With data center power consumption surging due to the growing demand for AI inference, Israeli semiconductor fabless company NeoLogic has unveiled its low-power AI server CPU technology on the European stage and begun discussions on commercialization. This move comes amidst a trend where energy efficiency and inference cost reduction are emerging as key challenges for AI infrastructure.

NeoLogic announced on the 30th that it participated as one of 15 representative companies of the Israel Pavilion at VivaTech 2026, Europe's largest technology and startup exhibition held in Paris, France from the 17th to the 20th.

On-site, AI server CPU and low-power processor design technologies were introduced, and technology partnerships and commercialization possibilities were discussed with representatives from global and European cloud service providers and semiconductor OEMs.

At this year's VivaTech, sovereign AI, AI infrastructure sovereignty, and data center power efficiency were discussed as major agenda items.

It was emphasized that AI competitiveness is expanding beyond competition in model training to competition in energy efficiency and inference costs linked to the reduction of operating expenses (OPEX) and capital expenditures (CAPEX).

According to the International Energy Agency (IEA), global power consumption by data centers is projected to more than double from approximately 415 TWh in 2024 to approximately 945 TWh in 2030.

Europe is also expected to see an increase of over 70% during the same period.

On the other hand, as the pace of power grid expansion fails to keep up with demand, a survey by global real estate services firm Savills cited power supply constraints and delays in grid connection as the reasons for the year-over-year slowdown in the growth of new data center capacity in Europe in 2025.

The AI server CPU currently under development by NeoLogic is a server-class processor platform targeting AI inference and machine learning workloads.

The company stated that its development direction is focused on lowering power consumption and operating costs concentrated on GPU-centric infrastructure, and that it is preparing to supply its first products to a select group of customers.

To date, it has raised a total of $16 million and was selected as a '2026 Technology Pioneer' by the World Economic Forum (WEF).

“Currently, inference accounts for about 90% of AI workloads, and data center customers are burning through their annual AI budgets faster than expected,” said Abby Meshika, co-founder and CEO of NeoLogic. “We aim to support cloud service providers and their customers in utilizing more affordable and sustainable AI inference solutions through cost-effective and power-efficient AI server CPUs.”