In a world strained by the energy demands of AI, one deep-tech startup is leading the way. Arago, a Paris and Silicon Valley-based company, has just closed an oversubscribed $26 million seed funding round to accelerate the commercialisation of a groundbreaking photonic AI chip. The chip, codenamed JEF, doesn’t rely on transistors or electrons. Instead, it processes data using photons, particles of light.
The funding round was co-led by Earlybird, Protagonist, and Visionaries Tomorrow, with participation from C4 Ventures and Generative IQ. It also attracted a high-profile list of angel investors with deep roots in semiconductors, machine learning, and enterprise software. Among them are Bertrand Serlet, a former VP at Apple and co-founder of Fungible, Christophe Frey of Arm, Olivier Pomel of Datadog, Thomas Wolf of Hugging Face, and Jack Abraham of Exowatt. That level of support speaks volumes not just about the novelty of Arago’s technology but about its commercial viability.
Behind Arago: from lab notebooks to a global vision
Arago was founded in 2024 by a trio of scientific minds: Nicolas Muller, Eliott Sarrey, and Ambroise Müller. Though still a young company, the team behind Arago has a rich pedigree spanning photonics, advanced electronics, applied mathematics, and machine learning. Their paths crossed in Europe’s photonic research circles and top-tier engineering institutions, eventually converging over a shared vision, rethinking AI’s hardware foundation from the ground up.
Arago’s 20-person team includes engineers and scientists recruited from some of the world’s leading tech companies and research labs, bringing together expertise from Apple, Nvidia, Microsoft, and other industry leaders. The company operates across three regions, France, North America, and Israel, positioning itself in global innovation hubs with access to top talent and critical semiconductor infrastructure.
Rather than trying to optimise around existing computing models, they started with a more radical question: What if the core mechanism of AI computing didn’t rely on electricity and transistors at all? This question laid the foundation for Arago’s flagship innovation.
Meet JEF: The photonic processor of tomorrow
At the heart of Arago’s mission is JEF, a novel kind of processor that trades electrons for photons. In conventional chips, data is moved and processed via electrical signals, generating significant heat and consuming enormous amounts of energy. JEF sidesteps this by using lasers to transmit and manipulate data. Because photons don’t generate nearly as much heat, the processor dramatically reduces the energy required to run demanding AI workloads.
JEF is built on a proprietary photonic technology, using lasers for data processing instead of transistors. Early internal benchmarks suggest that JEF could deliver up to 10 times lower energy consumption compared to leading GPUs, without compromising on compute performance or cost.
Even more remarkable is that JEF has been engineered to integrate seamlessly into existing AI infrastructure. It supports industry-standard software frameworks and can slot into conventional data centres without requiring expensive overhauls or proprietary ecosystems. This plug-and-play compatibility makes it especially attractive for developers, enterprises, and governments seeking to scale their AI operations without escalating energy use or hardware costs.
JEF is fully compatible with existing manufacturing processes and AI software stacks, allowing immediate deployment into today’s data centres. It also overcomes the historical technical barriers that have limited the performance and scalability of previous photonic or GPU alternatives.
Why light?
It’s easy to overlook the physics behind processors, but the truth is that current chips are reaching their thermal and architectural limits. As AI applications become more complex and ubiquitous, the industry faces a bottleneck not in innovation but in electricity. Massive GPU clusters now consume as much power as small towns, prompting concerns about environmental impact and sustainability.
Photonic computing offers a radically different path forward. Photons can travel faster than electrons and don’t suffer from resistive heating. While the concept isn’t new, Arago claims to have overcome the historical limitations that have plagued previous attempts, including the lack of software compatibility and scalability challenges. JEF doesn’t aim to replace every chip in existence but offers a compelling alternative for the most energy-intensive AI workloads.
Arago’s approach is not to replace all chips, but to provide a solution for the most energy-hungry AI workloads, offering a path to greener, more scalable AI infrastructure for industries such as healthcare, transportation, defence, and climate modelling.
Lean team, global ambitions
Despite having only 20 people on the team, Arago is already punching above its weight. Its staff includes engineers and scientists drawn from some of the world’s most respected tech companies and research labs. The team operates across France, North America, and Israel, positioning Arago in global innovation hubs with access to top talent and critical semiconductor infrastructure.
With the new funding, the company plans to ramp up product development, refine its chip for broader deployment, and expand its business partnerships. Arago is also actively hiring to grow its technical and operational teams in all three regions.
More than hype: The future of sustainable AI?
It’s tempting to dismiss new chip startups as part of the hype cycle around AI. But Arago stands out for its clear focus, scientific credibility, and pragmatic engineering choices. Rather than trying to dominate the entire AI stack, it’s offering a very specific value proposition: drastically cutting the energy footprint of AI at scale, without asking customers to rip and replace existing systems.
This approach could have significant ripple effects, especially as countries and enterprises increasingly scrutinise the carbon cost of AI development. In fact, if JEF performs as promised, it could become a key enabler of greener, more sustainable infrastructure in sectors ranging from healthcare and transportation to defense and climate modeling.
A bright path forward for Arago
Arago’s launch marks a turning point for energy-efficient computing. In an industry that often moves fast but conservatively, the startup’s blend of photonic ingenuity and real-world compatibility gives it a strong edge. With $26 million in the bank, top-tier backers, and a chip that defies convention, Arago isn’t just building new hardware. It’s rewriting the script on what AI infrastructure can look like in the 21st century.
In a future increasingly defined by light-speed computation and planetary constraints, Arago’s photon-first vision may prove not only timely but essential.
Nicolas Muller, Arago’s CEO and co-founder, said: “To build a product that’s not only high-performing but also truly usable, it’s critical to deeply understand the constraints of integrating a component based on a different compute principle into the broader ecosystem. We don’t have the luxury of waiting for the ecosystem to adapt – our technology needs to be compatible with everything from manufacturing processes to the AI software stack from day one.”
“Innovation in computing is one of the hardest deeptech challenges,” added Thong Le Hoang, co-founder of Visionaries Tomorrow. “Arago’s superpower lies in its first-principles vision of the future, combined with a deep understanding across digital, analog, and photonic approaches. The result is a unique hybrid architecture built on the latest industry breakthroughs and engineered to deliver step-change performance gains in the near term.”
Pierre Boudier, former Nvidia Fellow and Lead GPU Architect at Intel, said: “I’ve worked with Nvidia CEO Jensen Huang for nearly a decade. I know what it takes to succeed in this industry, and I see a lot of those same qualities in the Arago team.”
“Arago is creating a ‘DeepSeek moment’ for AI chips,” said Earlybird co-founder Hendrik Brandis. “This technology has the potential to defy the laws of AI computing, using only a fraction of the resources. The Arago team is executing towards this moment with unseen velocity and frugality.”
