Amazon's Alexa platform, which has 80,000 skills, has yet to have a breakout hit; interviews with developers show Amazon pays up to $5,000/month for a skill
Matt Day / Bloomberg : Tweets: @mattmday Tweets: Matt Day / @mattmday : Amazon's Alexa has 80,000 apps, from Spotify to trivia and phone finders. But is any one of them a transformative, you-must-buy-a-smart-speaker hit? http://www.bloomberg.com/... In search of the killer app for voice, new for @technology
Context & Ripple Effects
Two years of related coverage trace a platform growing fast but not deep: skills went from over 15K in mid-2017 to 40K by May 2018, when Amazon made in-skill purchases generally available, then to 56,750 US skills by end-2018, up 120% YoY. The monetization layer arrived on schedule — but as Bloomberg's developer interviews show, the economics still run on Amazon subsidies of up to $5,000/month per skill rather than organic demand.
First-order effects
- Developers building Alexa skills are effectively working on Amazon-funded contracts, not businesses — the $5,000/month payments are the clearest signal that in-skill purchases have not yet made third-party voice apps self-sustaining.
- Amazon is carrying the cost of proving out its own app-store model, paying to keep a long tail of trivia and utility skills alive while Spotify-class first-party integrations do the real selling of smart speakers.
Second-order effects
- Rivals face the same wall at smaller scale — Voicebot counted just 378 Google Home voice apps and 65 for Cortana when Alexa passed 15K — so no competitor can claim a working voice-app economy either, keeping the smart-speaker market a hardware-and-first-party-services race.
- If subsidy-funded skills dry up, the long-tail catalog becomes a liability: maintenance costs without engagement, pressuring Amazon to curate rather than grow the count it has spent years touting.
Third-order effects
- That curation pressure has materialized: by 2025, developers worry about monetization as Amazon prepares Alexa+ without many of the 160K skills accumulated under the old model — the open long-tail store giving way to a smaller set of AI-native experiences.
- The pattern suggests voice app stores will not replicate mobile app-store economics; ambient assistants are consolidating around platform-owner-built capabilities, with third parties relegated to narrow integrations.
The trend: Voice assistant platforms are abandoning the mobile-style long-tail app store in favor of curated, platform-built AI experiences, because neither scale nor subsidies ever produced a third-party killer app.