Samsung Electronics has started a strategic pivot away from Western graphics-chip technology by developing its own mobile graphics processor, a move that could gradually loosen the grip of U.S. semiconductor giants on artificial-intelligence hardware and shift competition in high-end smartphones.
The South Korean company has embedded its first internally designed graphics processing unit in the new Exynos 2600 processor slated for next-generation Galaxy smartphones, according to people familiar with the matter. Samsung has designed its own mobile processors for years but until now licensed GPU blueprints from U.S.-based AMD. The shift marks the opening phase of a long-term plan to reduce reliance on external intellectual property in one of the most critical components for modern AI computing.
Graphics chips, originally built for rendering video games, have become the workhorse of the AI boom due to their ability to perform thousands of calculations simultaneously. By bringing GPU design in-house, Samsung aims to control more of the technology stack that will define AI performance in phones, tablets and eventually other connected devices.
The GPU in the Exynos 2600 was developed by Samsung’s engineers but still rests on an AMD architectural license, the people said. Future generations of Exynos chips are expected to shed even that foundational layer in favor of fully Samsung-drawn designs. If successful, Samsung would enter an exclusive tier of companies that design GPUs from the ground up—a club that includes Nvidia, AMD, Intel, Apple and Qualcomm.
The implications ripple beyond Samsung’s supply chain. American chip designers have dominated the GPU field for decades, supplying the silicon behind everything from gaming consoles to the data-center servers that train large AI models. A top-tier device manufacturer building its own graphics technology could eventually alter licensing revenues, competitive dynamics and where AI processing happens—between the cloud and the device.
Samsung’s push coincides with the industry’s pivot toward on-device AI, which processes data locally rather than sending it to remote servers. Tasks like real-time photo enhancement, live translation and generative AI features require powerful, efficient graphics processors. Controlling the GPU design allows Samsung to tune performance and power consumption specifically for its hardware and software, much like Apple has done with its custom silicon.
The strategy mirrors a broader decoupling in the tech world, where leading hardware makers are internalizing core chip designs to capture more value and differentiation. Apple’s shift to its own M-series and A-series processors has redefined its product margins and performance, while Qualcomm now markets its Snapdragon chips as dedicated AI platforms. Samsung, the world’s largest smartphone maker, appears to be preparing for a future where mobile devices serve as primary AI interfaces.
The Exynos GPU project is spearheaded by Samsung’s System LSI division, the chip-design arm of its semiconductor business. The Exynos 2600 launch is widely viewed as an attempt to regain momentum in mobile processors after Samsung lost ground to Qualcomm in recent years in key markets like the United States.
Whether Samsung can turn its GPU ambition into market advantage will depend on execution. Real-world performance, energy efficiency and developer support will determine if Samsung’s graphics technology can compete with established players. Still, the strategic direction is set: in the AI era, control over GPU design is no longer a luxury, and Samsung is no longer content to rent that control from others.
