Oxylabs AI Predictions 2026: Five-Year Track of Early Calls and What the Future Holds

Oxylabs' sixth edition of AI predictions identifies the next key steps in AI-driven market transformation, as seen by our experts. We released our first AI predictions for the year 2021, along with our forecast for the big data industry. 

We recognised that only the internet has enough data to meet the growing demands of AI training. Predicting what’s next for AI was also part of our job, given our efforts to integrate AI innovations into web scraping as quickly as this technology evolves.

Looking back at five years of forecasts, one can find early mentions and elaborations of many trends and considerations that soon became everybody’s talking points:

• The importance of multimodal data and its handling

• AI on both sides of the cybersecurity battle

• Increased regulatory scrutiny (quite similar to what we had in web scraping just a few years prior)

• AI-powered developer tools as the driver for going from hype to ROI

• The rise of AI agents and agentic infrastructures

Here is a glimpse into 2026, according to our experts, with a broader discussion below.

Top 10 AI predictions for 2026

1. The AI bubble narrative will persist, but investment will continue flowing in 2026, driven by a steady stream of real-world use cases (Julius Černiauskas, CEO)

2. AI agents will move from helpers to full workflow operators with auditability, permissions, and governance built in by default (Ali Chaudhry, AI/ML Board)

3. Web scraping will increasingly be handled by coordinated AI agents that automate tasks end-to-end, democratizing access to public data (Juras Juršėnas, COO).

4. AI-native browsers will challenge Chrome’s dominance by acting as autonomous research and task-completion agents (Rytis Ulys, Head of Data & AI).

5. If access to public data becomes too restricted in Europe, companies may either build biased AI models on limited datasets or look for loopholes (Denas Grybauskas, CGSO).

6. AI will increasingly be embedded into systems of economic and social control, rapidly reshaping power structures (Adi Andrei, AI/ML Board).

7. On-device AI will redefine privacy, decentralising data collection and creating new, largely invisible forms of data extraction (Rytis Ulys, Head of Data & AI).

8. Copyright and “transformative use” will dominate AI legal debates as the industry explores mechanisms for attribution and creator compensation (Denas Grybauskas, CGSO).

9. CAPTCHA wars will rage on, with AI being used to both create and solve CAPTCHAs in competitive battles for data access (Juras Juršėnas, COO).

10. The crucial breakthroughs will be in integrating AI into business operations rather than in model advancements (Julius Černiauskas, CEO).

AI and web scraping: predicting the common future

No one today could seriously dispute that public web data is the lifeline of the AI industry. As one of the pioneering companies in web intelligence gathering, Oxylabs was positioned to see the intertwined trajectory of the two industries before it appeared on most radars.

A crucial step in our AI journey was the launching of Oxylabs AI/ML Advisory Board in 2020. Five seasoned AI and ML experts with career paths going everywhere from NASA to leading tech companies to teaching at world-famous universities were recruited to help us step up a notch in AI and ML implementation for web scraping and the development of unprecedented solutions.

Our goal, together with our board members, was to understand where AI and web intelligence were headed. Out of that, a tradition was born to share our thoughts with the broader public at the turn of the year.

Did we see it coming?

Looking back, some could say that the AI boom was inevitable. Before the rollout of game-changing tools in 2022, however, what seems obvious now wasn’t so obvious or at least widely accepted yet.

Nevertheless, at the start of 2021, Gediminas Rickevičius, our Senior VP, predicted that investment in AI and ML would skyrocket over the next 12 months. The breakthroughs of 2022 are a testament to the accuracy of this forecast.

Once they saw the tools that these breakthroughs resulted in, members of our AI/ML advisory board recognised their potential immediately and correctly predicted their upcoming pervasiveness. Some of them already saw from the outset the rise in interest in tools that can handle different types of input, such as text, images, and audio.

Today, such multimodal tools are on everyone’s mind. As correctly predicted by Ali Chaudhry, a member of our AI/ML advisory board, 2025 was a big year for them, especially for text-to-video models.

