TL;DR
- $510 billion flowed into startups globally in H1 2026 — more than all of 2025 ($440B) combined, and $135B above the previous half-year record (H2 2021: $375B)
- 70%+ of Q2 capital went to AI-focused companies, up from ~50% a year earlier
- OpenAI and Anthropic alone raised $217B — 43% of all global startup investment in H1 2026
- 7 of 16 billion-dollar rounds in Q2 went to frontier AI labs (OpenAI, Anthropic, DeepSeek, Moonshot AI, StepFun)
- Q1 2026 set a quarterly record at $305B; Q2 2026 was the second-largest quarter ever at $205B
- AI concentration is accelerating: everything else competes for a shrinking fraction of a growing pool
Introduction: The $510B Quarter-Half
Crunchbase published its H1 2026 Global Venture Report on July 2, 2026. The headline number — $510 billion in six months — reset every benchmark in venture capital history. What used to take twelve months now takes six.
(Source: Awesome Agents — AI Took 70% of Record $510B Venture Haul in H1)
The scale becomes clear when stacked against recent history:
| Period | Total Raised | Notable Context |
|---|---|---|
| H2 2021 | $375B | Previous half-year record (ZIRP peak) |
| Full Year 2025 | $440B | All VC deployed in all of 2025 |
| Q1 2026 | $305B | Largest quarter in VC history |
| Q2 2026 | $205B | Second-largest quarter ever |
| H1 2026 | $510B | New half-year record — topped 2021 by $135B |
AI's Lock on Capital
The concentration of venture dollars into AI is the defining structural feature of the 2026 market.
Q2 2026: 70%+ to AI
Over 70% of Q2 global startup capital went to AI-focused companies, Crunchbase data shows — up from roughly 50% in Q2 2025. In Q1 2026, the share was even higher: $242 billion of the $305 billion total (80%) went to AI.
This means the AI sector is growing faster than the overall venture market. The non-AI startup is competing for a shrinking fraction of a growing pool. In Q1 2026, non-AI startups shared just $63 billion — less than what Anthropic raised in its Series H alone ($65 billion).
The Megadeal Tier: An AI-Exclusive Club
Seven of Q2's 16 billion-dollar rounds went to frontier AI labs:
| Company | Round | Amount |
|---|---|---|
| OpenAI | Series ? | $122B (Q1) |
| Anthropic | Series H | $65B (May) |
| DeepSeek | Series ? | Undisclosed (China) |
| Moonshot AI | Series ? | Undisclosed (China) |
| StepFun | Series ? | Undisclosed (China) |
| Waymo | Round | $16B (Feb) |
| Together AI | Series C | $800M (at $8.3B valuation) |
(Source: Crunchbase — H1 2026 Global Venture Report)
Two Labs, Half the Market
OpenAI and Anthropic together raised $217 billion in H1 2026 — 43% of all global startup investment. OpenAI's Q1 round (estimated at $122B) was the largest single startup round in history. Anthropic's $65B Series H in late May pushed its valuation to $965 billion, briefly making it the most valuable private company.
(Source: Awesome Agents — Anthropic $65B Series H at $965B Valuation)
This level of concentration raises structural questions: what happens when two companies absorb nearly half of all venture capital? The answer depends on whether their IPOs transform that capital into public market liquidity or concentrate risk in two mega-cap positions.
Geographic Concentration
AI venture dollars are overwhelmingly American. Roughly 88% of AI startup funding in H1 2026 ($319B) went to US companies. China's share grew but remains concentrated in domestic frontier labs (DeepSeek, Moonshot AI, StepFun). Europe's share of AI venture capital declined relative to the US for the third consecutive half-year.
(Source: TechRiseUps — Why ~88% of AI Dollars Go to American Startups)
Emerging Trends in Agent Funding
Beyond the frontier labs, several patterns emerge in Q2 2026 funding:
AI Infrastructure
- Together AI raised $800M Series C at $8.3B valuation — the largest AI infrastructure round of Q2
- Compute providers and inference optimization startups saw 3× year-over-year funding growth
- Microsoft launched its own AI deployment company with a $2.5B commitment (July 2)
- AWS committed $1B to embedded AI engineers (July 1)
(Source: TechCrunch — Microsoft AI Deployment Company $2.5B)
Agent Platforms
- OKX AI Agent Marketplace launched with 150M users at $25B valuation
- Perplexity raised $200M for Comet browser (June 2026)
- Workday Agent Passport launched at DevCon
- Multiple YC W26 startups building agent infrastructure
The FDE (Forward-Deployed Engineer) Model
A new funding category emerged in Q2: companies that embed AI engineers directly into enterprise teams. AWS's $1B commitment and Microsoft's $2.5B deployment company are the largest examples, but dozens of smaller FDE startups raised rounds in Q2, reflecting enterprise demand for AI deployment services that pure SaaS cannot satisfy.
What Q3 2026 Looks Like
Several factors suggest Q3 will maintain or exceed Q2's pace:
- Anthropic IPO preparation — the S-1 confidential filing signals a public offering that would be the largest tech IPO of 2026
- OpenAI's funding trajectory — with revenue targets slipping (WSJ reported missed targets in June), OpenAI may need another round before its own IPO
- YC W26 companies graduating to Series A — 180+ startups from the Winter 2026 batch will raise follow-on rounds, many in AI agents
- Google's Gemini 3.5 Pro launch — the delayed flagship launch could trigger a competitive response from investors in competing model ecosystems
FAQ
Q: Is the AI funding bubble sustainable?
A: Valuation-to-revenue multiples for frontier labs are historically unprecedented. OpenAI's $300B+ valuation against revenue that missed targets in Q2 2026 (per WSJ) suggests froth. But the underlying revenue growth (OpenAI generating billions in API revenue, Anthropic with enterprise contracts across government and finance) exceeds what SaaS companies at similar valuations generated in 2021. The froth is real but may deflate through growth rather than crash.
Q: What does this mean for non-AI startups?
A: The non-AI startup is competing for approximately 20% of a market that grew 3× in two years. In absolute terms, non-AI venture funding is still higher than any pre-2024 year. The squeeze is relative, not absolute.
Q: Which AI sub-sectors are seeing the most funding growth?
A: Agent infrastructure (platforms, marketplaces, deployment services) showed the highest growth rate in Q2. Compute and inference optimization grew 3× YoY. Frontier pre-training (models) remains the largest absolute category but its share is stabilizing.
Further Reading
- Crunchbase — H1 2026 Global Venture Report
- Awesome Agents — AI Took 70% of Record $510B Venture Haul in H1
- Three AI Giants Ate 83% of February's $189B VC
- TechRiseUps — Why ~88% of AI Dollars Go to American Startups
- OpenAI Misses Revenue Targets — WSJ
- Together AI $800M Series C
Cet article a été initialement publié sur The Agent Report.
Top comments (0)