Fujitsu bets on sovereign AI, eyes $30 bn biz; B’luru lab to play key role
Fujitsu bets on sovereign AI to build a $30B business; Bengaluru lab with 400 researchers drives global AI stack, agents, Takane model, vision and next-gen CPUs.
Bengaluru: Fujitsu’s Bengaluru research centre, with a team of about 400 researchers, is playing a key role in developing the Japanese technology major’s global AI software stack as the company bets on sovereign AI, physical AI and next-generation computing to build a $30 billion business over the next eight to nine years.The company is building capabilities across compute, networking and software, with much of the supporting software development taking place in India. The researchers, many of them PhDs, contribute to technologies that Fujitsu plans to deploy across markets in Asia, Europe and the US, said Vivek Mahajan, CTO at Fujitsu.The company, which currently has a business of about $40 billion, aims to build the new business from a near-zero base, replacing a significant part of its existing revenue with technologies developed over the past five years.“Without compute, network and the software stack, you are not going to be able to develop and deploy AI at scale,” Mahajan said, adding that sovereign AI platforms would become increasingly important as businesses seek greater control over data, security and workloads.Fujitsu’s AI strategy rests on three areas: a multi-agent AI framework, its enterprise AI model Takane, and computer vision. Rather than develop its own large language models, the company works with model providers and focuses on frameworks that improve the accuracy and speed of enterprise AI agents.Fujitsu is expanding into physical AI, combining devices, software, computing and networks to enable robots to operate and coordinate more intelligently.On the computing front, Fujitsu has released a 2-nanometre CPU designed for power efficiency and confidential computing. Its next-generation Monaka-X processor is expected in about two years, with 256-core and 128-core versions featuring AI processing capabilities.Mahajan said demand was shifting from AI model training towards inference and agentic AI applications. The GPU-to-CPU ratio, earlier around eight to one, is moving towards one to one. Fujitsu is also pursuing a quantum computing roadmap targeting 250 logical qubits, alongside diamond-spin technology.The company is developing networking technologies to support data centres and AI radio access networks, while continuing to build hardware in Japan and much of its supporting software in India.Money habits form by age 7. Is your child learning the right ones? Join the TOI Money Masterclass.
Topics in this story
Gathered from external sources. Rights to this text belong to whoever originally published it.