AI strategy is becoming increasingly important for businesses to get right. But reaping the rewards of AI requires enterprises to overcome a significant challenge: developing the powerful, connected, highly-performant infrastructure needed to support the technology at scale.
We’re excited to announce that Interxion is partnering with Scan Business UK to offer AI-as-a-Service, powered by NVIDIA DGX™ Systems, through its London campus. This partnership enables enterprises to start experimenting with the power of DGX in a colocation environment, supported by Scan’s ecosystem of technology and software offerings that help simplify the AI journey.
Scan, a leading AI solution provider in the UK and Europe, will manage the installation comprised of DGX systems, providing enterprises everything they need to develop and scale their AI initiatives, from deep learning frameworks to architectural and engineering support. At the same time, enterprises get all the benefits of colocating in an Interxion data centre that offers the best IT environment to support over time leading infrastructure for AI. With access to a community of connectivity and cloud providers for inference low latency, enhanced security measures, and advanced power and cooling technology critical to support training of models at scale.
Specifically for supporting secure last-mile connectivity from Enterprise premises to the Interxion London Campus, customers will be able to use Unitas Global services. This will enable us to link up our London Campus to the Janet Network
Proven experience building AI infrastructure is a key success factor for AI projects, which is why Interxion is thrilled to be collaborating with both Scan and NVIDIA on this initiative. As one of the few companies that have built AI infrastructure at scale in the real world, Scan is uniquely positioned to help enterprises solve AI challenges.
“We’ve seen the cost of cloud creep up on many enterprises, especially when models are being run continuously and using vast computing resources,” said Ivan Coetzee, Head of AI Collaboration at Scan. “Through our partnership with Interxion, we’re looking forward to helping enterprises combat some of these challenges, enabling them to move from the AI development phase to data processing and inference at scale.”
Addressing Enterprise AI Challenges
Many enterprises started experimenting with AI in the cloud, with multiple scattered projects running at once across business units. The cloud can be very costly, however, especially for training and building models. This solution helps IT execs manage shadow AI, scale their initiatives with power density and alleviate cost burdens, without slowing down the organisation’s AI roadmap.
Other enterprises have taken a different AI path, perhaps experimenting with NVIDIA’s AI technology in their own proprietary data centre. But those on-prem data centres simply aren’t capable of handling the power density and scale needed to accelerate AI projects. Colocation is the answer. Interxion’s test drive environment enables enterprises to kick their AI initiatives into high gear, combining the scalability and performance of third-party data centres with the support of Scan’s ecosystem. Colocating in a secure data centre also provides peace of mind for enterprises worried about the security of their data, since the data never lives in a public cloud or shared infrastructure.
Together with Scan, Interxion is offering a trial for enterprises to test this AI-as-a-Service solution. Contact us to learn more.