Energy & water
By 2030, forecasts predict 3% of the world's energy, and enough water to feed 1.3 billion people, to be used every year by AI.
This isn't AI.
This is
natural intelligence;
in harmony with the world
and its people.
Currently raising our pre-seed round — write to Arun for details or to express interest.
Eykos is a startup grounded in its original research in philosophy, particle physics and molecular chemistry, and driven by humanitarian goals.
We are developing a radically new whole-stack paradigm for commerical intelligent systems, post-silicon and post-deep-learning, hoping to solve the present cost, energy and water crisis around AI.
New hardware, new theory, and a new, more sustainable and humanitarian superintelligent world.
Read our approach ↓
New hardware, new theory, and a new, more sustainable and humanitarian superintelligent world.
Read our approach ↓
01 / The research
A search for intelligence beyond its artificial reflection.
Our path uniquely begins with intelligence itself -- the metaphysics, the topology -- instead of with trial-and-error hyperscaled programming.
Eykos' research spans years, having investigated deep learning at its sparkling cutting edge, not for building on GPU clusters, but in order to understand what it is fundamentally about those components that creates the modern phenomenon. Now, we're bringing that together with our research in particle physics to create these pure systems.
02 / Why now
The cost of AI is being paid by the world.
AI has achieved extraordinary capabilities, but the trillion-dollar ecosystem is fragilely built around scaling and scaling further. Each generation demands more computation, harming the world in the construction and usage of the massive data centres that follow. We believe in intelligent systems that serve the world, not take away from it.
A technology meant to serve, harming
Where superintelligent systems should be used to improve the world, the unsustainable cost of producing them in the pre-Eykos way is far from it.
A human question
Distrust among large AI corporations is prevalent and rising. The architecture of intelligence will shape who controls it, who can afford it and where it can exist. Efficiency is not only an engineering problem -- it determines who intelligence belongs to.
03 / The whole stack
One theory, an entirely integrated future.
The current paradigm is built upon on chained monopoly, from the raw silicon to the language model. Eykos is the only brand hoping to replace the entire stack, removing reliance on the giants who control the silicon-to-chip chain, and creating hardware naturally built for our new theory and software.
Our physical systems and the principles of intelligence they run come from a single, shared idea -- theoretical, unified elegance and market independence in the same stroke.
04 / The mission
Sustainable intelligence for the world and its people.
Eykos' long-term mission is to promote harmony and humanitarian goals: the protection of the planet and a re-prioritisation of the individual in a world increasingly suffering from digital isolation.
We are labouring to create a paradigm of intelligence that is not only more capable, but more compatible with human happiness and the physical world it inhabits.
05 / People
Our team
The obsessive dedication and quality of our team is one of the cornerstones in our belief in the Eykos dream.
Arun
Arun is the founder, director and principal research engineer of Eykos, motivated by the humanitarian vision and theoretical elegance. His research in deep learning and the philosophy of intelligence spans six years. His preprint, Unifying Probabilistic Learning in Transformers, was published at 17 to positive reception. His early original work, beginning at 12, included dynamically-sized transformers, hyper-adjusting optimisation algorithms and custom Fourier attention mechanisms. Alongside this, Arun has developed a background in theoretical physics, relating his research in quantum gravity to intelligent systems. A Winchester College scholar, he achieved five A* grades at A-level, as well as distinctions in the national mathematics olympiad, and is about to read mathematics at Oxford.
Harry
Harry is the physical systems lead at Eykos, fascinated by patterns in the physical world and determined to use them for large-scale real-world solutions. A Winchester College scholar, he achieved distinctions in the national mathematics olympiad, top successes in the national chemistry olympiad and a top-20 performance in the national linguistics olympiad. He is reading chemistry at Oxford, where he earned a first-class result in his preliminary examinations.
Alex
Alex the systems and software engineer at Eykos, and is motivated by delivering efficiency to global solutions without draining global resources, replacing scale with ingenuity. His work focuses on theoretical computer science, and has worked in mathematical modelling, in which he has received international recognition as a winner of Siam's M3 Technical Computing Challenge, optimisation, and autonomous systems. A recipient of multiple national olympiad distinctions and also a Winchester College scholar, he is reading computer science at Oxford.
Write for anything.