In this episode of Builders, host Lee Haynes sits down with Gilles Backhus, Co-founder and VP of Product at Recogni, to discuss the future of AI hardware, energy-efficient computation, and what it takes to scale a global deep-tech startup.

Gilles shares how Recogni is rethinking chip design through logarithmic math to cut power consumption, why the next wave of AI innovation won’t come from bigger models but smarter architectures, and how culture and hiring shape deep-tech success.

If you’re interested in AI chips, autonomous systems, data center innovation, or building teams that last, this conversation goes deep into what’s coming next for AI hardware, inference efficiency, and sustainable compute.

(00:00) Intro: Meet Gilles Backhus, Co-founder of Recogni

(01:00) From electrical engineering to AI hardware innovation

(02:00) The idea behind Recogni and its mission

(03:45) How Recogni’s logarithmic math cuts power use

(06:00) The real story behind AI’s carbon footprint

(08:30) Power, scale, and sustainability in AI computing

(09:40) Building Recogni across the US and Germany

(10:45) Inside Recogni’s engineering culture and low attrition

(13:30) Partnering with Juniper Networks and networking breakthroughs

(14:50) What Recogni looks for when hiring engineers

(19:00) Balancing global teams and cultural differences

(22:30) From VP of AI to VP of Product: Gilles’ leadership shift

(27:00) Staying ahead in a fast-moving AI hardware market

(30:00) From autonomous driving to data centers: Recogni’s focus today

(32:00) The next 5–10 years in AI innovation

(36:50) The evolution of AI chip specialization

(45:00) The rise of edge computing and analog AI

(47:00) Startup lessons and emotional resilience

(52:00) Closing thoughts and key takeaways