Oxylabs builds the proxy, scraping, and browser-automation infrastructure companies use to collect public web data at scale. Its Oxylabs funding moment arrived on July 9, 2026, when Warburg Pincus invested $130 million at a $3.6 billion valuation — the company’s first outside capital since founder Julius Černiauskas started it in Lithuania in 2015. The core problem is simple: AI agents can’t do serious work with yesterday’s internet. They need current pages and reliable retrieval. They also need systems that don’t fall apart every time a website changes.
That’s why this deal matters more than the usual venture headline. Oxylabs wasn’t a young startup looking for product-market fit. It had already grown into a bootstrapped business with $350 million in annual recurring revenue and more than 350,000 technology teams using the platform. Its infrastructure handles billions of requests a day across e-commerce intelligence, cybersecurity, travel, and brand protection. Now it’s taking fresh capital because the agentic AI boom could turn a niche plumbing layer into a much bigger one.
What does Oxylabs do with web scraping and AI agents?
At the product level, Oxylabs is basically a full-stack public web data platform. A customer can start with the Web Scraper API Playground, enter a target URL or search query, and tweak things like user-agent type and JavaScript rendering. They can also adjust localization, filters, limits, and session settings, then export the result as code or data in formats like HTML, JSON, or Markdown. It’s a lot closer to “test, tune, ship” than the old way of building and babysitting scrapers by hand.
It also sells tools for the messy part of the job — interacting with pages that behave like applications, not documents. Oxylabs’ Headless Browser and browser-instruction features let developers automate clicking and scrolling. They can also fill forms and wait for elements. And because nobody wants to hand-write those flows every time, the platform can prefill instructions from a natural-language prompt and turn them into structured JSON for API use.
Then there’s the data cleanup layer. OxyCopilot can generate request code and parsing instructions, while dedicated parsers return structured JSON for targets like Amazon, Google, Google Shopping, YouTube, Bing, Walmart, Etsy, and Target. For AI workflows, Oxylabs also supports Model Context Protocol and Markdown output so raw HTML can be transformed into something Claude, GPT, and other models can actually use without a bunch of extra formatting work.
That matters because the customer experience has changed. Before, teams often had to stitch together proxies and browser automation. They also had to manage parsers, retries, and formatting on their own. Oxylabs is trying to collapse that stack into one platform — plus AI Studio, a no-code prompt-based tool it released in 2025 for simpler collection jobs.
Who founded Oxylabs and how big is it now?
How Oxylabs started
Oxylabs began in 2015, inside the broader Tesonet orbit, after a friend asked whether unused IP addresses could be rented to a data-collection company. Julius Černiauskas turned that question into a business, starting with datacenter proxies before expanding into a wider web intelligence platform. That origin story sounds almost too tidy, but it explains why Oxylabs has always looked more infrastructure-first than hype-first.
Černiauskas isn’t some random founder who discovered AI late. He studied math and statistics, worked in Lithuania’s early digital advertising scene, and did SEO. He started a digital marketing business, then later worked at Vinted and data company Exacaster. That background gave him a practical read on search data, marketing demand, and what businesses would actually pay for. That’s more useful here than a flashy AI résumé.
Oxylabs changed leadership in March 2026, with Černiauskas moving from founder-CEO to chairman and Vytautas Savickas taking over as CEO. Savickas is an operator with deep execution and product-building experience, which fits where Oxylabs is now: less “invent the category” and more “scale the machine without breaking it.”
Traction, fundraising, and competition
The company is long past the beta phase. It’s live, global, and profitable from its early years. It has topped $350 million in ARR, built a patent portfolio of more than 160 inventions, and serves a customer base that includes large enterprises around the world. Oxylabs also acquired Webshare in 2022 and ScrapingBee in 2025. That gave it stronger reach across both enterprise buyers and developers.
The Warburg Pincus deal marks a real change in financing strategy. It came through the firm’s Capital Solutions Founders Fund, values Oxylabs at about $3.6 billion, and gives the company more room to expand its network. It also gives Oxylabs room to build next-generation products and keep doing M&A. CFO Jurgis Rudgalvis has already said the company wants more corporate development opportunities that add technology and products to the wider platform.
