AI Startups Raised $12.8 Billion in July — Europe's Defense Tech Led the Charge

Sixty-four venture capital deals closed between June 14 and July 14, 2026, totaling $12.8 billion in disclosed funding. That's more capital in 30 days than most sectors see in a quarter—and it tells a story about where global VCs are placing their chips.

The headline is simple: AI is not slowing down. What's more interesting is what kind of AI, where it's being built, and the quiet shift happening in how capital flows through the venture ecosystem.

AI Takes 84% of Disclosed Capital

Technology and AI companies claimed $10.7 billion of the $12.8 billion deployed across the 64 deals analyzed. That leaves just $2.1 billion for everything else—healthcare, fintech, logistics, and the rest. The concentration is real.

But the deals themselves reveal fragmentation. Yes, there were mega-rounds like Helsing's $1.8 billion Series C for AI-powered defense technology, and Dream's $260 million for AI-native cybersecurity. But the median deal size tells a different story. More than half the deals announced no disclosed amount at all—a signal that smaller rounds are increasingly common as more capital flows through rolling funds and SAFEs rather than traditional tranches.

Big Rounds Are Back, But They're Strategic

Three things stood out in the deal pattern:

First, defensive tech is having a moment. Helsing, a German startup building AI for military and defense applications, raised $1.8B in a single round. That's not a venture capital story—that's a geopolitical one. European governments and NATO-aligned investors are betting on homegrown defense AI as strategic infrastructure. It's the kind of round that raises questions about how AI funding will evolve if geopolitical tensions continue to shape capital allocation.

Second, cybersecurity and AI-powered security tools are consolidating capital. Dream's $260 million round, the Israeli cybersecurity startup founded by former NSO Group CEO Shalev Hulio, mirrors a broader pattern: security layers around AI are attracting institutional capital at scale. As enterprises deploy more AI, they're also deploying security tooling—and VCs are funding both sides of that equation.

Third, smaller rounds are growing in frequency. Seed-stage deals and early-stage funding rounds make up the majority of the deal count, even though they represent a tiny fraction of capital. This is the long tail of venture—founders still raising $1-5M, struggling to get attention in a market obsessed with $100M+ outcomes.

The Undisclosed Majority

Here's a critical caveat: only 29 of the 64 deals disclosed their funding amount. The other 35 deals? Silent. Based on typical deal patterns, estimated undisclosed capital could push the true total above $28 billion for the period. That's not a bug in the analysis—it's a feature of modern venture. Rolling funds, secondary markets, and direct listings have fragmented the information we used to get from traditional rounds.

What we don't see affects how we interpret trends. The deals that make headlines (Helsing, Dream) skew our perception toward megadeals. But the real volume, by deal count, is happening in the $500K-$5M range, where founders are still grinding to prove product-market fit.

What Founders Should Watch

If you're building an AI company, three signals matter right now:

1. Vertical AI applications are attracting more capital than foundational models. Nobody is funding another GPT. But startups that apply AI to specific problems—defense, cybersecurity, logistics, healthcare—are still raising. The market is sorted by problem, not by model.

2. If you're in the US, you're competing against Europe and Asia for capital. Helsing is German. Dream is Israeli. TikTok's parent Bytedance is building AI infrastructure in China. VCs are increasingly geographic—betting on teams and policy environments, not just technology. Pick your geography intentionally.

3. Down rounds are becoming real. While mega-deals dominate headlines, the median Series A or B deal in this cohort reflects slower burn-rate expectations. Profitability and unit economics matter again in venture. Hype alone doesn't close capital.

What's Next

The July data suggests venture capital is bifurcating. On one end, a few massive AI bets ($500M+) in strategic domains: defense, security, infrastructure. On the other end, thousands of smaller rounds ($1-5M) in category-specific applications. The middle—the traditional $50-100M Series B—is less common than it was three years ago.

For LPs watching from the sidelines, the message is clear: AI funding isn't slowing. It's narrowing. Capital is flowing to defensible problem domains with regulatory moats (like defense tech) and to clear margin expansion plays (like security tooling). Generalist AI plays and consumer-facing AI startups are facing tougher fundraising, even as media coverage suggests otherwise.

The founders who thrive in this environment won't be the ones building the most capable models. They'll be the ones solving the most pressing problems for the customers willing to pay.