The AI Graveyard: Notable AI Projects and Startups That Shut Down

The rapid growth of artificial intelligence has created a wave of new products, startups, and experimental tools, but many of those projects have failed to survive. Some were shut down after larger technology companies introduced similar features, while others struggled with technical limitations, high costs, privacy concerns, or weak user adoption.

The result is a growing AI graveyard filled with products that attracted attention, funding, or large user bases before eventually disappearing or changing direction. A report from S&P Global Market Intelligence found that many corporate AI initiatives are eventually abandoned because companies face challenges including funding constraints, technical hurdles, competition, scaling difficulties, and limited demand.

OpenAI’s AI experiments face setbacks

Even one of the most prominent AI companies has experienced product setbacks. OpenAI attempted to expand ChatGPT into a broader workspace experience with a redesign that introduced separate Chat, Codex, and Work modes. The company’s updated approach was reported alongside its efforts to build ChatGPT into a more comprehensive productivity platform through ChatGPT Work and related initiatives.

The redesign received criticism from some users who found the new interface confusing. OpenAI later returned to the familiar ChatGPT experience after the changes created friction. The company’s own announcement about expanding ChatGPT’s role in professional work is available through OpenAI’s ChatGPT for work initiative.

OpenAI has also moved several standalone products into ChatGPT rather than maintaining them separately. The company’s Operator agent, designed to browse the web and complete tasks for users, became part of OpenAI’s broader AI assistant strategy.

The company has also increasingly integrated image generation features into ChatGPT, reducing the role of DALL-E as a separate destination. OpenAI notified users about managing their existing DALL-E images as it changed how it handled image creation features.

Apple’s Siri AI

Apple’s planned AI upgrade for Siri became one of the company’s most closely watched artificial intelligence projects. When Apple introduced Apple Intelligence in 2024, the company highlighted a more capable Siri designed to understand context, work across applications, and complete more complex tasks.

However, Apple delayed several Siri AI improvements after facing engineering challenges and software issues. Reports about the delays highlighted the difficulties Apple encountered while developing the upgraded assistant experience.

Microsoft’s Recall

Microsoft’s Recall feature for Windows PCs became another major AI project surrounded by controversy. Announced at Microsoft Build 2024, Recall was designed as an AI-powered memory feature that captured screenshots of user activity, allowing people to search through their previous computer actions.

The feature immediately raised privacy and security concerns because the captured information could include sensitive data such as messages, passwords, and financial details. Microsoft later delayed the rollout while making changes to improve security protections.

Microsoft explained its vision for Recall as part of its workplace productivity efforts, while outside reporting examined the privacy concerns surrounding the feature. Coverage from Microsoft, CNBC, and The Verge documented the debate around Recall’s security approach.

Notion Mail

Notion Mail launched in 2025 as an AI-powered email product designed to help users organise and automate their inboxes. However, the company later found that many users were relying on separate AI agents to manage email tasks instead.

As a result, Notion planned to discontinue the product after determining that the standalone email experience was not gaining enough long-term adoption.

Humane AI Pin

The Humane AI Pin became one of the most recognised AI hardware failures. The wearable device was designed as an alternative way to access AI services without relying on a traditional smartphone and attracted significant investor interest before launch.

Despite raising substantial funding, the product struggled with performance issues and customer adoption. Humane eventually shut down its AI Pin business, with most company assets acquired by HP.

Rabbit R1

The Rabbit R1 was another heavily promoted AI device. Introduced at CES 2024, the device was designed to perform tasks for users through an AI assistant. Rabbit reported strong early sales after launch, including 100,000 units sold.

However, early reviews criticised the device for unreliable performance, unfinished software, and limited integrations. Rabbit has continued developing the product, positioning the R1 as a device that can control computer-based tasks. The company’s product details remain available through Rabbit’s official R1 page, along with later updates shared by the company.

Huxe

Huxe was an AI audio application created by former Google NotebookLM developers. The service transformed written information into conversational, podcast-style audio experiences.

The company shut down as larger platforms expanded into similar AI audio features. More information about the product was previously available through Huxe.

Yupp

Yupp allowed users to compare responses from hundreds of AI models in one place. The platform included model comparisons, user voting, and other community features.

Despite attracting attention among AI enthusiasts, the company’s founders said the platform did not achieve sufficient product-market fit and eventually shut down. The service’s original platform was available through Yupp.

Figgs AI

Figgs AI allowed users to create and interact with customisable AI characters for role-playing and storytelling. The platform reportedly attracted a large user base before its developers decided that maintaining the free service was no longer financially sustainable.

The company later ended the service, with community discussions documenting the shutdown process and end-of-life announcement.

The growing list of discontinued AI products shows how difficult it is to turn promising artificial intelligence ideas into sustainable businesses. While the industry continues to attract investment and experimentation, many projects must compete with larger platforms that have more resources, established users, and can integrate similar features quickly.