Cybersecurity and content delivery network provider Akamai Technologies said that it is experiencing a surge in demand for Akamai Inference Cloud, little more than a week after its official debut on stage at the NVIDIA GTC Conference in Washington, DC. It reflects the numbers from the company’s third quarter of 2025, when Akamai achieved a revenue of $1.055 billion, marking a 5% increase compared to the same period last year.
This is why the company just announced Akamai Inference Cloud, a platform that redefines where and how AI is used by expanding inference from core data centres to the edge of the internet. Inference is the future of AI. Training teaches AI to think, but inference puts it to work. It turns models into real-time applications that reason, respond, and act. Inference delivers the experiences that make AI valuable. As a result, AI apps on Akamai Inference Cloud perform closer to users, respond instantly, and scale unbound.
The strong early interest in the Inference Cloud points to how fast organizations are moving from experimentation to execution and underscores growing demand for platforms that make it easier to operationalize AI in real-world environments. Organizations across industries are testing the purpose-built platform in production and expanding into new use cases in video, personalization, customer support, and consumer products.
With today’s AI inflection point, AI will not — it cannot — scale by doing more of the same,” said Adam Karon, Chief Operating Officer and General Manager, Cloud Technology Group, Akamai. “As an industry, we’ve become fixated on building ever-larger models in ever-larger data centres. Every headline trumpets billion-dollar contracts, data centre campuses measured in gigawatts, and record-breaking clusters of GPUs. But bigger isn’t better. Not any more. Centralization has become a constraint. Every request sent across continents wastes precious cycles, power, and money. Latency cripples user experience in trades, payments, games, and conversations. Power grids groan under the strain. Companies spend billions to train models, only to find that egress costs and seconds of delay render them impractical at scale.
Karon said that moving forward, the rise of intelligent agents, autonomous systems, and physical AI will trigger an inference explosion that dwarfs human-initiated requests by orders of magnitude. Every autonomous vehicle, robot, and smart system will become a persistent consumer of distributed edge inference, supplementing their local compute the way every smartphone today streams video it could never store locally. AI is breaking the decades-old cloud model as companies run up against rising inference costs and latency issues. CFOs are pushing back – and CIOs are quietly rebalancing workloads, doing everything they can to get closer to where data and users actually are. The bottleneck isn’t GPUs. It’s proximity.
Akamai says examples of current use cases include 8K video workflows. Harmonic is using Akamai Inference Cloud to deliver ultra-high resolution, multi-language video. With live video intelligence. Monks is using Akamai Inference Cloud to get the best shots from real-time multi-cam feeds. Akamai is working with a number of name-brand global retailers to help AI agent activities personalize product recommendations. Akamai is working with one of the world’s largest gaming companies to help enable context-aware chatbots across game modalities. Akamai is also working with mobile shopping companies to use AI inference to enable user controlled fitting room experiences with local photo and video. Finally, Akamai is working with one of the world’s largest toy makers to develop toys that learn, adapt, and interact.
“Live sports production requires processing at the edge – there’s simply no time for roundtrips to centralized data centres,” said Lewis Smithingham, EVP of Strategic Industries at Monks. Akamai Inference Cloud enables us to transcode 8K multi-camera feeds, process virtual reality content, and generate AI-powered play summaries in real-time, right at the edge. This enables us to use cards designed for video, with native 4:2:2 support. With Akamai Inference Cloud, we will accelerate the delivery of key capabilities, including identifying players and plays and delivering tactical insights to coaches while the game is still happening. That’s only possible by distributing advanced GPUs to the edge, and it will transform how we approach sports broadcasting and immersive fan experiences.”
This is why the company just announced Akamai Inference Cloud, a platform that redefines where and how AI is used by expanding inference from core data centres to the edge of the internet. Built with NVIDIA, Akamai Inference Cloud combines Akamai’s expertise in globally distributed architectures with NVIDIA Blackwell AI infrastructure and AI leadership to radically rethink and extend the systems needed to unlock AI’s true potential. Unlike traditional systems this platform is purpose-built to provide low-latency, real-time edge AI processing on a global scale by placing NVIDIA’s accelerated computing infrastructure closer to where data is created and decisions need to be made
“Personalizing live streams demands intelligence to be deployed closer to the user,” said Gil Rudge, Senior Vice President, Solutions and Americas Sales, Harmonic. “Using Akamai’s massive cloud infrastructure to run NVIDIA Blackwell cards at the edge is a game changer for AI-powered video innovation. Akamai Inference Cloud will allow us to run AI models locally, expanding the number of functions we can deliver cost-effectively within the same compute instance for faster response times, sophisticated personalization and more enriched video content.”