Relay, an AI-driven workflow automation service meant to compete with Zapier, ceased operations completely in August. The tool let users automate email and task flows via AI agents, yet competition from OpenAI, Google and other major platforms that embedded similar features made the five-year-old service lose its purpose as an independent offering.
More recently, chatbot platform Character.ai entered bankruptcy only five years after its debut.
Beyond these two, many corporate AI experiments never survive. S&P Global Market Intelligence reports that roughly 42% of AI projects are eventually dropped by their parent companies, citing reasons such as lack of funding, technical hurdles, competitive pressure, scaling problems, or weak user demand.
The outcome is an expanding AI graveyard. Below is a look at several prominent AI products, startups, launches and bets that have been terminated, redirected, or fell far short of expectations. Consider this not as a tally of failures but as a record of lessons for the industry.
OpenAI's Setbacks
Even the sector's giants face difficulties. OpenAI, one of the world's most valuable private firms, has experienced its share of missteps. A recent effort to turn ChatGPT into a "super app" saw an update on July 9 that merged separate "Chat," "Codex," and "Work" modes while rebranding the classic interface as "ChatGPT Classic."
Users quickly complained that the new layout was confusing and cluttered, prompting OpenAI to revert to the familiar ChatGPT design shortly thereafter.
OpenAI also retired several standalone apps, folding their capabilities into ChatGPT. The ChatGPT Atlas browser, an AI-powered web tool, was discontinued on August 9 after less than a year, with its best features absorbed into ChatGPT.
Similarly, the Operator AI agent, which could browse the web and perform tasks for users, was merged, and image generation was increasingly built directly into ChatGPT, reducing DALL-E's role as a separate service.
The company's video-sharing platform Sora, plagued by high operating costs and poor user retention, was shut down in March 2026.
Apple's Siri AI
When Apple introduced Apple Intelligence in 2024, an upgraded Siri was highlighted as a key feature, promising contextual understanding across apps and efficient task handling.
Repeated postponements of the new Siri, blamed on engineering problems and bugs, eventually led to a $250 million settlement over allegations about how Apple marketed the iPhone 16's AI capabilities.
The Siri AI finally appeared in the iOS 27 beta in July, rolling out to English-language users this month, with additional language support slated for later.
Microsoft's Recall
Unveiled at the 2024 Build developer conference, Microsoft's Recall was billed as an AI-powered "photographic memory" for Windows PCs, periodically capturing screenshots so users could later search their own activity history.
The announcement sparked immediate backlash over privacy and security, with critics warning that Recall could archive sensitive data such as passwords, private messages and financial information. Microsoft postponed the feature for nearly a year to reassess its safeguards.
Even after redesigning Recall, a cybersecurity researcher created a tool that could extract and display captured data, reigniting concerns about whether the underlying issues had been resolved.
Notion Mail
Launched in April 2025, Notion Mail was an AI-focused email solution aimed at organizing and automating inboxes. The company observed that users preferred separate AI agents for email handling, leading to a scheduled shutdown on September 22.
Humane AI Pin
The Humane AI Pin, a high-profile AI hardware attempt, sought to deliver AI functions via a wearable rather than a smartphone. The venture raised $230 million and attracted considerable attention.
Performance issues plagued the product, and a warning about a potential battery fire risk in its charging case forced Humane to halt sales. The company closed the AI Pin business in February 2025, with HP acquiring most assets for $116 million.
Rabbit R1
Unveiled at CES in January 2024, the Rabbit R1 was marketed as an AI companion capable of handling tasks for its owner. Rabbit claimed sales of 100,000 units shortly after launch.
Despite early volume, reviewers described the device as unfinished, unreliable and limited in useful integrations.
The firm continues to update the R1, now positioning it as a computer controller for agentic tasks, and has announced a new hardware effort called "Project Cyberdeck" aimed at a portable vibe-coding device.
Huxe
Huxe, an AI audio app created by former Google NotebookLM developers, transformed written content into conversational, podcast-style audio.
The company ceased operations in May 2026 as larger platforms, such as Spotify, began offering comparable AI audio tools to broad audiences.
Yupp
Yupp offered a free sandbox where users could compare responses from hundreds of AI models side by side, vote on outputs and earn cryptocurrency.
At its height, the platform supported testing of over 800 models from firms including OpenAI, Google and Anthropic. Founders said Yupp never achieved sufficient product-market fit, leading to its shutdown in March 2026.
Figgs AI
Figgs AI operated between 2023 and 2024, letting users create and interact with customizable AI characters for role-play and storytelling. The service attracted more than 1 million users, but its developers cited the cost of keeping it free as unsustainable, prompting closure.
