























































































































































































































































































































































































Programs for AI Startups Programs for AI Startups
How to design a program and its mentoring for startups building with AI.
What this presentation covers
What this presentation covers
What this presentation covers
Where the money goes
Capital is moving fast into AI
Private investment in AI grew 44.5% in 2024, led by the United States.
- United States $109.1B$109.1B
- China $9.3B$9.3B
- United Kingdom $4.5B$4.5B
Teach, do not just host
Teach founders, do not only host them
At Start-Up Chile, funding and desks alone showed no effect on performance.
- Funding and desks no effect found
- Plus training higher performance
How accelerators help
Broad, intensive and paced advice
Top accelerators sped ventures up with many advisers, frequent sessions and a fixed pace.
- Broad mentors, customers, alumni, peers
- Intensive many sessions in a short time
- Paced one calendar for the cohort
The mentor pool
A mentor pool built for AI
Pair technical depth with product, data, regulation and sales experience.
- Machine learning models and evaluation
- Data sources, rights, quality
- Regulation AI Act, privacy
- Go to market pricing, first buyers
Compute and data
Line up compute, data and pilots
Founders struggle to buy these alone, so the program brokers them.
The EU AI Act
What the AI Act offers start-ups
Article 62 asks member states to support SMEs and start-ups in four ways.
- Priority access to regulatory sandboxes
- Training tailored to start-ups
- Channels for advice and questions
- A seat in standard setting
Responsible AI
Build risk management in early
The NIST framework organises AI risk work into four functions.
- Govern roles and policies
- Map context and impact
- Measure test and monitor
- Manage act on the risks
Measure progress
Track progress, not only funding
Follow evidence of demand and of a model that holds up in use.
- Pilots signed pilots per startup
- Revenue paying customers
- Model quality agreed test results
- Runway months of cash left
Sources
The research behind this presentation
- Hallen, B. L., Cohen, S. L. & Bingham, C. B. (2020) Do accelerators work? If so, how? Organization Science, 31(2), 378-414
- Gonzalez-Uribe, J. & Leatherbee, M. (2018) The effects of business accelerators on venture performance: Evidence from Start-Up Chile The Review of Financial Studies, 31(4), 1566-1603
- Stanford HAI (2025) The 2025 AI Index Report: Economy Stanford Institute for Human-Centered Artificial Intelligence, Stanford University
- European Parliament & Council of the European Union (2024) Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act), Article 62 Official Journal of the European Union, L series, 12 July 2024
- National Institute of Standards and Technology (2023) Artificial Intelligence Risk Management Framework (AI RMF 1.0), NIST AI 100-1 U.S. Department of Commerce, Gaithersburg
Design for how AI startups grow
Teach, pace the advice and prepare founders for regulation early.