Web Summit 2026 opened in Lisbon on November 9 with artificial intelligence agents and China’s growing technology presence setting the agenda. But amid demonstrations of what AI can do, founders and policymakers raised a more consequential question: who will benefit when software can increasingly act on people’s behalf?
The four-day conference, running through November 12, brought that question into discussions of government services, workplace practices and the prospects for new businesses competing online.
China’s Technology Ambitions Take the Stage
Web Summit founder and CEO Paddy Cosgrave argued that China would produce many of the next generation’s leading technology companies, citing AI, robotics, vehicles and energy.
“The superstars of the future are not coming from California; they’re coming from China,” he said, according to Euronews.
Cosgrave said more than 50 figures from Chinese technology companies and research laboratories would appear across 24 stages. He also highlighted a record presence of Portuguese startups. Both were descriptions of the event’s programme, while his prediction about China’s future leadership remained his assessment.
His example was European fintech Revolut, which made self-hosted Chinese GLM models the default for roughly 2,000 engineers while retaining access to other AI tools.
Portugal Frames AI Around Public Benefit
Lisbon Mayor Carlos Moedas urged attendees to look beyond technical achievement toward useful innovation. He emphasised diversity and collaboration between people with different expertise, including the translation of university science into products that can improve lives.
His example was CellmAbs, the Portuguese biotechnology venture co-founded by Nuno Prego Ramos and researcher Paula Videira. Its 2024 agreement with BioNTech involved rights to a preclinical antibody candidate and related technology, not a proven cancer cure or a confirmed sale of the entire company.
Gonçalo Matias, Portugal’s minister responsible for state reform, extended that practical focus to public administration. He described five priorities: simplifying procedures before digitising them, reducing licensing delays, improving healthcare resource allocation, building dependable digital infrastructure and developing an “agentic state.”
Under that proposed model, authorised AI agents could coordinate agencies when citizens move home, have children or start businesses, avoiding repeated forms and requests for the same information. Matias presented this as a direction for government reform, not a service already operating at that scale.
Manus Tests How Employees Work With Agents
In a conversation with Web Summit co-host Casey Lau, Manus AI co-founder and chief product officer Tao Zhang described handing a list of tasks to Manus, which organises the work, retrieves context from connected tools and prepares fuller instructions for colleagues.
He also detailed a change to engineering recruitment: applicants are asked to solve a problem with Manus over four hours, after which interviewers examine how they used the agent. The exercise assesses collaboration with AI, not simply conventional coding ability.
Zhang discussed Manus 2.0 and Cue, an agent app with its own email, phone number and user-controlled spending capabilities. Both products had been announced in September, before the summit.
Cheaper Software, Tougher Competition
A panel moderated by CNBC senior technology correspondent Arjun Kharpal examined the economic implications. Lovable co-founder and CEO Anton Osika argued that AI-assisted creation lowers the barriers separating an idea from a working product, while insisting that companies must show measurable benefits rather than rely on novelty.
Cloudflare co-founder and CEO Matthew Prince identified the counterweight: an agent choosing services for its user may prefer established businesses with more available information. A small company could therefore find software easier to develop but customer acquisition harder, even with a competitive product.
Prince also questioned whether heavily financed AI infrastructure providers could sustain their economics, anticipating consolidation among some operators.
Richard Socher, co-founder and CEO of Recursive and You.com, described research systems that generate ideas, implement experiments and learn from their results. His vision pointed toward faster scientific work, not a demonstrated system capable of unrestricted self-improvement.
As Web Summit continues through November 12, the opening’s central test is whether increasingly capable AI can deliver better services and viable new businesses without concentrating opportunity among companies already dominant.
