TechCrunch published a running list of dead AI products on October 24. It fits a grim statistic: S&P Global Market Intelligence estimates that about 42% of corporate AI initiatives end up abandoned, for reasons from funding and technical trouble to weak demand. The list sorts into different ways to die: some products were absorbed, some were outcompeted by the platforms, and some failed on their own terms. The full list is at TechCrunch.
Absorbed into the flagship
Most of OpenAI’s disappearances were not failures in the usual sense. ChatGPT Atlas, a stand-alone AI-powered web browser, lasted less than a year before OpenAI discontinued it on August 9, folding its most useful features into ChatGPT. Operator, an agent that browsed the web and completed tasks for users, went the same way. Image generation moved inside ChatGPT too, shrinking DALL-E from a destination into a button. Sora, OpenAI’s video-sharing platform, shut down in March 2026 after high operating costs and weak user retention.
The one consolidation that did not work was the July 9 redesign that tried to turn ChatGPT into a broader super app, with separate Chat, Codex, and Work modes and the familiar version renamed ChatGPT Classic. Users called the new interface confusing and cluttered, and OpenAI rolled the change back soon after. The episode showed the limit: users accept a feature moving inside the flagship, but reject a flagship rearranged around it.
Beaten by the platforms
Relay lasted five years as an AI workflow automation tool, an alternative to Zapier. It shut down entirely in August 2026. Its problem was plain: OpenAI, Google, and other large platforms built similar automation into tools people already used.
Huxe, the AI audio app whose team Apple later hired, shut down in May 2026 for a related reason. Built by former NotebookLM developers, Huxe turned writing into podcast-style audio. Spotify kept expanding its own AI-powered audio tools to an audience Huxe could never match. Yupp, a free playground for comparing responses from hundreds of AI models side by side, shut down in March 2026 after never finding strong enough product-market fit, its founders said. At its peak it let users test more than 800 models.
Failed on their own terms
The hardware entries are the most expensive lessons. Humane raised $230 million for the AI Pin, a wearable meant to replace the smartphone with AI features. The product struggled with performance, Humane warned customers to stop using its charging case over a potential battery fire risk, and the company shut the business down in February 2025, with most assets going to HP for $116 million.
Rabbit’s R1, unveiled at CES in January 2024 as an AI companion that would perform tasks for its owner, sold a reported 100,000 units soon after launch. Early reviews called it unfinished, with unreliable performance and few useful integrations. Rabbit keeps updating the R1 as a computer controller and has announced Project Cyberdeck, a portable device built for vibe-coding.
Two software bets from the biggest companies belong in this group as well. Apple unveiled Apple Intelligence in 2024 with an improved Siri as a highlighted feature, then repeatedly pushed the new Siri back over engineering issues and bugs. The delay contributed to a $250 million settlement over claims about how Apple marketed the iPhone 16’s AI capabilities. Siri AI reached the iOS 27 beta in July 2026 and began rolling out to English-language users this month. Microsoft’s Recall, announced at Build 2024 as an AI-powered photographic memory that screenshotted everything on a Windows PC, drew immediate backlash, was delayed nearly a year, and faced fresh questions when a researcher built a tool that could extract data Recall had captured.
The legal pile-up
Character.ai stands apart because its end came through courts and regulators as much as through the market. Founded in 2021 by former Google researchers Noam Shazeer and Daniel De Freitas, it raised $150 million at a $1 billion valuation in 2023. In 2024, Google paid a reported $2.7 billion to license its technology and bring its founders back, leaving the startup to carry on without them. Families then sued, alleging the chatbots contributed to teen suicides and self-harm, and Disney sent a cease-and-desist letter over unauthorized use of its characters. Character.ai barred under-18 users from open-ended chats late last year and, with Google, agreed to settle several suits in early 2026. It was not enough. Character.ai’s bankruptcy filing in early October cited legal, regulatory, and operational headwinds. Disney is now the opening bidder, offering $20 million for a license to some of its intellectual property. The app still runs, and the sale is expected to wrap up by late November or early December.
For users, the lesson is practical. A feature you rely on can move inside a flagship app, as Atlas did, or vanish outright, as Relay did. The graveyard keeps growing because the industry launches products faster than it finds durable reasons for them to exist apart from the platforms. Checking whether a tool survives on its own merits, and keeping a backup for anything critical, is cheap insurance.
