Fujitsu has just released a new technology: based on multiple AI agents capable of collaborating securely and quickly. This represents a significant step towards the future of supply chains. The expectation is that this project will help the company manage inventory, respond to crises, and optimize processes. The project will only begin actual testing in 2026, but the company already intends to demonstrate how intelligent automation can go far beyond simple tasks.
What makes Fujitsu’s new AI different from anything currently on the market?
This Fujitsu project is a major collaboration between AI agents from different companies. This was something difficult to imagine, mainly because of privacy, technical standards, and governance disputes between these companies. But Fujitsu’s technology allows these agents to exchange information securely and react suddenly, such as to an unexpected surge in demand or crises generated by natural disasters. This will be put into practice in the first tests, especially within the supply chain of the pharmaceutical company Rohto, which is partnering with the Institute of Science Tokyo.
This project emerges at a time when the use of AI in companies is growing, but still faces obstacles such as high computational demand and difficulty in customization. It is precisely because of these obstacles that Fujitsu is investing in lighter and more specialized AI models, capable of operating with less energy, yet still with good accuracy. The company’s goal is to enable agents to work together without relying on complex resources or systems.
Another important part of the project is governance. Fujitsu supports initiatives that strive to create secure, collaborative environments between companies and strengthen the competitiveness of Japanese industry. This is a way for the company to include standardization, safety guidelines, and constant monitoring among the agents within its project.
How Fujitsu is making AI more optimized, faster, and more practical
To ensure Fujitsu’s AI ecosystem functions correctly, the company has developed technologies that can reduce the size of AI models. One of the most impactful advancements is the “1.0-bit Quantization Technology,” which can compress language models by up to 94% without significant loss of accuracy. This technology allows advanced systems to operate on smaller devices, but with less energy consumption. This change could open doors to applications in robotics, edge devices, and industrial automation.
The company also created something called “specialized distillation” techniques, which basically work like intelligent filtering, inspired by the human brain. The goal of this technique is to reinforce information that is truly useful and discard unimportant information. This will make smaller models faster and more efficient. This technology will be essential to help agents make quick decisions within a supply chain.
A new model for complex operations
The future of AI envisioned by Fujitsu in this project is one that doesn’t depend on a single agent, but on multiple agents, each specialized and working together as a team. The “Multi-AI Agent Framework” is a project where each agent can have its own domain, such as logistics, demand forecasting, inventory, or transportation. In this way, they negotiate with each other to arrive at the best decisions.
Collaborative AI: When multiple agents make decisions together
With this new AI, which will operate using a multi-agent scheme, Fujitsu isn’t just introducing a new technology or a new way of thinking about supply chains.
By combining specialized agents, lighter models, and organized data, the company is investing in a future where complex decisions will be made intelligently, cooperatively, quickly, and securely. With tests scheduled for 2026, the company is expected to reveal how much of this project it can realize. But what is already clear is that the global market is about to begin a new chapter in intelligent automation.
