YC Demo Day Investment Pattern Analysis: W24 → S24 → W25 → S25 → W26
Analysis date: 2026-07-24 Purpose: Analyze which categories attract the most investment after Demo Day, to reverse-engineer where YC is pushing
1. Top Funding Categories by Batch
W24 (Winter 2024) — 260 companies
| Rank | Category | Share | Key Trend |
|---|---|---|---|
| 1 | AI / Generative AI | ~33% (86 companies) | 2x vs W23, 3x vs W21. AI infrastructure, data curation, AI scribes |
| 2 | Cross-border Fintech | ~8% (30 companies) | International payments, currency exchange, B2B remittance. Targeting India, Southeast Asia, Africa |
| 3 | GovTech / Legal AI | Growing | Government contract automation, compliance |
| 4 | Climate / Weather AI | Small number | AI-based weather forecasting, precision agriculture |
Notable points:
- Fintech share plummeted from 24% (W22) to 8% (W24) — down to a third
- Number of LatAm founders dropped sharply to 1 (from a historical peak of 33, as focus shifted to AI)
- 85% of AI founders are US/Canada-based
S24 (Summer 2024) — Demo Day 1 & 2
| Rank | Category | Key Startups |
|---|---|---|
| 1 | AI Infrastructure | Hamming AI (AI voice agent testing), Simplex (synthetic datasets), Silurian (AI weather models) |
| 2 | Defense / Robotics | Theseus (anti-jamming drones), Azalea Robotics (airport baggage automation) |
| 3 | Space Tech | Lumen Orbit (space data centers), Spaceium (space refueling stations) |
| 4 | AI Semiconductors | Exa Laboratories (AI chip 28x more efficient than Nvidia H100) |
| 5 | Construction / Infrastructure AI | SchemeFlow (construction permitting automation, used by 400+ projects) |
Notable points:
- AI remains dominant, but the mix is shifting from “AI applications” to “AI infrastructure/hardware”
- Defense/robotics/space hard-tech share increased significantly
- Exa: raised $17M from Lightspeed, Nvidia, and YC
W25 (Winter 2025) — 160 companies
| Rank | Category | Key Startups |
|---|---|---|
| 1 | AI Agent Infrastructure | Abundant (agent teleoperation), Browser Use (web-navigating AI tool) |
| 2 | Defense Tech | Splash (autonomous patrol boats, 800-mile voyages) |
| 3 | Robotics / AgTech | Red Barn Robotics (“Roomba for weeds,” signed $5M LOI) |
| 4 | EdTech AI | GradeWiz (AI grading automation) |
| 5 | AI Developer Tools | Nextbyte (AI coding interview evaluation) |
Notable points:
- “Tools for AI agents” emerged as a new category
- Investment concentrated on agent infrastructure rather than standalone agents
- Strengthening of autonomous defense systems (drones, boats)
S25 (Summer 2025) — 160+ companies
| Rank | Category | Key Startups | Funding/Valuation |
|---|---|---|---|
| 1 | AI Agent Infrastructure | Dedalus Labs (“Vercel for AI agents”), Autumn (“Stripe for AI startups”) | High investor demand |
| 2 | AI Developer Tools | Keystone (AI fixes production bugs, turned down a 7-figure acquisition offer) | 20-year-old founder |
| 3 | Defense Tech | Perseus Defense (counter-drone mini-missiles) | Requested for DoD pilot |
| 4 | Fintech / AI Infrastructure | Autumn (billing/payments infrastructure for AI apps) | Used by 40 YC startups |
| 5 | EdTech AI | Pingo AI (AI language conversation tutor) | $250K/month revenue, 70% monthly growth |
| 6 | Insurtech AI | Solva | $245K ARR just 10 weeks after launch |
Notable points:
- “Tools for the AI economy” is the core theme
- AI agents are evolving into AI agent infrastructure
- GetASAP Asia: led by General Catalyst, the highest valuation in the batch
W26 (Winter 2026) — ~190 companies
| Rank | Category | Key Startups |
|---|---|---|
| 1 | AI Infrastructure / Security | Silmaril (AI agent prompt injection defense), Crosslayer Labs (AI scam detection) |
| 2 | Defense Tech | Milliray (small drone tracking radar) |
| 3 | Robotics / Data | Asimov (training humanoid robots using human motion data) |
