We exhibited at 4YFN this year, the startup event co-located with MWC Barcelona, and it turned out to be one of the most productive events we have attended since founding Lokutor. We went in with a technical demo and left with concrete enterprise conversations, exactly the kind of validation you cannot get from typical startup channels.
What We Set Out to Do
For a deep-tech startup building proprietary infrastructure, the hardest part is not making the technology work. The hardest part is getting in front of the right enterprise decision-makers who understand what you are building and are ready to move.
At 4YFN, we had three clear goals:
- Validate enterprise demand for CPU-native Voice AI through live demonstrations with corporate decision-makers and technology buyers.
- Convert technical credibility into early commercial traction through pilot discussions or letters of intent with enterprise partners.
- Build visibility within the European deep-tech and VC ecosystem ahead of our seed round.
What We Showcased
We brought a production-grade demonstration of our voice AI stack to the exhibition floor. Not a mocked-up prototype, but our actual flow-matching TTS engine (Versa) and streaming ASR system, running on standard CPUs in real time.
The core pitch was simple and visceral: the same voice AI quality that competitors run on expensive NVIDIA GPUs, we run on commodity hardware. Every person who saw the demo could see the latency numbers on screen, hear the naturalness of the voices, and understand the implications: drastically lower infrastructure costs and the ability to deploy voice AI anywhere without GPU dependency.
Throughout the event, we engaged in focused technical and business conversations with enterprise innovation leaders, CTO-level contacts, and corporate technology buyers. We also highlighted our partnerships with the Barcelona Supercomputing Center (BSC) through the AI Factory Acceleration Program and the NVIDIA Inception Program, which reinforced the credibility of our technical approach.
What Happened After
The outcomes from 4YFN were measurable and concrete.
Enterprise pilot discussions. We advanced one conversation from a cold introduction to an active exploration phase with a corporate partner. This is a significant jump in the enterprise sales cycle, where moving from first contact to qualified discussion can typically take months.
Meaningful technical conversations. We completed in-depth discussions with 3-4 enterprise innovation leads and corporate technology buyers who were genuinely interested in what we are building and how it could fit into their product roadmap.
Qualified leads. We established a pipeline of follow-up commercial discussions with companies that understand the technical differentiator and see the business case.
Stronger partner positioning. Our presence at 4YFN strengthened our position within both the BSC AI Factory Acceleration Program and the NVIDIA Inception Program, by demonstrating real enterprise traction alongside technical excellence.
Why This Matters
Deep-tech startups often fall into a trap: you build something genuinely novel, but the market does not know how to evaluate it. You end up spending all your time explaining the technology instead of selling the outcome.
4YFN gave us a stage where enterprise buyers already understood the value of infrastructure innovation. The conversations we had were not “what is Voice AI?” but “how fast can we deploy this?” That shift is everything when you are trying to move from technical demo to commercial traction.
We also learned something important about positioning. When you say “CPU-native voice AI,” the first reaction is confusion. Why would you run AI on a CPU? But when you show the demo, the latency numbers, and the cost implications, the reaction flips. It goes from confusion to “why is no one else doing this?” That gap between confusion and conviction is where enterprise deals live.
What Changed in Our Go-to-Market
4YFN helped us refine how we talk about what we do.
Before the event, we led with the technology. We talked about flow-matching models, spectral gating, and prosody-aware turn-taking. After 4YFN, we lead with the business outcome: voice AI that costs a fraction of GPU-based alternatives, runs on your existing infrastructure, and deploys in minutes instead of weeks.
The technology still matters. But the pitch now starts with the problem we solve, not the models we built to solve it.
Gratitude
We want to thank the 4YFN team for curating an event that actually connects startups with the right people. Too many conferences are networking theater. 4YFN was different. The conversations we had were substantive, the connections were real, and the outcomes were tangible.
We also want to thank our partners at the Barcelona Supercomputing Center and NVIDIA Inception for the credibility and support that helped open doors at the event.
What Is Next
We are continuing to build on the momentum from 4YFN. The enterprise conversations we started at the event are progressing, and we are deepening our engagement with the corporate partners we met.
If you are building voice-powered products and want to explore what CPU-native voice AI can do for your platform, reach out to us at contact@lokutor.com or visit app.lokutor.com to get started.