Key Takeaways

  • Cliff Weitzman admits avoiding B2B voice APIs was the biggest strategic mistake in Speechify's history, allowing ElevenLabs to capture the enterprise market.
  • Initial text-to-speech models are not commodities; they act as the distribution wedge for compound products like conversational voice agents.
  • Harry Stebbings warns against the Postmates effect: entering a market third behind well-funded leaders like ElevenLabs and Sierra.
  • Weitzman plans to counter entrenched competitors by targeting enterprise volume with a 10x compute cost advantage from buying Nvidia GPUs directly.

The Disagreement

Cliff Weitzman built Speechify into a consumer powerhouse for text-to-speech. But when ElevenLabs launched, it attacked B2B developers directly, leaving Speechify behind in enterprise revenue.

“B2C customers pay a lot less than B2B customers,” Weitzman told Harry Stebbings. “So 11 Labs huge credit to them leapfrogged us because they sell to B2B. Well historically we've only sold to B2C. 100% is on me. The biggest strategic mistake I made in the history of Speechify.”

Weitzman assumed that standalone voice models would commoditize quickly. He thought selling raw APIs was a race to zero. What he missed was how AI labs work.

“What I didn't understand is that the point of an AI lab like Speechify or like 11 Labs is to continuously innovate and the first product that you release is your wedge that gets other people to then later use your other technology,” Weitzman explained.

Now, Speechify is pushing into enterprise B2B voice agents. Stebbings pushed back hard on the timing.

“I think Sierra running behind them chasing and they're doing a decent job of it, but they've got Bret and they've got Sequoia and Green Oaks and every royalty of Silicon Valley behind them and they're still running behind chasing 11 Labs,” Stebbings argued. “I just think being third, the Postmates effect, is never a good market to be in when I could be the dominant consumer brand.”

Stebbings argued that enterprise voice is consolidating fast. Being third behind ElevenLabs and Sierra leaves you fighting for scraps while burning margins.

Weitzman rejected the Postmates comparison. His view is that voice AI is not a winner-take-all market, enterprise demand is uncapped, and lower infrastructure costs will reset the board.

“The best way to lose is not to be in the race. Be in the race,” Weitzman answered.

Who's Right (and When They're Wrong)

Stebbings is right about enterprise sales inertia. Once developers build on ElevenLabs or companies deploy Sierra for customer support, switching costs rise fast. Enterprise contracts do not churn just because a rival offers a slightly cheaper API. Being the third option in a venture-backed category usually destroys capital.

Weitzman is right about the nature of the product wedge. AI products do not stay static APIs. They are distribution pipes for next-generation models. If Speechify stayed strictly in consumer reading apps, its technical capabilities would eventually fall behind competitors who collect massive developer telemetry and enterprise cash flow.

Speechify can win market share only if its 10x cost reduction is real and passed directly to high-throughput enterprise buyers. If voice generation becomes a high-frequency background utility across games, phone support, and media apps, raw compute economics will beat early brand preference.

What to Do With This

Review your product roadmap this week. Identify which feature you treat as a standalone revenue line that is actually a distribution wedge for a larger workflow. If you are selling only to consumers because B2B feels like a commodity trap, price out an API tier tomorrow to capture enterprise developer demand before a competitor locks them in.