Dubai: Positron AI Ltd, a US-based developer specializing in next-generation AI inference infrastructure, has made its first international expansion by establishing a presence in the Dubai International Financial Centre (DIFC). As a leading global financial hub in the Middle East, Africa, and South Asia, DIFC presents a strategic location for Positron AI's regional growth ambitions.
According to Emirates News Agency, the acquisition of a DIFC license marks a significant milestone in Positron AI's expansion strategy. The company has successfully raised over US$300 million, including a US$230 million Series B funding round, demonstrating strong financial backing. Positron AI has also achieved unicorn status and deployed its inference technology with a major US hyperscaler, underscoring its growing influence in the AI industry.
AI inference, which involves the deployment, operation, and scaling of large language models under real-world constraints, is driving global demand for computational resources. Positron AI's move to DIFC reflects its commitment to innovation in AI infrastructure, focusing on power-efficient, deployable inference solutions that cater to the increasing need for scalable AI capabilities. By establishing itself in DIFC, Positron AI aims to engage closely with industry leaders, regulators, and partners, while contributing to Dubai's ambition of building competitive digital and AI capabilities on a global scale.
Mohammad Alblooshi, CEO of the DIFC Innovation Hub, expressed enthusiasm about Positron AI joining the Dubai AI Campus' ecosystem. He emphasized that Positron AI's approach to AI inference infrastructure aligns with DIFC's evolution into an AI-native jurisdiction, positioning the Centre as a launchpad for future industries and a model for AI governance and innovation.
As the focus of AI demand shifts from model training to scalable model inference, Positron AI is poised to offer a next-generation alternative to traditional AI hardware providers. The company's proprietary infrastructure addresses critical industry challenges such as high power consumption and memory constraints, providing a cost-effective and energy-efficient solution.
Positron AI's first-generation server, Atlas, is designed to deliver large language model (LLM) inference for models with up to 500 billion active parameters, offering performance comparable to DGX-H100 systems but with reduced power usage and cost. The upcoming product, Titan, set for release in the first quarter of 2027, will further enhance Positron AI's ability to run frontier models with superior performance and power efficiency compared to NVIDIA's Rubin and Blackwell GPU systems.