Competition is tough, and Oxylabs isn’t alone. Bright Data pitches a broad web data platform with 400 million-plus residential IPs and 660-plus prebuilt scrapers. It also has a dataset marketplace and browser infrastructure for AI agents. Zyte leans on its roots in Scrapy and sells an all-in-one scraping API with managed extraction and AI scraping features. Apify comes from a different angle — cloud-hosted “Actors” for scraping, browser automation, and AI-agent workflows. Oxylabs’ edge is its push for a more enterprise-grade bundle around compliance and proxies. It also wraps in headless browsing, parsing, and AI-friendly output in one stack, rather than just offering raw access or developer components. The legacy alternative, honestly, is still a mess of in-house scrapers, patchwork proxy vendors, and teams wasting time fixing breakage every week.
Why does Oxylabs funding matter right now?
Because this isn’t rescue money.
Oxylabs had already proved people would pay for its infrastructure without venture backing. So taking outside capital now suggests management thinks the shift from traditional scraping to agentic web access is big enough to justify changing a formula that had worked for more than a decade. Warburg isn’t betting on a concept deck. It’s betting on a business that already has customers, compliance muscle, and a global delivery network.
Savickas put it plainly: “The next generation of AI won’t be powered by static indexes that only capture yesterday’s internet.” That’s the whole thesis. If AI agents are going to monitor markets, compare products, complete tasks, and pull live information from the open web, the boring infrastructure layer becomes a lot less boring.
Warburg’s reasoning was straightforward too. Allison Ross, a principal at the firm, pointed to Oxylabs’ technology, expansive network, and enterprise relationships. Translation: the investor likes that Oxylabs already looks like a scaled supplier, not a science project.
There’s also a Lithuania angle here. Oxylabs is the second unicorn to come out of the Tesonet pipeline, and this valuation puts it among the country’s most valuable tech companies. That’s not just local bragging rights. It tells later-stage investors that serious infrastructure businesses can be built in Europe and sold globally without following the usual Silicon Valley playbook.
What market is Oxylabs funding betting on?
The direct market is getting big fast. Grand View Research valued the global web-scraped-data segment of the alternative data market at $2.54 billion in 2025 and projects it will reach $35.89 billion by 2033, a 36.9% CAGR. Those are huge numbers for what used to be treated like a back-office tooling category.
The second shift is even more telling: machines are already dominating large parts of the web. Thales said in its 2026 Bad Bot Report that bots made up more than 53% of all web traffic in 2025, up from 51% the year before. Not every bot is an AI agent, obviously. But the direction is hard to miss. The web is increasingly being read, monitored, and acted on by software, not people.
That’s why Oxylabs’ timing feels sharp rather than opportunistic. Training a model on archived pages is one market. Building systems that can reliably access live, changing websites in production is another one entirely. That second market is where a lot of enterprise AI spending is likely to pile up next.
Should you watch Oxylabs funding from here?
Yes.
Oxylabs funding isn’t about proving the business works. That part already happened. The question now is whether Oxylabs can turn a decade of web-data infrastructure into the default layer enterprises trust when AI agents need the live web — fast, structured, and without constant breakage. If that happens, this won’t look like a late growth round. It’ll look like a pretty early claim on a much bigger market.
Read how Elevation Capital raised a $500M Fund IX and launched a $400M Holdings vehicle to back India's next generation of AI startups from seed to pre-IPO.
FAQ about Oxylabs funding
- What is the Oxylabs funding deal? Oxylabs raised $130 million from Warburg Pincus on July 9, 2026, at a valuation of about $3.6 billion. It was the company’s first outside investment since its 2015 founding, and the money came through Warburg Pincus Capital Solutions Founders Fund.
- How does Oxylabs actually collect web data? Oxylabs combines proxies and a Web Scraper API. It also uses headless browser automation, parsers, and AI tools like OxyCopilot in one workflow. A developer can test requests in the Playground, automate actions like clicks or form fills, and return data in JSON, Markdown, or MCP-friendly formats for AI systems.
- Who founded Oxylabs? Julius Černiauskas founded Oxylabs in 2015 and moved to chairman in March 2026. Before building the company, he studied statistics, worked in digital advertising and SEO, launched a marketing business, and held roles at Vinted and Exacaster. Vytautas Savickas now runs the company as CEO.
- Is Oxylabs an AI company or a web scraping company? It’s more accurate to call Oxylabs a web data infrastructure company that’s being pulled deeper into AI. Its products still sit in the web scraping and proxy category, but features like headless browser support, AI-generated parsers, and MCP output make it increasingly useful for AI agents and enterprise retrieval systems.
