The AI graveyard: a running list of projects and startups that didn’t make it
Sep 15, 2026, 12:00 PM · TechCrunch
TechCrunch opens an AI graveyard with Relay’s shutdown and an S&P figure that about 42% of corporate AI initiatives get abandoned—less scoreboard, more field notes on what failed.
Why it matters
Lauren Forristal’s TechCrunch piece marks Relay—an AI workflow automation Zapier alternative—shutting down Monday after larger platforms folded similar agents into their suites. It cites S&P Global Market Intelligence that about 42% of AI initiatives are abandoned by corporate parents, for funding, technical, competition, scaling, or demand reasons.
A running graveyard is healthy industry memory. Boom cycles erase failure; lists like this keep the tuition visible.
Builders and buyers both need the autopsy, not only the Series B photo.
From the desk
We’re glad someone is keeping the tombstones.
Relay’s fate is the classic wedge-product death: useful until the platforms you integrate with ship “good enough” natively. The 42% abandonment stat—if you trust S&P’s methodology—says most enterprise AI theater never graduates. That’s not an argument against useful AI. It’s an argument against confusing pilots with production.
We’re for ruthless prioritization: ship narrow tools that own a workflow platforms can’t casually clone, or sell to them. The harm of ignoring the graveyard is capital and attention burned on clones while real infrastructure and safety work go underfunded.
I’m watching which categories TechCrunch adds next—and whether abandonment rates fall as tooling matures or stay high as hype outruns problem fit.
Context
Relay shut down around September 15, 2026 per TechCrunch. S&P Global Market Intelligence is cited for the ~42% corporate AI initiative abandonment figure.
Who feels it
- AI startups
- Platform risk is existential for thin wrappers—differentiate on data, distribution, or regulation-hard workflows.
- Enterprise buyers
- Budget for kill criteria up front; 42% abandonment should be planned for, not mourned.
- Investors
- Graveyard entries are diligence exhibits for “why won’t Google/OpenAI eat this?”
What to watch
- Updates to TechCrunch’s running list and common failure patterns.
- Whether S&P refreshes the 42% figure with clearer methodology.
- Acqui-hires versus hard shutdowns as the dominant end state.
Companies: OpenAI