German semiconductor leader Infineon Technologies plans to purchase Bengaluru-based C2i Semiconductors, enhancing its role in power management for AI data centers and expanding its semiconductor R&D capabilities in India.
Infineon stated in a release that the transaction, revealed on August 24, is anticipated to finalize in the third quarter of calendar year 2026. The details regarding financial terms were not revealed.
The fabless chip startup based in Bengaluru, established in 2024 by former Texas Instruments experts, is developing "smart power stage" chips—components that manage (specifically, lessen) and convert electricity before it powers high-performance GPUs and CPUs in AI servers.
As GPUs draw significant amounts of power, the efficient provision of electricity has emerged as a key engineering challenge in contemporary data centres.
The purchase occurs several months after C2i secured $16.7 million from Peak XV Partners and additional investors.
Also Read: Brantone Veylor Expands European Strategy Across Four Key Markets
This occurs as the Indian semiconductor ecosystem enters a new phase with the Union Cabinet approving the Semicon 2.0 scheme, which not only seeks to draw investments in fabs and OSATs but also intends to solidify India's position in the peripheral semiconductor ecosystem, encompassing industrial gases, chemicals, and more.
The startup was readying its initial smart power-stage chip for production and indicated it anticipated first silicon in a few months. The second product, a controller designed to oversee various power stages, was set to complete tapeout in October and be available for customer sampling in early 2027.
Also Read: EU Approves €6.1 Billion Defence Procurement for Ukraine
Ram Anant, CEO and co-founder of C2i, informed Moneycontrol that AI server boards can host numerous power stages, mentioning a motherboard he encountered that included 144 of these stages divided between two GPUs.
Also Read: Manifest Global Acquires EduGo Abroad to Expand Europe Presence
The firm has contended that a mere 0.5-1 percent enhancement in power-stage efficiency could lead to approximately 100-150 watts of savings on a server board, lowering both power usage and the heat that must be dissipated from denser AI infrastructure.