Insights from CB Insights — State of Venture FY2025
Venture capital in 2025 has returned in size, but not in structure. Funding volumes rebounded sharply after two constrained years, reaching approximately $469 billion globally. Yet this recovery masks a deeper shift. Capital is no longer broadly distributed across sectors, stages, or regions. Instead, it is tightly concentrated around a narrow set of technologies, geographies, and operating models.
The State of Venture FY2025 does not describe a cyclical rebound. It documents a structural transition in how innovation is financed, scaled, and controlled. The implications extend well beyond venture portfolios. They affect innovation strategy, infrastructure planning, geopolitical exposure, and long-term resilience across industries.
A rebound defined by concentration, not diversification
At first glance, 2025 looks like a return to form. Total venture funding rose materially from 2023–24 lows. However, deal count continued to decline, falling to roughly 29,500 transactions globally, far below the peak years of the previous cycle. The market is placing fewer bets, but those bets are significantly larger.
Two patterns dominate the data. First, artificial intelligence captured approximately 48% of all venture funding, the highest share ever recorded. Second, mega-rounds of $100 million or more accounted for roughly two-thirds of total capital deployed. Together, these dynamics reshape risk. Returns are increasingly driven by scale exposure rather than portfolio breadth, and downside scenarios are more correlated than in prior cycles.
This shift marks the end of venture as a dispersion engine for experimentation. It is now a scale-driven allocation system, where access to compute, infrastructure, and late-stage balance sheets determines competitive outcomes.
AI moves from experimentation to industrial deployment
The most visible force behind this concentration is AI. Funding into AI companies reached approximately $226 billion in 2025, spanning foundation models, physical AI, robotics, and the infrastructure layers that support them. What matters is not simply the volume, but the maturity profile of funded companies.
CB Insights’ Commercial Maturity framework highlights a decisive transition. Capital is flowing away from exploratory experimentation and toward companies classified as “deploying” — firms with production-ready systems, clear customer adoption, and the ability to integrate into enterprise and industrial workflows. Innovation is being benchmarked less by novelty and more by operational readiness.
This helps explain why Large Vision Models and Vision-Language Models rank among the highest-momentum markets by Mosaic score. These are not abstract research plays. They sit at the intersection of software, hardware, and physical environments, with clear deployment paths in manufacturing, logistics, healthcare, and defense-adjacent sectors.
For innovation portfolios, this reframes the question. The challenge is no longer whether to engage with AI, but which layers of the stack offer durable strategic control, and which depend on external infrastructure that may become constrained or politicized.
Infrastructure becomes the primary bottleneck
One of the most underappreciated signals in the report is the rise of infrastructure liquidity. Hardware, compute, hosting, and energy-adjacent platforms are attracting capital — and exits — at scales that increasingly rival application-layer companies.
Large funding rounds for infrastructure providers such as Lambda and Crusoe, alongside landmark acquisitions like the $40 billion Aligned transaction, illustrate where value is accumulating. These assets sit upstream of software innovation. They control access to compute, power, and physical capacity, turning infrastructure into a strategic chokepoint.
This has two consequences. First, application-layer companies face rising dependency risks, as access to compute becomes both expensive and geographically concentrated. Second, infrastructure providers gain disproportionate bargaining power, shaping pricing, availability, and long-term innovation trajectories.
From a strategic perspective, infrastructure is no longer a background consideration. It is a primary constraint that shapes which innovation strategies are feasible at scale.
Geography reasserts itself as a strategic variable
The geographic distribution of capital reinforces these dynamics. Approximately 70% of global venture funding flowed to the United States in 2025, totaling around $328 billion. Europe and Asia posted modest growth — roughly $68 billion and $53 billion respectively — but not enough to materially rebalance global exposure.
This concentration amplifies geopolitical sensitivity. Policy shifts, regulatory decisions, or capital market disruptions in the US now have outsized effects on global innovation pipelines. At the same time, regional exit dynamics are diverging. While US markets remain dominated by large private rounds and selective M&A, parts of Asia — particularly India and China — are showing stronger IPO activity for technology companies.
This divergence matters for liquidity planning. Regions with active public markets offer different risk-return profiles and timing options than those dependent on private capital recycling. Innovation strategies that ignore these differences risk misaligning growth plans with realistic exit pathways.
Capital behavior signals a structural change
Perhaps the clearest signal that this is not a temporary phase comes from investor behavior. The report shows that top-performing “smart money” investors are no longer diversified across sectors. They are effectively running concentrated AI-focused portfolios.
This represents a structural change in capital formation. Specialization is replacing generalism, and narrative coherence around AI and infrastructure has become a prerequisite for accessing top-tier capital pools. For companies seeking strategic partnerships or long-term funding, this raises the bar. Alignment with dominant capital theses now matters as much as standalone performance.
It also increases systemic risk. When capital, talent, and infrastructure converge around a narrow set of themes, shocks propagate faster and further.
Human capital follows capital gravity
Talent dynamics reinforce these patterns. Headcount analysis in the report shows clear clustering around specific size bands correlated with commercial maturity. High-performing companies tend to scale teams in disciplined increments aligned with deployment readiness, rather than rapid, speculative hiring.
As capital concentrates geographically and technologically, talent follows. AI hubs attract disproportionate shares of technical labor, deepening regional imbalances and raising retention challenges elsewhere. For organizations building innovation capacity, this creates a dual constraint: access to skilled talent and the ability to retain it amid global competition.
The emerging risk profile
The cumulative effect of these shifts is a new risk landscape. AI concentration risk is high. Infrastructure dependency risk is rising. Geographic exposure is increasingly asymmetric. Valuations at the top of the market assume sustained dominance and uninterrupted access to compute and capital.
These are not reasons for retrenchment. They are signals that resilience must be designed into innovation strategies from the outset. Optionality is migrating from startups to capital providers and infrastructure owners. Strategic autonomy now depends on how these dependencies are managed.
Strategic implications
The State of Venture FY2025 reframes venture capital as a system operating under tighter constraints. Growth remains possible, but it is conditional. Innovation strategies must account for infrastructure bottlenecks, geopolitical exposure, and capital concentration simultaneously.
The era of broad-based diversification has given way to an era of selective dominance. Success increasingly depends on choosing where to place concentrated bets, understanding the dependencies those bets create, and building buffers against correlated risks.
For organizations navigating this environment, the priority is not speed, but structural clarity: clarity on where value accrues, where constraints sit, and how exposed current strategies are to forces outside direct control.
