TikTok has overtaken Google and Amazon to become the world’s most scraped website, reflecting how the boom in artificial intelligence is reshaping demand for online data.
According to Decodo’s second annual Most Scraped Websites report, the short-form video platform recorded a 321% surge in scraping traffic over the past year, rising from outside the top 10 to claim the top spot in 2025. The shift highlights how AI developers are moving away from traditional, text-heavy sources and increasingly targeting platforms rich in video, audio, and social interactions.
The report found that video and social media platforms now account for 38% of all scraping activity, overtaking search engines, which make up 24%, and e-commerce platforms at 22%. YouTube, South Korea’s Coupang, and the academic publishing site ScienceDirect were among six new entrants to the top 10 list, marking the sharpest annual shake-up since Decodo began tracking scraping activity.
Industry analysts say the rise of TikTok and other multimedia platforms underscores a profound transformation in the data economy. Training next-generation AI models requires vast quantities of diverse and high-quality datasets, and companies are increasingly turning to multimodal content — video, audio, images, and text — to gain an edge.
“Data might have been the new oil in 2006, but in 2025 it’s the fuel that powers artificial intelligence,” said Gabrielė Verbickaitė, Senior Product Marketing Manager at Decodo. “AI systems have an appetite for fresh, varied training data at unprecedented scale.”
Scraping — the automated extraction of data from websites — has long been a source of tension between technology companies, data providers, and regulators. While some platforms embrace controlled access through APIs, many others restrict scraping due to concerns over intellectual property, user privacy, and server strain. The surge in demand for TikTok’s content highlights the growing value of short-form, user-generated videos as training material for AI models designed to mimic human communication, creativity, and decision-making.
Decodo’s report suggests that organisations prioritising broad and diverse content inputs will be better positioned to innovate as multimodal AI reshapes entire industries, from media and retail to healthcare and education.
The findings also raise questions about how platforms will balance the commercial opportunities of data licensing with the need to protect users’ rights and safeguard their content from misuse. With AI’s appetite for data showing no sign of slowing, the competition over who controls — and who profits from — the world’s digital information is set to intensify.
