- Lumen and IBM are teaming up to take watsonx and AI inferencing capabilities to the edge of the network
- Lumen's CEO told Fierce proximity to the enterprise is key for both security and speed
- IBM is beefing up its GenAI capabilities and thinks network players will play an increasing part in getting AI to the edge
Somehow, Lumen Technologies has gone from a company seemingly on its deathbed to one at the center of every AI announcement. Case in point, a freshly announced partnership with IBM that positions Lumen at the leading edge of the shift from centralized AI training to inferencing at the edge.
“This is all about the power of proximity and bringing AI capability right to the point of where data is created so that we can actually get to real time inferencing,” Lumen CEO Kate Johnson told Fierce in an exclusive interview. “It’s a pretty big deal.”
The new agreement will see Lumen embed IBM’s watsonx technology in its more than 50 edge data centers across the U.S. The idea is not only to cut down on latency to enable real-time AI applications (with Lumen claiming it can deliver sub-5ms turnaround time) but also to boost security and privacy in ways that might entice more customers from regulated industries like financial services and healthcare to lean into AI solutions.
“You don’t want to send this huge tranche of data up to some public cloud somewhere and wait for the answer to come back. That’s too costly from a data egress perspective and there’s too much latency associated with that,” Johnson said.
She added that the longer the distance the data has to travel and the more connection points it crosses, the more opportunities there are for the data to be intercepted. Thus, a shorter distance and fewer connections offers a pathway that is safer by design.
Where AI wants to be
For its part, IBM is also beefing up the AI capabilities it offers through watsonx.
Alongside the Lumen news, IBM announced updates to watsonx including a suite of prebuilt AI agents for human resources, procurement and sales tasks; tools to allow enterprises to build their own agents in under five minutes; integration with more than 80 enterprise applications (Adobe, Salesforce, Workday, etc.); and agent orchestration tools.
Oh, and it debuted a new agent catalogue for its watsonx Orchestrate platform that will streamline access to over 150 IBM and partner-built agents.
During a call with media, IBM executives said agents will play a key role in how AI moves to a distributed model. And when it comes to how agents get to the edge, IBM CEO Arvind Krishna said network providers like Lumen “will become a much closer part of these delivery mechanisms.”
Indeed, Fierce asked Johnson about the broader opportunity before Lumen on this front. While she didn’t want to stray too far from the IBM announcement, she acknowledged that the network platform it’s building depends on the existence of an ecosystem of third party intellectual property. So, in short, there’s “more in store on the horizon,” she said.
Bonus points
There’s one other element to the Lumen-IBM deal that will perhaps goes under appreciated.
By partnering with IBM and co-packaging its network connectivity with IBM’s software, Lumen is aiming to make its network more consumable for enterprises. Today, when enterprises pick a software solution, the networking piece is usually sold separately. The idea here is to eliminate that separate step entirely so it’s all one transaction.
According to Johnson, that strategy is part of a play to close what she called the “readiness gap” that’s preventing enterprises from tapping into the full value of AI.
But the whole setup, of course, also means that IBM will be doing quite a bit of selling for Lumen.
And if Lumen manages to replicate this deal with, say, a hyperscaler or two that it’s already cozy with? Well, that would certainly be one way to widen the competitive moat with other networking providers, wouldn’t it.