| 4 | AI Developer Tools | Sonarly (monitoring alert noise reduction), Superset (running 100+ coding agents concurrently) |
| 5 | AI Wearables | Button Computer (voice-command app control wearable, founded by ex-Apple team) |
| 6 | Energy / Climate | Terranox AI (AI-driven uranium deposit exploration, supporting nuclear power) |
S26 (Spring 2026) — Latest batch
| Rank | Category | Key Startups | Funding/Valuation |
|---|---|---|---|
| 1 | Defense Tech | 9 Mothers (counter-drone, $1.6M revenue, $35M contract expansion) | $200M+ valuation |
| 2 | AI Developer Tools | Ploy (automated website building/marketing) | $27M seed (First Round, YC led) |
| 3 | AI Agents | Tasklet (natural-language workflow agents), Arga Labs (AI agent digital twins) | High investor demand |
| 4 | AI Security | Silmaril (AI agent security infrastructure) | Multiple startups at $175M+ valuations |
| 5 | Space Tech | Dispatch (return satellites for space-manufactured products) | — |
Notable points:
- At least 2 startups reached $175M+ valuations
- One startup reached $200M+ — among the highest in YC history
- Repeat founders command a valuation premium
2. Investor Sentiment Analysis
Categories Investors Are Excited About (ranked)
- AI Agent Infrastructure — #1 for three consecutive batches. Top premium goes to “agents that build agents”
- Defense Tech — Rose sharply starting W25, reached $200M+ valuations by S26. DoD contracts are the key traction signal
- AI Developer Tools — Steady demand every batch. Coding agents, bug fixes, test automation
- AI Security — An emerging category. Prompt injection defense, AI scam detection, etc.
- Robotics / Autonomous Systems — Hardware + AI combination. Agriculture, logistics, defense
Categories Investors Are Avoiding
- Traditional Fintech — Share plummeted from 24% to 8%. Cross-border is the exception
- Consumer Apps (general) — Pure consumer apps without AI see declining interest
- Web3 / Blockchain — “Noticeably absent” since S24
- General SaaS — Non-AI-native SaaS sees declining investment appeal
Shifts in Funding Structure
| Item | 2020-2021 (Bull Market) | 2024 (W24) | 2025-2026 |
|---|---|---|---|
| Seed round size | $16-27M | $1.5-2M (tiny seed) | $3-27M (polarized) |
| Post-money valuation | $75-150M | ~$15M | $175-200M+ (top startups) |
| Investor sentiment | “Invest no matter what” | “Avoid small rounds” | “Premium for proven traction” |
| Lead investor | Required | Angel-centric, leads avoided | Top startups get major leads like First Round |
Key insight:
- Mid-tier startups: $1.5-2M tiny seeds → investors avoid (can’t hit their 10% ownership target)
- Top startups: $27M+ seeds, $175M+ valuations possible
- Polarization is intensifying: heaven if you have proven traction, hell if you don’t
3. YC’s Strategic Direction (What It’s Pushing)
YC Request for Startups (Official)
- Artificial Intelligence (AI) — The foundation of every category
- Spatial Computing
- Climate Tech
- Bringing Manufacturing Back to America
- Defense Technology — “New defense technology rooted in U.S. military R&D”
Key Statements from YC Partners
Garry Tan (YC CEO):
- “The San Francisco Bay Area is the optimal location for AI success”
- Expanded from 2 to 4 batches per year (Spring and Fall added starting 2025)
- Shifted Demo Day to an in-person event (inviting 1,500 investors)
- In-person eligibility: only decision-making investors who invested $50K+ in a YC company within the past 2 years
Ankit Gupta (YC Partner):
- Announced the “YC AI Stack” — partnerships with 24 AI tooling companies, offering $25K+ in free credits
- Positioning AI development infrastructure as the core of the YC ecosystem
Where YC Is Actually Pushing (Data-Driven Reverse Engineering)
2022: Fintech + consumer apps + general SaaS
↓
2023: Early AI + fintech begins to shrink
↓
