Akamai Technologies has taken a decisive step with the launch of Akamai Inference Clouda platform that redefines how and where AI runs. Their proposal is to bring inference, the phase where AI models process and apply their knowledge, directly to the Internet Edgethe point closest to the user and the devices.
The company, recognized for its global content distribution infrastructure, proposes a new architecture designed for Real-time AIminimizing latency and maximizing responsiveness. This movement is supported by technology NVIDIA Blackwellallowing Akamai to combine its distributed networking expertise with the most advanced accelerated computing on the market.
Akamai Inference Cloud: AI closer to the user
In the new generation of smart applications, from digital assistants to instant decision-making systems, speed and local context are essential. Akamai Inference Cloud addresses that need by placing processing power on globally distributed nodes. Thus, inference stops depending on large centralized data centers and runs right where the data is generated.
The objective is to offer an immediate interaction between machines and peopleopening the door to more seamless and personalized digital experiences. According to the company, this technology “enables intelligent and autonomous inference, capable of adapting to the environment in milliseconds.”
Akamai Inference Cloud redefines how and where AI runs
The launch has had the support of the top managers of both companies. “The future of artificial intelligence involves adopting the same logic that made the Internet a global, ubiquitous and scalable network. Thanks to NVIDIA’s infrastructure, we will bring AI inference to thousands of locations around the world, offering faster, smarter and safer decisions,” he says. Tom LeightonCEO and co-founder of Akamai.
For its part, Jensen Huangfounder and CEO of NVIDIA, emphasizes that “inference has become the most demanding phase of AI, because it requires real-time reasoning on a planetary scale.” He added that “by bringing inference closer to users, Akamai and NVIDIA are opening the door to a new generation of intelligent and scalable applications.”
An architecture designed for generative and autonomous AI
Akamai Inference Cloud combines NVIDIA RTX PRO serversequipped with Blackwell Server Edition GPU and Bluefield-3 DPU, along with software NVIDIA AI Enterprise. All of this is integrated into Akamai’s global infrastructure, which already has more than 4,200 locations around the world. This network allows distributed AI to work homogeneously, regardless of the geographic point in which the inference is executed.
The result is a platform capable of supporting Generative AI, predictive analytics and intelligent agents with minimal latency. Furthermore, the architecture is prepared to incorporate the new DPU Bluefield-4which strengthens the security and performance of processed data from the core to the edge.
Applications in key sectors
Among the practical applications, Akamai highlights the possibility of extending the enterprise AI factories to the edgeenabling smart commerce agents and hyper-personalized digital experiences. These systems will be able to adapt to user behavior and intent in real time.
The platform is also positioned as a strategic resource for financial, industrial and mobility sectorswhere decisions must be made in milliseconds. Thanks to its low latency, it can be used to detect fraud, optimize secure payments or coordinate automated workflows.
On the physical side, Inference Cloud offers crucial support for autonomous vehicles, industrial robots and smart urban networksin which security depends on an instantaneous response of the AI to the environment. With this solution, Akamai seeks to ensure that AI can “respond at the speed of the physical world.”
