|
Therml
Silicon for the inference era. The world needs ten times more inference than it can build.
|
80 % |
|
(re)conceive - Kickstarting the efficiency era of AI.
(re)conceive is building the infrastructure for efficient, distributed AI. The company focuses on making intelligence faster, lighter, and more economical by...
|
79 % |
|
Nodarus, Computation Without Transport
The v42 targets 10x to 100x lower energy per token by holding model weights as persistent physical states on a topological substrate. Room temperature. Matur...
|
78 % |
|
LiteSolve: Massive compute, without the data centre
Forecasts, simulations and parameter sweeps mean running the same work thousands of times. LiteSolve runs it on compute that already exists and sits idle, fo...
|
78 % |
|
Threeum
Threeum is a frontier company making self-accelerating, self-organizing physical AI: the model and the machine designed as one system.
|
78 % |
|
Kannaki Labs | Neuromorphic software and hardware
Kannaki Labs builds biologically plausible neuromorphic software and hardware for real-time learning, without the energy and water cost of conventional AI.
|
78 % |
|
DeltaChip — In-Memory Compute for AI
DeltaChip designs analog and digital in-memory compute for AI inference, built on device physics, analog design and full-flow silicon capability.
|
78 % |
|
Natural Intelligence
Oscillatory AI Systems that mimic nature's power & efficiency. Unlike today's neural networks, oscillatory networks harness the way that biological systems c...
|
78 % |
|
nivedita ai — Research & Core Architecture
A research monograph detailing architectural systems utilizing dynamic inference-time logit steering, runtime token distribution tracking, and compute-optima...
|
78 % |
|
ENGRAM STRATEGIC | Algorithmic Defense Topologies
Deterministic simulation infrastructure for optimizing cross-border defense assets. Genetic placement algorithms for saturation resilience.
|
77 % |
|
QBoltz — Where machines think with photons
Photonic compute for the future of autonomy.
|
77 % |
|
ScalarState Technologies — Research & development
Independent R&D practice in hardware-efficient computation: edge computing, edge inference, neural network acceleration on FPGA, and sparse methods.
|
77 % |
|
Wetware Intelligence: A Bio-Digital Manifesto
|
77 % |
|
Aspheric — Programmable Photonics. Computing at the speed of light.
Aspheric is a pre-seed company building programmable photonic computers — a slab of lithium niobate, programmed with light, designed to be reconfigured into...
|
77 % |
|
Nakatomi Systems
|
77 % |
|
coprime: frontier-class AI on hardware you own
Factored AI: a belief network plus a reasoner, on hardware you own. Frontier-class performance at a fraction of the cost — private, provable, built for the e...
|
77 % |
|
Lathilda — Other ways to compute
Lathilda explores other ways computers can be built, programmed and experienced: alternative computational substrates, execution models and public instruments.
|
77 % |
|
ChipVat — a computer that cools itself by being sealed
A computing machine whose modules sit on a lattice inside a sealed vessel of supercritical CO2, pressurised by its own waste heat. No pump, no compressor, no...
|
77 % |
|
The Open Machine
Transformer-native AI systems where logical model scale and physical execution cost no longer have to grow together.
|
77 % |
|
Kurosh Kavoosi
Kurosh Kavoosi. Electrical and computer engineering, machine learning, and quantum computing. Research, projects, awards, and certifications.
|
77 % |
Detected social media networks