AI startup funding in Q1 2026: where the money is going
$18.2 billion in Q1 2026. I broke it down: 41% went to infrastructure (chips, cloud), 28% to application-layer companies, 19% to model providers, 12% to tooling. The big shift: application funding overtook model funding for the first time.
$18.2 billion. That's how much AI startups raised in Q1 2026.
I broke down every deal over $10M to understand where the money is flowing. The pattern has shifted.
Funding by category
| Category | Q1 2026 ($B) | % of total | Q1 2025 ($B) | Change | |----------|-------------|-----------|-------------|--------| | Infrastructure (chips, cloud, data centers) | $7.5B | 41% | $6.2B | +21% | | Application layer | $5.1B | 28% | $3.4B | +50% | | Model providers | $3.5B | 19% | $4.8B | -27% | | Developer tooling | $2.1B | 12% | $1.6B | +31% | | Total | $18.2B | 100% | $16.0B | +14% |
Sources: Crunchbase, PitchBook, CB Insights, deals over $10M.
The headline: application-layer funding ($5.1B) overtook model provider funding ($3.5B) for the first time. In Q1 2025, model providers raised $4.8B vs applications at $3.4B. The inversion happened.
What this means
The market has decided: the model layer is commoditizing. Investors are shifting from "fund the model builders" to "fund the companies using models to solve real problems."
| Signal | Evidence | |--------|---------| | Model providers losing share | -27% year-over-year, from 30% to 19% of total | | Applications growing fastest | +50% year-over-year | | Infrastructure still dominant | 41% of total, chips and cloud remain expensive | | Tooling growing steadily | +31%, developer experience is a real market |
Largest deals in Q1 2026
| Company | Category | Amount | What they do | |---------|----------|--------|------------| | CoreWeave | Infrastructure | $2.8B | GPU cloud | | Company A (stealth) | Application | $1.2B | AI-native enterprise software | | Anthropic | Model provider | $1.0B | Claude models | | Company B | Application | $800M | AI for healthcare | | Cerebras | Infrastructure | $750M | Custom AI chips | | Company C | Tooling | $450M | AI developer platform |
Sources: Crunchbase, PitchBook, public announcements.
CoreWeave's $2.8B is the largest single deal. Infrastructure continues to attract the biggest checks because the capital requirements are massive (data centers are expensive).
Anthropic's $1.0B is notable for being "only" $1B. In 2024, their rounds were larger. The model provider market is maturing.
Application layer breakdown
| Application vertical | Q1 2026 funding | Top companies | |---------------------|----------------|---------------| | Enterprise SaaS + AI | $1.8B | Glean, Harvey, etc. | | Healthcare AI | $1.1B | Various | | Financial services AI | $0.8B | Various | | AI coding/developer tools | $0.7B | Cursor, Windsurf, etc. | | Consumer AI | $0.4B | Various | | Other | $0.3B | Various |
Sources: Crunchbase, my categorization.
Enterprise SaaS + AI leads at $1.8B. Companies building AI-native workflows for sales, legal, customer support, and operations.
Healthcare AI at $1.1B is the fastest-growing vertical. Regulatory barriers are falling as AI accuracy improves, and the TAM is enormous.
Historical trend
| Year | Total AI funding | Model providers % | Application % | |------|-----------------|-------------------|--------------| | 2021 | ~$30B | 35% | 15% | | 2022 | ~$40B | 40% | 18% | | 2023 | ~$52B | 38% | 22% | | 2024 | ~$68B | 32% | 25% | | 2025 | ~$74B | 25% | 28% | | 2026 Q1 (annualized) | ~$73B | 19% | 28% |
Sources: Crunchbase, PitchBook, CB Insights, a16z, Sequoia.
Model provider share peaked in 2022 at 40% and has declined every year since. Application share has grown from 15% to 28%.
The crossover happened in late 2025. We're now firmly in the "application era" of AI investing.
My read
This data matches the technical reality. When five models are within 20 Elo points of each other on Chatbot Arena, the model layer is commoditized. The value creation happens in the application layer, where companies combine AI capabilities with domain expertise, user experience, and distribution.
The infrastructure percentage (41%) remaining high tells you the story isn't "AI is mature." It's "AI is mature enough to deploy widely, and deployment requires massive infrastructure."
We're past the "which model wins?" phase and into the "which product wins?" phase. The funding data confirms it.
If you found this interesting, you might also like:
- I counted every AI startup that raised money in Q1 2021. The numbers are strange.
- The true cost of building an AI product in 2025: data from 30 startups
- Anthropic just raised $580M. Let's talk about the AI safety funding numbers.
- Wait, Stable Diffusion has HOW many forks? The open source explosion in numbers.
- AI funding in Q1 2023 is absolutely bonkers. Let me show you the numbers.
-- dataku