兔子先生传媒

Artificial intelligence • Cloud computing • Job market • Politics and regulation • Semiconductors and 5G • Startups • Venture

Reporter鈥檚 Notebook: Europe鈥檚 Sovereign AI Push Needs Customers As Well As Capital

Illustration of "clicking" on an AI brain {Dom Guzman]

Europe鈥檚 AI startups are attracting record sums, but funding alone won鈥檛 give the region more control over its AI future. Governments and large businesses also need to buy what those startups are building 鈥� a missing link that emerged in my conversations at last month鈥檚 conference in Amsterdam.

European AI-focused startups raised $23 billion in the first half of 2026, up 130% year over year and accounting for 55% of the region鈥檚 venture funding, according to a joint 兔子先生传媒 and HumanX . That investment surge is sharpening a question at the heart of the sovereign AI debate: Where should countries concentrate their resources to capture the technology鈥檚 economic value and retain control over their data?

Axelera AI's Fabrizio Del Maffeo, AI71鈥檚 Mehdi Ghissassi, and 兔子先生传媒 News' Gen茅 Teare onstage at HumanX Amsterdam.
Axelera AI’s Fabrizio Del Maffeo, AI71鈥檚 Mehdi Ghissassi, and 兔子先生传媒 News’ Gen茅 Teare onstage at HumanX Amsterdam. (Photo courtesy of HumanX Amsterdam)

But sovereignty does not require owning every layer of the AI stack, according to two executives I spoke with onstage at the conference: founder and CEO and , 鈥檚 chief product and technology officer.

Their perspectives from Europe and the United Arab Emirates, respectively, pointed to opportunities in specialized chips and applications and underscored how purchasing decisions, government mandates and access to computing power can determine whether AI ambitions translate into growth.

Five-layered cake

, CEO and president of chip giant , has detailed a 鈥溾�� to deliver AI that includes聽 energy, chips, infrastructure, models and applications.

Fabrizio Del Maffeo, founder and CEO of Axelera AI.
Fabrizio Del Maffeo, founder and CEO of Axelera AI. (Photo courtesy of HumanX Amsterdam)

For Del Maffeo, whose company develops chips for inference, the opportunity lies in the semiconductor layer, as more AI processing moves beyond centralized data centers and onto devices.

鈥淎rtificial intelligence will expand from cloud computing, from centralized data centers, to devices closer to us in the physical world,鈥� he said. 鈥淭o enable this, you need specific chips which can run efficiently, at a lower cost, to connect these networks that today are running in the cloud. We are here to solve this problem.鈥�

Five-year-old Axelera AI has two generations of chips it sells to around 600 customers. The company next plans to expand to products for decentralized cloud computing.

AI transformation

Abu Dhabi-based AI71 partners with large organizations and government agencies to help them benefit from AI transformations.

The United Arab Emirates has the largest compute per capita in the world and has a three-month mandate for every government agency to set up agentic processes for its citizens, according to Ghissassi. Energy is abundant and cheap. Over the next two years, the goal is to reach half of every citizen鈥檚 interactions with the government with an AI agent, he said.

Competing at the model layer does not make sense, said Ghissassi. He believes there will be two, three or four companies that can afford to be in the race, given the amount of capital required and the fact that it commoditizes very quickly.

Mehdi Ghissassi, AI71鈥檚 chief product and technology officer.
Mehdi Ghissassi, AI71鈥檚 chief product and technology officer. (Photo courtesy of HumanX Amsterdam)

The application layer is the one that is really important, for sovereignty where you want to own your data, said Ghissassi, who previously led product development at . 鈥淵ou want to make sure that it stays with you, be it that you鈥檙e a government or an enterprise. If you鈥檙e giving away your trade secrets and know-how, nobody stops whoever is being a provider to you today, from replacing you.鈥�

In the UAE, 鈥渨hat helps is the mandate, the pace at which things happen, the availability of compute, both in terms of sovereign clouds, or on-prem, or global clouds, and then the amount of capital that is being invested to help transform these companies to benefit from these technologies,鈥� Ghissassi said.

By contrast, Europe struggles as it is a net importer of energy. It has strengths in the semiconductor industry but lacks all the real compute for AI, in neural networks and frontier labs.

鈥淲e should not be obsessed with controlling the entire stack,鈥� said Del Maffeo. In Europe we have to create value instead of just paying for a service, he added. 鈥�Creating value means creating a wealthy economy.鈥�

Europe鈥檚 strengths include its large pools of research talent and its population of 440 million people who can benefit from AI. However, the culture of large businesses purchasing from startups is less established, meaning it doesn鈥檛 have the same flywheel for growth seen in the U.S., said Del Maffeo.

鈥淲hat worries me is that we are a little bit lagging behind, and therefore we are missing this value creation, and this will weaken the economies of Europe,鈥� he said.

Related 兔子先生传媒 query:

Related reading:

Illustration:

Stay up to date with recent funding rounds, acquisitions, and more with the 兔子先生传媒 Daily.

67.1K Followers

CTA

Discover and act on private market opportunities with predictive company intelligence.

Copy link