The AI Graveyard: Every Startup and Product That Didn't Survive

Chatbot platform Character.ai filed for bankruptcy just five years after launching. Its collapse joins a growing AI graveyard of shut-down products, pivoted startups and abandoned corporate bets — and 42% of AI initiatives get abandoned by their corporate parents, according to S&P Global Market Intelligence.

The reasons vary: insufficient funding, technical challenges, competition, difficulty scaling, or weak user demand. Here's a record of the most notable casualties, and what they say about where AI goes from here.

How did OpenAI stumble?

Even the most valuable private companies in the world aren't immune. On July 9, OpenAI updated its app into a broader "super app," combining separate "Chat," "Codex" and "Work" modes and renaming the familiar version "ChatGPT Classic." Users found the redesign confusing and cluttered, and OpenAI rolled it back soon after.

The consolidation strategy has worked better than the redesign. ChatGPT Atlas, a stand-alone AI-powered web browser, lasted less than a year before being discontinued on August 9, with its useful features absorbed into ChatGPT. OpenAI did the same with Operator, an AI agent that browsed the web and completed tasks on a user's behalf, and has increasingly built image generation directly into ChatGPT, shrinking DALL-E's role as a separate destination.

Sora, OpenAI's video-sharing platform, struggled with high operating costs and user-retention challenges. It shut down in March 2026.

What happened to Apple's Siri?

Apple unveiled Apple Intelligence in 2024 with an improved Siri as a headline feature. The pitch promised a Siri that understood context, knew what was happening across apps, and got tasks done effectively. Apple repeatedly delayed the release, citing engineering issues and bugs. That delay contributed to a $250 million settlement over claims about how Apple marketed the iPhone 16's AI capabilities.

Siri AI finally appeared in the iOS 27 beta this past July. The AI-powered version rolls out to English-language users first this month, with other languages to follow.

Which AI hardware failed hardest?

Humane raised $230 million for its AI Pin, a wearable designed to deliver AI without a smartphone. The product struggled with performance, then Humane warned customers to stop using its charging case over a battery fire risk. Humane shut down the AI Pin business in February 2025, with HP acquiring most assets for $116 million.

The Rabbit R1, unveiled at CES in January 2024, sold 100,000 units shortly after launch, Rabbit said. Early reviews called it unfinished, with unreliable performance and limited integrations. Rabbit keeps updating the R1 as a computer controller for agentic tasks and recently announced "Project Cyberdeck," a portable device built for vibe-coding.

Why did smaller AI apps shut down?

Relay, a five-year-old AI workflow automation tool built as a Zapier alternative, shut down entirely in August. As OpenAI, Google and other larger platforms built similar automation directly into their own tools, Relay struggled to hold on to a reason to exist as a stand-alone product.

Notion Mail launched in April 2025 as an AI-focused email product. Notion noticed users running separate AI agents on their inboxes instead, and the product is scheduled to shut down on September 22.

Huxe, an AI audio app built by former Google NotebookLM developers that turned written information into podcast-style audio, shut down in May 2026 — partly because larger platforms like Spotify began offering similar AI audio tools to existing audiences.

Yupp let users compare responses from more than 800 AI models side by side, including models from OpenAI, Google and Anthropic, and earn cryptocurrency for voting on answers. Its founders said Yupp never achieved strong enough product-market fit, and it shut down in March 2026.

Figgs AI operated from 2023 to 2024, letting users create customizable AI characters for role-play and storytelling. It drew more than 1 million users, but its developers said keeping the service free became too expensive to sustain.

What does the graveyard teach?

Microsoft's Recall shows the pattern cuts across Big Tech too. Announced at the 2024 Build conference as an AI "photographic memory" that periodically screenshots a user's activity for later search, Recall triggered immediate privacy backlash over its potential to archive passwords, private messages and financial details. Microsoft delayed it nearly a year and redesigned it — yet a cybersecurity researcher recently built a tool that extracts and displays Recall's captured data, reviving questions about whether the underlying problem was ever fixed.

The lesson isn't a scorecard of flops. It's that distribution beats novelty: when OpenAI, Google and Spotify fold AI features into platforms with existing audiences, stand-alone products lose their reason to exist — and the next wave of founders should price that risk in from day one.