In the latest white paper, Tachyum demonstrates supercharged LLM (Large Language Models) training in only 4 bits. Training LLMs is essential for advancing a business’s AI capability, but currently demands significant capital investment in hardware, as well as vast operational expense in time and energy costs over days to months of training. These barriers can be lowered by model quantization, which reduces the compute and memory footprint of deep learning models, allowing them to run faster and more economically.

Tachyum has demonstrated an algorithm to perform LLM training in the quantized 4-bit FP4 format, instead of the standard FP32 or BF16 training formats. This algorithm dramatically reduces memory and compute requirements, delivering faster, more cost-effective, and more energy-efficient training.

“With AI model sizes doubling every 3 to 6 months, there is an ever-growing need for greater efficiency to narrow the ratio between model sizes and processing times. With models moving from 1T to 10T parameters – with the goal to achieve human brain processing of about 100T synapses – traditional FP32 or BF16 formats simply are a barrier to the AI revolution. By quantizing to the 4-bit FP4 format, we continue to transform the economics of AI by delivering industry-leading performance, cost and power efficiency required to properly train LLMs.”

Dr. Radoslav Danilak

Tachyum ported the eBPF Just-In-Time (JIT) compiler on the Prodigy Universal Processor software emulation platform. eBPF is a technology that can run sandboxed programs in a privileged context, such as the operating system kernel. It is used to extend the capabilities of the kernel safely and efficiently without requiring changes to the kernel source code or having to load kernel modules. The software emulation system is an important part of ensuring that existing applications can run optimally on Prodigy processors and verifies that they are fully able to reap the benefits of high performance, low power and lower total cost of ownership when compared to running in traditional data centers.

Tachyum had the opportunity to take part in the Export Forum 2025, organized by the Ministry of Foreign and European Affairs of the Slovak Republic in cooperation with the Council of Slovak Exporters on October 15, 2025, in Tále, Slovakia. The event gathered Slovak companies with export potential. The goal was to provide Slovak business entities with the opportunity to actively dialogue with state institutions supporting export. The forum offered a unique opportunity for individual G2B meetings with economic diplomats working at the Slovak diplomatic missions around the world.

Tachyum’s team had productive discussions with several economic diplomats from the Slovak diplomatic missions in the regions of Europe, MEA and Asia. Tachyum remains dedicated to expanding and creating strong business connections in these regions, recognizing that AI and computing capacities are increasingly central to the national strategies of many countries. Tachyum’s base technology enables the proposed outcomes with reduced CAPEX and OPEX, lowering both technology investment and energy consumption.

Tachyum is preparing for mass market penetration of Prodigy and appreciates events such as the Export Forum, which are very helpful in actively seeking opportunities.