2024 (W24/S24): AI explosion + defense/robotics rise + fintech shrinks to 8%
↓
2025 (W25/S25): AI agent infrastructure + defense strengthens + hard tech
↓
2026 (W26/S26): AI security + defense at $200M+ + agent-economy tools
YC’s 3 core bets:
- The AI agent economy — the full stack of building, testing, securing, and billing agents
- Defense / national security tech — geopolitical necessity + high contract value + rapid growth
- AI + the physical world — robotics, autonomous systems, space, energy (not just software, but hardware too)
4. Implications for New Applicants
The Startup Profile YC Wants
| Element | Description |
|---|---|
| AI-native | A product where AI is a core function, not an add-on |
| Part of the agent economy | Building AI agents, or infrastructure/tools for agents |
| Fast traction | Revenue growth rate matters (Pingo AI: 70%/month, Solva: $245K ARR in 10 weeks) |
| Defense / hard tech | Potential for government contracts, solving physical-world problems |
| US-based | Bay Area is optimal. Remote is possible but the center of gravity is the US |
| Repeat founders | S26 saw repeat founders command $175M+ valuations |
The Startup Profile YC Avoids
| Element | Description |
|---|---|
| General fintech | Share has plummeted unless it’s cross-border |
| Web3 / blockchain | Nearly disappeared |
| SaaS without AI | Interest declines without an “AI-powered” label |
| Non-US-based (outside AI) | Region-specific startups (LatAm, Southeast Asia, etc.) are shrinking |
| Slow growth | Cases where even a tiny seed can’t find investors |
Application Strategy Recommendations
- AI agent infrastructure is the hottest category — target one of: agent development, testing, security, deployment, or billing
- Defense tech can command high valuations — but requires government-contract traction
- Revenue traction is key — revenue growth rate is the biggest lever at Demo Day
- Being Bay Area-based is advantageous — YC explicitly emphasizes this
- “AI + X” positioning — AI plus a specific domain (healthcare, construction, agriculture, defense) differentiates better than pure AI
- Combine hardware + AI — robotics, autonomous systems, wearables, etc. earn a premium when combined with the physical world
5. Tracking Funding Success Rates by Category
High Follow-on Funding Rate (within 6 months of Demo Day)
| Category | Example Funding | Notes |
|---|---|---|
| AI Infrastructure | Exa: $17M (Lightspeed, Nvidia) | Large seed/Series A right after seed |
| AI Developer Tools | Ploy: $27M seed (First Round) | Repeat-founder premium |
| Defense Tech | 9 Mothers: $200M+ valuation | $35M DoD contract expansion |
| Cross-border Fintech | Fintoc: $7M, Cleva: $120K/month revenue | Only niche cross-border plays succeed |
| AI Data | Lago: $22M, Datacurve | Data curation / synthetic data |
Low Follow-on Funding Rate
| Category | Phenomenon |
|---|---|
| General fintech | “Definitely shrinking” — declining investor interest |
| Web3 / blockchain | Nearly extinct |
| General consumer apps | Hard to fund without an AI angle |
| General SaaS | Often stalls at the tiny-seed stage ($1.5-2M) |
6. Key Conclusions
What YC Pushes = What Investors Fund
YC’s Request for Startups and actual Demo Day investment patterns are remarkably aligned:
- The full AI agent economy stack — the surest bet. Growing every batch
- Defense / national security tech — surged starting in late 2024, reached $200M+ valuations by 2026
- AI + the physical world — robotics, autonomous systems, space, energy
- AI security — an emerging category in 2026
Action Items for Applicants
- Clearly position how your product connects to AI agents
- Maximize revenue traction before Demo Day
- Confirm you’re relocatable to the Bay Area
- Explore a defense/hard-tech angle if it’s plausible for your product
- If you’re a repeat founder, emphasize that