Global Venture Funding Hits Record $510B in H1 2026

$510 billion in six months is not a recovery story. It is a concentration story. Crunchbase says global venture funding hit a record in H1 2026, already above the $440 billion invested in all of 2025, with OpenAI and Anthropic alone accounting for $217 billion, or 43% of the total. For founders and operators, the headline number is less useful than the allocation map: capital is moving, but not evenly.

The market is open, but only at the top

Crunchbase data shows H1 2026 set a new half-year record for startup investment, beating the prior H2 2021 peak of $375 billion. Q1 carried $305 billion. Q2 added $205 billion across more than 5,000 startups, making it the second-largest quarter on record after Q1.

That sounds broad. It is not.

The defining mechanic is concentration:

• OpenAI and Anthropic: $217 billion combined in H1.
• Share of total H1 funding: 43%.
• AI-focused companies: more than 70% of global startup capital in Q2.
• Billion-dollar rounds in Q2: 16 companies, totaling $108.6 billion.
• Share of Q2 funding in those rounds: 53%.

This is not a normal venture cycle where more liquidity lifts all stages. It is a capital stack forming around compute, foundation models, AI infrastructure, robotics, defense, and healthcare. The spreadsheet says the median founder should not read “record funding” as “easier funding.” They should read it as “investors have a narrower mandate and larger checks for fewer companies.”

Liquidity is back, but underwriting has changed

The useful part of the Crunchbase report is not just funding volume. It says IPOs and acquisitions accelerated in Q2, creating the strongest exit market since the 2021 boom. The same report says the largest IPO ever for a venture-backed company and the largest startup acquisition ever both occurred in Q2, both involving SpaceX: a public listing at a $1.77 trillion value with $75 billion raised, followed by a stated intent to acquire Anysphere, maker of Cursor, for $60 billion.

Treat that as a market signal, not a template. These are not comps most companies can use. They are outliers large enough to distort the dataset.

For boards, the practical move is simple: rebuild the financing plan around two separate markets.

One market exists for companies that can credibly sit inside AI infrastructure, frontier AI, defense, robotics, healthcare, or adjacent hard-tech budgets. That market has large checks and strategic urgency.

The other market is everything else. It may benefit from better exit conditions, but it will not automatically inherit frontier-AI multiples. A generic “AI-enabled” slide will not clear diligence if gross margin, retention, deployment cost, and customer concentration do not support it.

Europe and LatAm show the same filter

The pattern is not only U.S.-led, though the U.S. remains dominant. Crunchbase says two-thirds of Q2 startup capital went to U.S.-based companies, down from 83% in Q1 and in line with Q2 2025 proportions.

Europe is showing the same narrowing. EU-Startups reports that European VC fundraising in 2026 is concentrated in specialist funds targeting DeepTech, DefenceTech, AI, FinTech, quantum, BioTech, climate technologies, and early-stage software infrastructure. The ten largest identified vehicles account for about €3.29 billion in capital. Kembara leads that list with a reported €750 million first close for a planned €1 billion DeepTech fund, aimed at growth-stage European DeepTech companies, including Series B and Series C. E2D, created by Earlybird and AVP, reported a €500 million fund focused on Europe’s DefenceTech and dual-use scaling gap.

LatAm has a smaller data point in this pack, but it points in the same direction: FinTech Global reports LatAm FinTech investment hit a five-quarter high in Q1 2026 amid investor optimism. No further deal-level detail is confirmed here, so the responsible read is limited: regional capital appetite appears to be improving in at least one vertical, not across the whole startup market.

The operator checklist is blunt:

• If raising now, benchmark against the capital lane you actually occupy, not the global headline.
• If pitching AI, show where the cost curve improves or where revenue quality changes. Labeling is not enough.
• If planning an exit, separate strategic scarcity value from ordinary SaaS multiples.
• If sitting outside the favored sectors, extend runway before assuming the record market reaches you.

Verdict: venture funding is liquid again. It is not democratic.