Supertrace AI builds ‘AI NOC’ as networks face looming talent crunch

- Supertrace AI is building an “AI NOC” to speed detection, root-cause analysis, triage and resolution across complex networks - It’s gaining traction, already monitoring “about 100,000 devices” across ISP, enterprise and data center customers - With engineers aging out, it’s hoping AI can help pick up the slack as networks grow Supertrace AI is pitching network operators on a new kind of helper for an industry grappling with an expanding footprint and fewer people to keep it running. The two-year-old startup is building what CEO Mahir Kalra described as “an AI NOC,” targeting internet service providers, data centers and enterprise campus networks. In a nutshell, Supertrace wants to make it easier to monitor and troubleshoot large, heterogeneous networks — the kind that can sprawl across thousands of devices and a tangle of vendors. Among other things, Supertrace’s AI NOC is designed to enable instant incident detection, root cause analysis, triaging and resolutions. Part of the pitch is consolidating “a lot of different panes of glass” that NOC teams currently toggle between during incidents, from vendor tools to ticketing systems and customer service data. But context is an even more important key to what Supertrace is doing. “Every network basically has their own Frankenstein's monster of like 22 network vendors,” Kalra said. These include the usual suspects like Juniper, Cisco, Arista and Nokia as well as Adtran, Ciena and Ubiquiti. Supertrace has built integrations with all these vendors to create a normalized layer across systems. But more importantly, it has created a way to feed its AI model the context specific to the network it's working on. “Over time, we’re building a network’s memory and really getting a full understanding of ‘when there’s an outage today, what are your runbooks?’” he explained, adding his team refers to this memory internally as Agentic Recall. By doing so, Supertrace is enabling AI agents to automatically executes those steps when an outage is detected to speed triage. Kalra said Supertrace is already gaining some traction in the market, with the company monitoring “about 100,000 devices” across ISPs, enterprises and data center customers. He added its business is pretty evenly spread across these three kinds of customers, but is geographically skewed to the U.S. Growing networks, shrinking workforce That momentum comes as network teams confront persistent workforce pressures. Kalra argued the company’s long-term vision is “to have fully self-healing networks,” but he acknowledged operators are hesitant to let AI agents make production changes. Supertrace is therefore taking a cautious path: “Our entire focus has been on read only, access,” he said, with an agent that can suggest “a bunch of commands” while keeping a “human in the loop.” But while cultural change and building trust in AI agents is hard, Kalra argued that operators may soon need – not just want – to supplement their workforce with AI. Concerns around an aging workforce have plagued the telecom industry for years – an Opengear report from 2023 warned that most CIOs expected a quarter of their network engineers to retire within the next five years. For those keeping track, that would mean by 2028. In a recent conversation with Fierce, networking company Meter suggested that by taking the grunt work off engineers’ shoulders, AI could make network ops more “fun” and lure in a younger cohort. But judging from Fierce’s observations at recent industry conferences, there doesn’t yet seem to be much of an injection of new blood. Kalra pointed out that network expansions in response to AI demand are compounding the problem. Supertrace’s bet is that reasoning models, paired with strong context, can help surface root cause faster — and preserve institutional knowledge as veteran engineers retire. Asked if its technology might end up taking jobs, Kalra said that to date no one has been impacted or fired. “We have this huge demand but then the supply of people that can actually service this is going down,” he said. “The tools that we're building now are helping our customers grow their network without having to spend more money on the NOC itself…We are squarely focused on helping our customers grow their network and keep their P&L in check.”