We also saw the proliferation of AI agents and multi-agent systems coming this year. Juras Juršėnas, our COO, predicted that the next generation of AI-assisted developer tools will become mainstream. Juras also saw the cat-and-mouse game in AI-powered cyber threats coming, with both fraudsters and security experts increasingly using these tools, as well as the growing mark of on-device generative AI tools.

Of course, what we could predict with the greatest confidence was the growing use of AI agents in web scraping. To be fair, our own work to adapt every useful AI innovation to streamline public web data gathering made these less about forecasts and more about giving a heads-up.

What about the bubble?

The question of whether AI is overhyped was raised as early as 2023, and we were among those who raised it. In a way, we were right in expecting more ROI scrutiny. AI talk did steadily turn from initial wonder to questions of ROI in 2024.

However, despite discussions, these ROI worries never stopped investors from trusting the industry. Adi Andrei believes that the influx of money, without any proof, kept the hype going, even as the bubble was set to burst in 2025. Of course, the bubble didn’t burst, but Adi remains adamant in his scepticism.

According to Adi Andrei, in 2026, “Venture capital, media narratives, and corporate PR will continue to inflate expectations - creating a speculative bubble that may eventually burst with catastrophic consequences for markets and jobs.“

Whether this is true is up for discussion and, as you shall see, not all of us would put our money on a bursting bubble. However, Adi is far from alone in raising  these concerns and inviting us to prepare for the possibly grim future. While no one knows what will happen with the bubble and when, one thing is close to certain - the “bubble talk”  will continue next year.

Regulatory roll continues

Since 2021, Denas Grybauskas, now our Chief Governance and Strategy Officer, has been directing readers’ attention to crucial cases shaping the future of data access. Since then, the importance of web data to AI has entered the public consciousness.

We have accurately predicted that legal scrutiny of the AI industry will intensify and that this new focus will bring public data collection practices into the spotlight. Regulation of one industry increasingly affects another. What sometimes still slips from common realisation is that regulations meant for AI data access will often affect many traditional industries that depend on public data.

The intensification of legal battles was also something to be expected. While we thought legal cases might bring some clarity to the field, at the moment, it still seems murky in battles that will decide the future of open access and AI. Regulatory action is still on the rise, and we expect the prioritisation of responsible AI training and transparency to bring us closer to balanced regulation that respects ownership rights while protecting open access.

Our predictions for 2026

It is unlikely that AI bubble talk will stop anytime soon. A bubble bursting next year is just as implausible. The incremental improvements and continuous adoption and implementation of AI in everyday workflow should be enough to keep the industry going strong.

Juras has been accurately predicting the growing adoption of AI by developers and data scientists for years. Data parsing has been one of the most revolutionary AI applications for public web data collection. Increased automation means fewer resources required and easier access to web intelligence for businesses of all sizes.

Adi sees the current AI revolution in a broader context of power struggles. While he has a unique perspective, we all agree that the current battles for access to data have huge implications for the future of markets.

One way to balance AI innovation with safety, rights, and responsibility is to place greater emphasis on transparency and auditability. With increased internal controls, we hope to see more concrete proof that AI delivers measurable value in specific areas.

The importance of balanced regulation is most clearly felt in Europe, which risks being left behind in the AI race. As there are no signs that China will let regulation stop its AI goals and the USA is taking moves to restrain regulation rather than AI, the EU is also starting to reconsider its initial approach.

For the public web data collection industry, AI developments that are revolutionising web browsing open new frontiers and challenges. Browsing assistants are essentially web scrapers with extra features. It is fascinating to see how the merging of AI and scraping technologies, something we worked on to improve the results of our business clients and  Project 4β partners, is now being employed to revolutionise web browsing for everyone.

Summing up

Predicting the future is a thankless job. You are either wrong, or everyone will say that it was obvious all along. Yet, part of business leadership, and a very important one, is to see patterns in the markets and decide which matter the most. We have no special powers, only experience and data. But that’s enough to see the future as it unfolds.