Social media site X, music streaming service Spotify and AI chatbot ChatGPT were among several platforms down for thousands of users.
Facebook, AWS (Amazon Web Services), bet365, Canva, BrightHR, and the multiplayer game League of Legends also suffered outages, according to Downdetector.com.
More than 10,000 people reported issues linked to Cloudflare, a web infrastructure company, on Tuesday morning.
The issues were first reported around 11am, with several websites, including X, coming back online temporarily before suffering further problems attributed to an "internal server error on Cloudflare's network".
Cloudflare, which provides network and security services for many online businesses in order to help their websites and applications operate, says around a fifth of all global websites use some of its services.
The company said in a server update that it was "experiencing an internal service degradation" and that some services may be intermittently affected.
It added: "We are seeing services recover, but customers may continue to observe higher-than-normal error rates as we continue remediation efforts."
At 1.09pm, Cloudflare said it had identified the issue and was implementing a fix.
The outage coincides with Cloudflare's scheduled maintenance at its SCL (Santiago) data centre on Tuesday.
Graeme Stuart, head of the public sector at Check Point, a cybersecurity firm credited with creating the first firewall, said: "Cloudflare going down today sits in the same pattern we saw with the recent AWS and Azure outages. These platforms are vast, efficient and used by almost every part of modern life."
"When a platform of this size slips, the impact spreads far and fast and everyone feels it at once," he explained.
Read more from Sky News:
Paris's Louvre closes gallery weeks after heist
Hacker ordered to pay back over £4m in Bitcoin
Mr Stuart said the reported outages did not happen because each organisation failed on its own, but because "a single layer they all rely on stopped responding".
"Many organisations still run everything through one route with no meaningful backup. When that route fails, there is no fallback. That is the weakness we keep seeing play out," he said.
"The internet was meant to be resilient through distribution, yet we have ended up concentrating huge amounts of global traffic into a handful of cloud providers."
(c) Sky News 2025: X, Spotify and ChatGPT among those hit by major outage
Man, 33, admits selling substance online to assist with suicides
Princess of Wales calls on businesses to value 'time and tenderness' in rare speech
Man who stabbed Glasgow neighbour in parking space row jailed
Fresh police appeal over disappearance of teenager in Stornoway