At Rise of AI 2026 in Berlin, Fabian Westerheide, co-founder of Rise of AI and founding partner at AI.Fund, opens the conference's 10th anniversary edition with a retrospective on a decade of AI. He argues 2016 was the industry's real turning point, when deep learning research (DeepMind, reemerging neural networks running on Nvidia chips) took AI from a niche topic to a political one, with China's rise and OpenAI's founding both dating from that period. He traces a shift in the conference's own recurring themes from discussing AI's potential to discussing implementation, scaling and, increasingly, technological sovereignty. Westerheide names lessons learned over ten years: transformation inside companies and governments moves slower than expected, regulation keeps expanding, and capital availability, more than talent or ecosystems, decides which regions' AI companies scale, with Europe lagging the US and China. He says jobs are shifting rather than disappearing and that AI agents are becoming core infrastructure. He closes with a call to start from concrete problems rather than technology, and not worry about being first. A short on-stage interview follows about the event's history and its founders.
Key takeaways
- Westerheide dates AI's shift from niche research topic to political issue to 2016, citing DeepMind, Nvidia hardware and China's AI investment.
- The conference's core theme has moved from discussing AI's potential toward implementation, scaling and now technological sovereignty.
- Westerheide says capital availability, not ecosystems or talent, decides which regions' AI companies scale, and Europe underinvests.
- He argues AI is turning work reorganization rather than elimination, with small teams increasingly managing many AI agents.
- His advice is to start from a concrete business problem and pilot a solution rather than starting from the technology itself.
Detailed notes
Ten years of Rise of AI
Westerheide opens by framing Rise of AI's 10th anniversary as both a business and a personal milestone, comparing the conference to his marriage and calling it 'our first baby' with his wife Veronica. He says the event started with no business plan and no investors, built on the belief that it mattered to bring the right people into one room once a year, and that it grew into a family-run business, something he calls 'very German' about family-driven entrepreneurship.
2016 as the turning point
Westerheide argues AI was a niche topic before 2016, discussed mainly at small machine-learning meetups. He identifies 2016 as the year that changed everything: DeepMind and the reemergence of neural networks, made possible by running older 1980s algorithms on Nvidia chips, turned AI from a research curiosity into a political issue. He says China's rise from an irrelevant AI player to a leading one, backed by an announced 180 billion investment, dates from this period, as does the founding of OpenAI as a response to it. He says companies that took AI seriously in 2016 are on stronger ground today than those that started later.
From potential to deployment, then to sovereignty
Reviewing ten years of conference agendas, Westerheide says recurring themes such as ethics, bias, jobs and systems have stayed constant, but the framing has shifted: early discussions centered on AI's potential and 'the singularity is close'; attention then dropped during COVID even as the technology kept improving; and today's central questions are about implementation, scaling, and, increasingly, sovereignty, a topic he says was not discussed ten years ago. He says despite the hype, adoption remains a fraction of AI's potential, and that the pace of model and tool change now outstrips most people's ability to track it.
Lessons: patience, capital, and pride
Westerheide lists lessons from ten years of the conference. Transformation inside companies, industries and governments takes longer than expected, and regulation tends to expand alongside innovation rather than ease. Capital, more than ecosystems or talent, determines which companies and regions scale; he says Europe has capital but keeps it in retirement funds rather than investing it in the future, while the US and China pour money into AI infrastructure and research. He also argues Europe should let go of the need to reinvent technology from scratch, pointing to Chinese models like DeepSeek as proof that iterating cheaply on existing approaches can succeed.
Jobs, agents, and AI as infrastructure
Westerheide says the fear of AI replacing jobs, a recurring theme since the conference's first years, is now visible in labor-market shifts, but argues work is being reorganized rather than eliminated, with new roles emerging and companies increasingly running hybrid teams of a small number of people managing many AI agents. He recalls describing brain-machine interfaces and an 'internet of agents' ten years ago; the technology behind agents, he says, now largely exists under different names, and AI is becoming infrastructure that every company depends on to stay competitive, not just a tool.
Closing call to action and Q&A on the event's history
Westerheide closes by urging attendees and viewers not to try to solve AI holistically but to find a specific problem, start with a pilot, and let the problem drive the choice of technology rather than the reverse. He says being first matters less than staying consistent, citing the US and China as proof one does not need to be first to do good work. In a following on-stage conversation with Daniel, managing director of a partner organization, Westerheide credits his wife Veronica with the event's early logistics, including bartering the first venue near the Deutsches Technikmuseum, and describes a printed 10-year 'chronicle' and manifesto distributed at the event, alongside contributed greetings from long-time attendees.
Quotes
Capital decides who scales.
Fabian Westerheide
Everything takes slower than we want and think.
Fabian Westerheide
It's not about being first.
Fabian Westerheide
Questions answered in this talk
Why does Fabian Westerheide consider 2016 the turning point for AI?
According to Fabian Westerheide at Rise of AI 2026, 2016 was when deep learning research such as DeepMind, running on Nvidia chips, moved AI from a niche research topic into a political issue, coinciding with China's rise as a major AI investor and OpenAI's founding.
What lessons has Fabian Westerheide learned in ten years of running Rise of AI?
According to Fabian Westerheide at Rise of AI 2026, transformation inside companies and governments takes longer than expected, regulation keeps expanding alongside innovation, and capital availability, more than talent or ecosystems, determines which regions' AI companies scale, with Europe currently lagging the US and China.
What does Fabian Westerheide say about AI and jobs at Rise of AI 2026?
According to Fabian Westerheide at Rise of AI 2026, jobs are shifting and new roles are emerging rather than work disappearing outright, and companies are moving toward hybrid organizations where small teams manage large numbers of AI agents.