Bright Data

From the Hola VPN to Luminati: how the company was born

Ofer Vilenski and Derry Shribman first founded Hola, a VPN service based on a peer-to-peer (P2P) network in which users share their internet bandwidth with each other. In 2014, out of the same network technology, the two founded Luminati Networks, a separate business that commercialized the power of this network, this time as a proxy service for businesses and not as a free privacy tool for individual consumers. The idea was that the Hola users who had unknowingly "contributed" part of their home internet connection had in effect created a huge proxy network that any paying customer could route browsing requests through. This model later also drew privacy criticism, since Hola users were not always fully aware that their computer was being used as a relay point for the traffic of business customers.

2017: separation and sale to EMK Capital

In 2017 Luminati Networks was fully separated from Hola, and a majority of its shares was sold to the British private equity fund EMK Capital. The amount first reported by various media outlets was about $200 million, but according to a later Calcalist report the final amount actually came to only about $165 million. The gap shows how early estimates of private deals sometimes change after the full details come out. Since 2017 EMK Capital has been the company's controlling owner, while the company kept operating under the day-to-day management of the original Israeli team. In March 2021 the company changed its name from Luminati Networks to Bright Data. The move was meant to separate it from the negative associations attached to the name "Luminati" and to position it as a legitimate, professional data platform, not just an anonymous proxy network.

What a "proxy network" actually is, and why AI needs one

A proxy network lets you send a browsing request to a website through an IP address that is not the original user's, so the website "sees" a different source of the request each time. For a company that wants to collect information from millions of web pages, this is critical: most websites block a single IP address that sends them thousands of requests in a row, suspecting it is a bot and not a person. According to the company's website, Bright Data offers a network of more than 72 million real residential IP addresses through which requests can be routed, so that from the visited website's point of view this looks like the activity of thousands of different human users and not a single bot. In the AI era, this ability has become even more essential: large language models need huge amounts of text for training, and much of it is collected from the open internet with exactly such tools.

The Suno affair: what leaked and what is alleged

In November 2025 the AI music company Suno was breached in a supply-chain attack, but the results were revealed to the media only in July 2026, when the site 404 Media published a report based on source code and documents leaked from the company. According to the report, the leaked files documented a systematic, large-scale scraping operation. A single file, labeled "youtube_music", logged more than two million tracks; other files documented tens of thousands of hours scraped from Deezer, from Genius and from the commercial music library Pond5. According to the same reports, part of the scraping was carried out using outside proxy services, including Bright Data, to get around the visited sites' protections against systematic automated access. It is important to note that, as of this writing, Bright Data has not published a detailed public statement denying or confirming these specific claims.

Round Hill Music's lawsuit against Suno and Bright Data

In August 2026 the music publisher Round Hill Music filed a lawsuit in the federal court for the Northern District of California against Suno and against Bright Data together, as a co-defendant. The complaint claims that Bright Data's Scraping Browser deliberately renders web pages without headers and footers, the place where copyright and ownership information is usually shown. Of the services Bright Data provided to Suno, the complaint says, in its own words, that they are "only good for copyright infringement." While the amounts claimed from Suno and from Anthropic in separate lawsuits are estimated in speculative terms of up to $1 billion, against Bright Data the lawsuit demands a defined, focused amount: 500 musical works at $150,000 each, for a total of $75 million for contributory infringement. It should be stressed that these are claims that have not yet been decided in court.

The wider debate: web scraping for AI training

The Suno affair is not an isolated case but a prominent example of a wider debate that follows almost every major AI company in the world: how far is it allowed to scrape seemingly public content from the internet to train a commercial model, and how do you tell when such scraping crosses into copyright infringement? Bright Data, for its part, presents its services as a technologically neutral tool for collecting public information and not as a tool meant for infringement. This position is similar to that of other network infrastructure companies that provide generic tools to their customers. Whether such an infrastructure provider bears legal responsibility for what its customers do with the tools it gave them is exactly what the court in California will have to decide, and even those who are not party to the lawsuit are following it as a possible precedent for the whole industry.