South Summit 2026 (Courtesy of South Summit)
South Summit, co-organized by IE University, is one of Europe's leading startup and technology gatherings. Held annually in Madrid, it brings together entrepreneurs, investors, universities, governments, and corporations in search of the next wave of innovation.
Japan Forward attended this year's edition to ask two questions: can Europe catch up in the AI race, and what can Japan learn from it?
The event ran from June 3 to 5 at La Nave under the theme "AI Convergence," reflecting how deeply artificial intelligence has spread across fields ranging from education and finance to national defense.
That widespread adoption also carries a geopolitical dimension. As AI becomes embedded in critical infrastructure, countries and companies can no longer rely entirely on foreign technology and platforms.
Europe, however, remains behind in frontier AI development. According to Stanford's AI Index, most leading model development remains concentrated in the United States and China, while US private investment in AI significantly exceeds that of China and Europe combined. Japan faces a similar problem: it has strong industrial technology and research capacity, but has yet to produce a globally dominant AI platform.
Can Europe Catch Up?
At the summit, Mikayel Vardanyan, co-founder of Picsart, was blunt about Europe's position. He said that it was "far behind" the US and China, particularly in developing large foundation models—general-purpose AI systems trained on vast datasets, such as ChatGPT.
However, he also argued that the opportunity is no longer limited to building such models, as they have "already commoditized."
Picsart itself is proof of this. Founded in Armenia, the company grew into a global digital creation platform, demonstrating that startups from smaller ecosystems can still scale internationally and compete on a global stage.
In his view, future opportunities lie in applying AI to specific problems—such as design tools, healthcare, and education. While a handful of large players may dominate foundation model development, he argued that much of the value in the AI era will come from companies focused on practical, real-world applications.
Do Rankings Even Matter?
Rafif Srour, Vice Dean of IE School of Science and Technology, framed the issue from an educational and institutional perspective.
She also argued that Europe's challenge should not be reduced to a simple ranking against Silicon Valley or Beijing.
"The question is not whether we've lost the race," Srour said. "The question is which race do we want to win?"
Developments in Europe suggest this is a valid framing. Its opportunity may be narrower but more targeted. Rather than trying to replicate the US or China in every segment, Europe is investing in public-interest computer infrastructure through initiatives such as AI Factories and future AI Gigafactories, alongside efforts like InvestAI to mobilize capital.
It also benefits from strong industrial foundations, including companies like ASML and a deep advanced manufacturing base, where AI can be embedded directly into production processes rather than developed solely as consumer-facing applications.
Just as importantly, Europe's emphasis on governance and trusted regulation could become a competitive advantage in high-liability sectors such as healthcare, finance, public services, and defense.
Avoid Fragmentation
Based on insights at the summit, Japan's lesson from Europe seems to be two-sided: learn from Europe's ambition for sovereignty, but avoid Europe's slow, fragmented execution.
The importance of policy centralization and clear decision-making authority was highlighted by Hian Goh, founding partner of Singapore-based venture capital firm Openspace Capital.
Goh pointed to Singapore's approach, where AI is treated as a whole-of-government issue rather than the responsibility of a single ministry. Given AI's wide-ranging impact across society, Singapore has established a central, council-like structure to bring stakeholders together and debate competing views.
However, Goh also cautioned that coordination should not turn into another layer of bureaucracy. Countries, he said, need "a centralized place where all views are heard," but once debate concludes, "you've got to move."
Secure Sovereignty
Another crucial ingredient in the AI race is securing sovereignty in semiconductors and compute infrastructure. Compute sovereignty spans three dimensions: supply chain control, ownership, and territorial jurisdiction, meaning hardware is located within a country so national laws apply.
Goh argued that AI has become an "existential force" for industry and society, making chips, which are essential components of AI systems, a potential single point of failure.
He further pointed out that Japan is already moving in this direction through the semiconductor manufacturer Rapidus, describing it as "the prime example" of what other regions, including Europe, should take seriously to secure silicon sovereignty.
In April 2026, the Japanese government approved an additional ¥631.5 billion ($3.95 billion) in support for Rapidus, bringing total public assistance to ¥2.354 trillion ($14.7 billion), with mass production targeted for fiscal 2027. Goh said this effort should be more explicitly connected to AI strategy.
Promote Wider Adoption
The third recommendation is to raise AI adoption in Japan's creative and business sectors. Mikayel Vardanyan, co-founder of Picsart, noted that Japan has long been one of the company's top markets, suggesting a strong existing appetite among Japanese users for digital creative tools.
He said generative AI should not be seen only as a threat to creative industries, as it can also give creators and companies new ways to produce content and experiment more quickly. His warning for Japan was blunt: if adoption remains low, "you won't be able to survive."
AI in Education: Proceed with Care
IE's Vice Dean Rafif Srour argued that higher education should be rebuilt around a "fundamentals plus applied AI" model. Under this approach, IE University limits the use of AI in foundational courses so that students first master the underlying logic before integrating AI into applied, multidisciplinary, industry-linked projects.
She warned that introducing AI too early can create a "false impression of knowledge," where students receive fluent answers without real understanding. This, she said, undermines the development of "intellectual or cognitive patience"—the ability to sit with difficult problems, clarify assumptions, and work through reasoning step by step.
She also suggested that Japan could benefit from more IE-style labs that bring together faculty, students, and industry partners to work on real-world challenges.
Create a 'Risk-Taking Zone'
Goh recommended that countries create protected spaces for risk-taking, but not by copying Silicon Valley. Instead, each society should build its own environments for experimentation that fit its own characteristics and culture.
He pointed to Singapore's Block 71 as one example, and Israel's broader national culture of risk-taking as another. Japan, he argued, does not need to become California, but it needs places, rules, capital, and institutions that allow people to experiment, where mistakes are treated as "setback, not failure."
RELATED:
Spain and Japan Seek a New Strategic Balance
Author: Daniel Manning
