An edge AI “easy button” for enterprise digital transformation
In their own effort to better (and more profitably) serve enterprise customers, CSPs are going through a number of network overhauls designed to deliver customizable compute, connectivity and services directly to business locations. Similarly, the enterprises targeted for consumption of this new class of CSP offerings, are themselves going through various digital transformation efforts. So how does AI fit into this complexity, and is there a way for CSPs to deliver something today that enterprises want? Has AI finally made the long-running edge conversation more relevant and tangible?
We asked experts from Dell Technologies and Intel—who are jointly going to market, alongside partners, with an edge AI solution—during the recent Telco AI Forum 2.0, available on demand here. Spoiler alert: yes, the consensus is that AI emboldens edge, the technology necessary to do it is ready to go, and CSPs delivering edge AI along with connectivity as a managed service is a very real route to market.
Dell’s Andrew Vaz, vice president of product management in the Telecom Systems Business, laid it out: “For something to be successful you need a bit of a convergence—a convergence of technology, the cost point to be something reasonable, as well as there to be a set of business needs. We’re actually hitting that point.” Specific to the AI piece, he said edge inferencing and locally running small models are delivering real business value today. “We’re actually seeing real deals materializing, combining all these technologies.”
From the Intel side, Cristina Rodriguez, vice president of the network and edge group and general manager of the wireless access networks division, called out the “symbiotic relationship” between private 5G and AI. “You have the opportunity to collect the data close to where the data is being generated…and then take actions on that data.”
Dell and Intel’s joint offering includes:
- Optimized and verified edge AI applications gathered from a range of ISVs and optimized for Dell’s PowerEdge XR8000 server platforms which feature Intel Xeon processors
- The opportunity to right-size AI investment by addressing use cases like computer vision that have clear return on a short timescale and don’t require discrete GPUs. By aligning the tech stack with the desired outcome, enterprises can materially reduce TCO
- Because enterprises use a variety of connectivity technologies—wired, Wi-Fi, or private cellular, for instance, Dell and Intel have optimized the platform for multi-connectivity environments
- As it is optimized and validated, the Edge AI Program is easy to deploy and market ready, which connects directly to time to value
- And the solution set draws on the extensive global partner ecosystems of both companies, meaning there are a variety of solutions across many industries and use cases.
Expanding on the CPU/GPU point, Rodriguez said, “You need the right tools for the problems at hand.” GPUs, she said, will obviously play a fundamental role in the AI world but, “What we see here in enterprise, you can address the AI enterprise needs greatly…by CPUs…All the applications don’t need the large language models…You don’t need to spend extra money getting a GPU there because you don’t need it.”
Vaz’s colleague Suresh Raam, director of telecom hardware partnerships and AI ecosystem, told us during an interview during Mobile World Congress Las Vegas that the approach developed with Intel “is giving [enterprises] an easy button.”
