dynamic partially reverses this mechanism. Mega-rounds benefit overwhelmingly to companies that have already proven their model—OpenAI, Stripe, SpaceX in another sector—or to players close to major Silicon Valley ecosystems and a few Asian hubs. The capital concentrated there consolidates dominant positions rather than contesting them.

In public markets, research by Carl Benedikt Frey and co-authors showed that previous automation waves benefited more companies with access to capital for transformation than those without it. The same mechanism operates today at the venture capital scale: startups capable of attracting mega-rounds integrate AI as a competitive lever, those struggling to raise remain isolated from the productivity dynamic this wave promises.

This is one of the tensions raised by the question of capital taxation in a mutating economy: when gains concentrate in a handful of positions, redistribution and reinvestment instruments are tested. Venture capital, in theory, is supposed to play the role of a recirculation pump—take risks on what doesn’t yet exist. If it reorients toward what already exists, who takes over?

What Actors Attempting to Rebalance Are Doing

Concentration is not universally accepted as inevitable. Several categories of actors attempt to create counterweights, with different logics.

Funds specializing in emerging markets—Partech Africa, Flourish Ventures in sub-Saharan Africa, Kaszek in Latin America—maintain deliberate focus on early stages outside major technology hubs. They work with smaller tickets, local teams, and an investment thesis founded on access to under-penetrated markets rather than head-to-head competition with OpenAI. Their share of the global market remains modest, but their role in creating local entrepreneurial fabric is documented: Partech Africa co-financed several of Africa’s unicorns from 2015-2025.

Public investment funds play a growing role in Europe. Bpifrance deploys €2 to €3 billion annually in equity across all stages—startups, SMEs, mid-market companies, and large enterprises. For early-stage startups alone, direct investment is substantially lower, around €500 to €600 million according to first-half 2025 data, with an explicit policy of financing seed and regions outside Île-de-France. The British Business Bank plays a similar function in the United Kingdom. These institutions were designed precisely to correct market failures that private capital doesn’t address—what economists like Philippe Aghion have formalized in their work on innovation and Schumpeterian growth: public investment finances stages too risky or too long for private capital, complementing the market.

In the United States, the Small Business Investment Company Program of the Small Business Administration backs funds investing in rounds under $5 million. Its track record is mixed, but its existence signals that the question of small business financing is not ignored by public authorities.

Micro-funds and specialized seed funds-of-funds have also proliferated since 2020. Vehicles like First Round Capital, Precursor Ventures, or Y Combinator continue to bet on very early stages with tickets of $500,000 to $3 million. Y Combinator now runs four batches per year as of 2025, each counting approximately 140 to 200 startups—recent cohorts ranging from 144 companies for Spring 2025 to 197 for Spring 2026—maintaining a pace that stands in sharp contrast to the concentration logic of large funds. These actors don’t compensate for the mega-round aspiration, but they preserve a form of diversity in the ecosystem.

The Geography That Freezes

Capital concentration has a geography. California as a whole absorbs approximately 60% of U.S. venture capital, with a large portion going to the Bay Area—San Francisco and Silicon Valley together representing roughly 40 to 45% of the national total, with Silicon Valley proper at around 30 to 35% depending on sources. New York captures approximately 15%. The rest of the United States shares the remainder. In Europe, London, Paris, and Berlin concentrate the majority of significant rounds. This geographical inequality is not new, but mega-round dynamics accentuate it: a Boston or Austin startup can still access competitive Series A rounds, but a Cleveland or Glasgow startup has seen its chances diminish.

This asymmetry pairs with a sectoral asymmetry. Startups in biotech, deep tech, energy, or hardware suffer chronic financing shortage despite long-term prospects. Their development cycles are longer, their physical capital needs greater, and their risks less legible than those of pure software models. The pivot to AI has further reinforced this distortion: capital flows toward what can show traction quickly—namely language models and software applications exploiting them.

The paradox is that the American economy displays real vitality in its most advanced segments, driven by rising B2B productivity, yet this vitality masks impoverishment of the ecosystem’s intermediate layers. You can have a surface-level performing innovation economy with structural underinvestment in its own regeneration.

What the Long Series Says About Trajectory

Is current concentration a peak or a durable trend? Long-term data provides elements of response without permitting certain prediction.

Between 2010 and 2020, the share of mega-rounds in total global financing progressed nearly continuously. According to CB Insights, this share reached approximately 49% in 2020, with the 60% threshold first crossed in Q1 2021. The jump observed in 2025-2026, bringing this ratio to 81%, represents acceleration of the trend, not a rupture of different nature. What’s new is the speed.

