Lokutor Selected as a Finalist at South Summit 2025

Update, 12 October 2026: this post calls Lokutor “CPU-native” and says it runs without a GPU. We no longer describe Lokutor that way. We build small, efficient speech models of our own, and they are small enough to run in real time on an ordinary CPU; that is the proof, not the product. Our hosted cloud now serves them on GPUs to carry more calls per server, and the models are the point. See our models.

We are excited to share that Lokutor has been selected as a finalist at South Summit 2025, one of the most prestigious startup events globally. This year, thousands of startups from over 100 countries applied, and we are honored to be among those selected to present on the main stage in Madrid.

What is South Summit?

South Summit is the leading innovation platform connecting the global entrepreneurial ecosystem. It brings together startups, investors, and corporations from around the world. Being selected as a finalist is a recognition of the technology we are building and the vision behind it.

Why We’re Going

The voice AI space is crowded with solutions that promise the world but deliver high latency, high costs, and vendor lock-in. Lokutor is taking a fundamentally different approach: rebuilding the entire voice stack — turn-taking, noise cancellation, orchestration, STT, LLM, and TTS — for commodity CPU hardware.

At South Summit, we will be demonstrating our platform live — showing real-time voice AI that runs on standard hardware without a single GPU in sight.

Come Say Hello

If you are attending South Summit 2025 in Madrid, come find us at the startup exhibition area. We would love to show you what we are building, hear about your use case, and discuss how CPU-native voice AI can transform your product.

We will be sharing more details about our participation soon. Follow us on X for real-time updates.