We've been tracking 571+ AI companies across 200+ data sources for months. One pattern keeps showing up: the companies with the most press coverage are often not the ones with the most actual technical output.
We call it the hype/reality gap.
What is the hype/reality gap?
It measures the ratio between a company's press coverage volume and its actual technical output — GitHub commits, arXiv research papers, versioned product releases, and community engagement.
The result is a score that tells you whether a company is shipping or just announcing.
- 🏆 = significantly under-hyped (more reality than hype)
- ✅ = balanced
- ⚠️ = moderate hype gap
- 🤡 = significant hype gap (way more press than actual output)
This week's data
Biggest momentum move: Thinking Machines Lab jumped +597 places in 7 days — from rank #660 to #63 — on just 6 tracked mentions. What makes this signal interesting: those 6 mentions were all substantive technical coverage from independent publishers, not PR announcements. That pattern typically precedes larger moves.
Under-hyped companies this week (🏆):
- Western Digital: reality score 14.6, hype score 0.1 — consistently shipping, barely covered
- ServiceNow: significant technical output relative to press presence
- Open WebUI: active open-source development, low media profile
Significant hype gaps this week (🤡):
Several well-known names you'd recognise. The point isn't to name and shame — it's that the gap is measurable and consistent.
Why this matters for developers
If you're deciding which AI platform to build on, or which company's research to follow closely, press coverage is a terrible signal. A company can generate enormous coverage from a single announcement that has no code behind it yet.
The technical output signal — commits, papers, releases — is harder to fake and tends to predict sustained momentum better.
The data
All of this comes from SectorHQ — a real-time tracker that monitors 571+ AI companies across 200+ sources and updates every 5 minutes.
If you want the weekly signal digest in 3 minutes, BandPass covers the most interesting movements each day, with the investor lens and data context.
Curious what people's intuitions are about the hype/reality ratio in AI right now. Which companies do you think are most over or under-hyped relative to what they're actually shipping?
Top comments (0)