This concentration level is not without precedent in capital’s history. It resembles the structure of railroad financing in the 1880s, when major infrastructure projects absorbed the majority of available capital, leaving little room for other investments. That cycle ended with saturation, valuation reversal, and capital reorientation toward new sectors. It took several decades.

The question of irreversibility is real. If a generation of potential founders concludes it’s impossible to raise without orbiting the major AI models, some simply won’t launch. The cost of this non-event figures in no statistics: you don’t count startups that were never created. But research on innovation clusters—particularly Daron Acemoglu’s work on technology and market power—suggests that financing concentration tends to self-reinforce: where capital flows, talent follows, and talent attracts more capital. The inverse also holds: ecosystems losing attractiveness over five to ten years struggle to regain it.

This finding is not a verdict. Technology cycles have a saturation logic. When major AI models consolidate positions and reduce their external capital needs—which may take three to seven years according to trajectories of previous technology infrastructure waves—funds must find new bets. The question is whether early-stage financing ecosystem will still be there to propose them. Dynamics of fragmentation in the global economy could paradoxically play a role here: partial deglobalization of financing could push regional capital to reorient toward local opportunities major American funds overlook.

What Founders Are Doing While Waiting

Facing traditional financing scarcity, part of founders adapt. Bootstrapping—building without external capital—has gained legitimacy since 2022, particularly in software, where infrastructure costs have fallen and no-code or low-code tools allow faster product validation. Communities like Indie Hackers or incubators specializing in bootstrapping document this trend: several hundred companies have reached $1 to $5 million in annual revenue without raising institutional funds.

This adaptation has its limits. It works in light software, not in deep tech, biotech, or energy. It favors founders already possessing personal capital or networks. It doesn’t solve long-cycle innovation financing.

Alternative funds also play a growing role: revenue-sharing, convertible notes, revenue-based financing. These instruments offer less dilutive conditions than classic equity for startups already generating revenue but struggling to raise. Capchase, Pipe, and several competitors offer these products in the United States and Europe, with documented customer-base growth since 2024.

These patches don’t erase the fundamental problem. They show founders aren’t passive, and the ecosystem produces responses at the margins of constraint. But an innovation economy structurally dependent on founder ingenuity to compensate for institutional financing absence isn’t an optimal innovation economy.

The real question is this: when the current concentration cycle reverses—and cycles eventually do reverse—will early-stage financing institutions, regional funds, public seed mechanisms have maintained sufficient capacity to irrigate the next wave? Or will the entire tissue need rebuilding from scratch, with the years of latency that implies?

Sources

  1. CB Insights, State of Venture Q2 2026, via Venture Curator, https://www.venturecurator.com/p/what-16000-companies-reveal-about
  2. PitchBook, Series A round data 2024-2026 (quarterly report, no stable URL)
  3. Bpifrance, 2025 annual report (bpifrance.fr)
  4. Small Business Administration, Small Business Investment Company Program (sba.gov)
  5. Daron Acemoglu and Simon Johnson, Power and Progress, PublicAffairs, 2023
  6. Carl Benedikt Frey, The Technology Trap, Princeton University Press, 2019
  7. CB Insights, State of Venture Q2 2026 (primary report), https://www.cbinsights.com/research/report/venture-trends-q2-2026/
  8. OpenAI – Official announcement of $40B round (March 2025), https://openai.com/index/march-funding-updates/
  9. Wikipedia – xAI Funding History, https://en.wikipedia.org/wiki/XAI_(company)
  10. NVCA 2026 Yearbook (PitchBook data), https://nvca.org/2026-nvca-yearbook/
  11. NYC Comptroller – VC Investment Report (October 2025), https://www.osc.ny.gov/press/releases/2025/10/dinapoli-nyc-area-countrys-second-largest-market-venture-capital
  12. Bpifrance – Official Investment Page, https://www.bpifrance.fr/nos-solutions/investissement
  13. SBA – SBIC Program Official, https://www.sba.gov/funding-programs/investment-capital
  14. Y Combinator – Official Directory, https://www.ycombinator.com/companies
  15. Dealroom – Europe Deep Dive (real-time data), https://dealroom.co/guides/europe
  16. CB Insights – State of Venture 2024 (seed valuations), https://www.cbinsights.com/research/report/venture-trends-2024/
  17. OpenAI – Announcement $122B round March 2026, https://openai.com/index/accelerating-the-next-phase-ai/
  18. xAI – Series C Announcement $6B (December 2024), https://x.ai/news/series-c
  19. CB Insights Q4 2020 – Mega-round share 2020, https://www.cbinsights.com/research/report/venture-capital-q4-2020/