With 100M devices sold, Alexa has achieved some level of product-market fit, but Amazon is still looking for Alexa's strategic place within its business
Amazon's Alexa has been a huge, impressive and unexpected achievement. Amazon created a category from scratch and left both the AI leader Google …
Context & Ripple Effects
The milestone itself is not news — Amazon SVP Dave Limp had already put the number on record in his strategy interview confirming 100M+ Alexa devices sold. What Benedict Evans adds is the framing: category creation at this scale is an achievement that even left Google, the AI leader, playing catch-up on distribution, yet Amazon still cannot say what Alexa is *for* strategically.
The surrounding coverage sharpens the puzzle. The platform has 80,000 skills but no breakout hit, and Bloomberg reporting found Amazon paying developers up to $5,000/month to keep skills alive — demand-side success without a supply-side economy. Meanwhile Amazon keeps widening the hardware surface, most recently with $129.99 Echo Buds carrying Alexa beyond the home.
First-order effects
- Amazon is carrying Alexa as a cost center: it subsidizes both the devices and the developer ecosystem (up to $5,000/month per skill) while the strategic payoff — commerce, services, data — remains undefined.
- Skill developers are building on subsidized economics; without organic revenue, the long tail of 80,000 skills depends on Amazon's willingness to keep paying.
Second-order effects
- Google, despite leading in AI research, is forced to compete on distribution rather than model quality — shipping Assistant into every Android handset and partner device to match an installed base Amazon built by selling hardware at a loss.
- Each new Alexa endpoint like Echo Buds raises the stakes for adjacent categories (wearables, autos, appliances), pressuring device makers to license Alexa or justify their own assistants.
Third-order effects
- If no monetization layer emerges, voice platforms consolidate around whoever owns the endpoint and can absorb the losses — turning assistants from app-store-style ecosystems into embedded features of larger businesses.
- The pattern points toward ambient computing being judged on installed base and integration depth rather than skill counts or raw AI capability, a structural inversion of how the mobile app economy rewarded third-party developers.
The trend: Voice assistants have entered the distribution phase of AI, where owning the endpoint matters more than model leadership — but Alexa's 100M devices show adoption can outrun any business model.