OxyCon 2026: Experts Warn the Open Web Is Closing, Just When It Matters Most

As the public web keeps disappearing behind rising walls, speakers at the industry’s flagship conference laid out what businesses, journalists, and researchers stand to lose, and what can still be done about it.

Vilnius, Lithuania — September 30, 2026: The question of whether the open web can survive the AI era took center stage at the latest edition of OxyCon 2026, the premier conference for the public web data collection industry. Organized by Oxylabs, a leading web intelligence collection platform, this year’s event was the biggest yet, with 18 speakers on stage and thousands more watching online from around the world.

Across ten sessions and two panel discussions, a single tension ran through the day: demand for public web data is rising fast, while access to that data is getting harder. A June 2026 global survey of 500 senior business decision-makers, conducted by Censuswide for Oxylabs, found that 68% expect their web data needs to grow over the coming year, even as more of the web closes itself off.

“The world needs us to speak louder for the open web,” said Vytautas Savickas, Group CEO of Oxylabs, who opened the conference by making the case for why the issue matters now more than ever.

That concern came into sharp focus during one of the day's most anticipated sessions, a panel moderated by Denas Grybauskas, Chief Governance and Strategy Officer at Oxylabs, on whether the web is closing and what it could mean for businesses, journalists, and researchers. Paul Bradshaw, who teaches data journalism at Birmingham City University and works as a consulting data journalist with the BBC Shared Data Unit, reframed the debate, arguing that it is less about whether the web is closing and more about which parts of it need to stay open. Carl Miller, founder of the Centre for the Analysis of Social Media at Demos and author of The Death of the Gods, described a field watching its access to data shrink just as its analytical tools have finally caught up, warning that “the data picture is becoming bleaker and bleaker, more distant.” Krishna Sood, who leads product legal, IP, and regulatory affairs at frontier AI developer Black Forest Labs, pointed to a widening asymmetry in who gets to access data, across both policy and technology.

If governance was one pillar of this year’s OxyCon, the other two were AI-native extraction and adaptive infrastructure, and here the mood was more hopeful. Rather than treating AI as a shortcut past the hard problems of web scraping, speakers kept returning to a more grounded view of where it actually helps.

Andrius Kūkšta, Tech Lead at Oxylabs, showed how an LLM can turn an unfamiliar website into a finished dataset in about half an hour, but only when scraping know-how is built into the model itself, the idea behind the company’s new AI-Guided Datasets solution. “It's not the LLM helping you, it's you helping it,” he said. Giedrius Šteimantas, Director of Scraping Engineering at Oxylabs, made a similar point from the production side, explaining that AI at Oxylabs works around the pipeline, keeping access reliable and helping parsers adapt, rather than touching extraction directly. As he put it, “good AI makes engineers happy, who in turn build great things.”

A recurring lesson was that the fastest, most reliable systems usually do the least unnecessary work. Etienne Ellie, Lead Developer at ScrapingBee, urged engineers to “reach for the browser last, not first,” while Emilis Strimaitis, Head of Innovation at Hostinger, described how deliberately giving his team's AI search agent less to work with made the product simpler and more effective. Yenny Cheung, VP of Engineering at Bluefish AI, argued that as coding becomes a commodity, planning is what determines quality, cautioning that “we can't throw away quality just to make way for speed.”

The two panels captured the day’s balance. One, on the so-called agentic web, landed on cautious optimism, with panelist Hocine Amrane reminding the room that “whatever you do, at the end of the day, there's a human who will step in.” The other, on the future of open data access, made clear these questions are not going away any time soon.

The sessions pointed to the same conclusion Oxylabs has been raising all year: open access to public web data is more valuable than ever in the age of AI, and keeping it open will take deliberate effort from everyone who relies on it.

On-demand videos of every OxyCon 2026 session are now available for free. Watch them on the official conference website to explore the full range of insights on AI-native scraping, adaptive infrastructure, and the future of the open web.

About Oxylabs

Established in 2015, Oxylabs is a web data infrastructure provider enabling companies to use web intelligence at any scale. Continued innovation, an extensive patent portfolio, and a strong focus on ethics have helped Oxylabs become a global leader in the web intelligence industry. In 2026, Oxylabs secured its first outside investment from Warburg Pincus at a $3.6 billion valuation, becoming the second bootstrapped unicorn in the Tesonet portfolio. For more information, please visit: https://oxylabs.io/

Media Contacts

Kristina Sprindyte
Email: press@oxylabs.io
Website: https://oxylabs.io
Phone: +370 655 34419