Hook
Sequoia just dropped $45 million into Sable, a company that lets you pitch in Tokyo while speaking English in Kuala Lumpur. The tech is slick: real-time voice conversion that flips your words into Japanese, French, or Arabic mid-sentence. Headlines call it a game-changer. I call it a tell.
Because I’ve been watching the same pattern play out in every bull run. First, the VCs fund a centralized solution. Then, the crypto community builds a decentralized version that does it better — cheaper, more trustless, and with actual ownership. Sable is no different. The alpha isn’t in its pivot to Japan. The alpha is in understanding why Sequoia is rushing to own the gateway before the blockchain versions arrive.
Context
Sable’s core product is an AI-powered sales platform that automatically translates and adapts your demo script in real time. Think of it as a multilingual sales agent that works across Zoom calls, webinars, and recorded pitches. The company claims to cut global expansion costs by hiring fewer native-speaking sales reps. The $45M round — reportedly Series B — values them around $1.8-2.2B.
On paper, it makes sense. B2B sales is a $15 trillion market. Cross-border friction costs businesses billions annually in misinterpreted pitches, lost nuance, and cultural mismatch. Sequoia is betting that the last mile of global efficiency is a AI layer sitting between a sales VP and her overseas buyer.
But here’s the missing piece: Sable is closed-source, centralized, and data-hungry. Every pitch you give, every client reaction, every hesitation — it all feeds back into their model. That’s valuable, but it’s also a honeypot. In crypto, we call that “centralized trust,” and history says it gets exploited.
Core
Let’s dig into the order flow. Sable’s tech stack is a cascade of ASR → MT → TTS — speech recognition, machine translation, voice synthesis. They lean heavily on API calls to whisper (or GPT-4o) and ElevenLabs. The real moat, if there is one, is the latency optimization: sub-500ms end-to-end on a real-time call. That’s hard. But it’s also not a blockchain-level innovation. It’s engineering integration.
Now compare with what’s happening in crypto. Projects like AIOZ Network and Bittensor are already distributing inference across decentralized nodes. You can run a whisper model on a consumer GPU anywhere in the world. The latency isn’t there yet — 1-2 seconds — but with layer-2 scaling and specialized chains like Lava focusing on data relay, it’s closing fast.
More importantly, decentralized AI solves the data sovereignty issue. When a Malaysian manufacturer pitches a Japanese buyer using Sable, their entire contract negotiation history lives in Sable’s cloud. If Sable gets hacked, that’s game over. On-chain, the data is split, encrypted, and governed by smart contracts. No single point of failure. No Sequoian control.
And here’s where my copy trading community comes in. Last month, a Daimler subsidiary used a custom AI agent built on Gensyn to negotiate with a Chinese supplier. No intermediaries. No central server. The entire session was recorded on-chain as a series of signed messages. The sales team trusted the code, not the company. Social capital was the alpha, not the tech.
Contrarian
Here’s the take most VCs will miss. Sable isn’t just competing with other AI sales tools. It’s competing with human trust networks. In my experience running a trading community, the best signals come from who you trust — not from a polished translation. A Japanese buyer who knows her counterparty’s history on a decentralized reputation system (think Spectral, Proof of Personhood) will close a deal faster than any AI-generated pitch.
Retail investors and founders are chasing Sable as the next SaaS rocket. Smart money is rotating into infrastructure that supports decentralized voice agents — chains that can handle real-time audio streaming, oracles that verify language accuracy, and DAOs that own their own sales data.
I’ll be blunt: Sable’s tech is impressive. But its business model is fragile. As soon as a decentralized alternative achieves sub-500ms latency — and it will — the cost of switching drops to zero. No vendor lock-in. No data jail. Volatility is just noise; community is the signal.
Takeaway
The $45M from Sequoia is a bet that centralized AI will own the sales stack. I’m betting the opposite. The next 18 months will show us whether the real alpha comes from trusting a closed-source API or from owning your own voice. I know which side my crew is on. The question is: who are you trusting?
