AI workloads are putting new demands on networks, while AI tools are also being used to manage those networks. Cisco technical marketing director Kshitij Mahant argues that companies should plan both sides together, rather than leave their teams to compare notes once a quarter.
Watch Desk analysis
What happened
In a Forbes Technology Council article published on 9 October, Mahant describes a two-way relationship: networks must carry the large, synchronised data flows used in AI training, while AI can help operators spot congestion, interpret telemetry and automate network responses.
His practical recommendations are specific. He urges organisations to give network designers and AI workload schedulers shared visibility, coordinate accelerator purchases with network upgrades and capacity planning, and regularly measure outcomes such as job completion time and time to detect and fix problems.
Why it matters
A slow network link can hold up a distributed AI job while other accelerators wait. Meanwhile, AI systems used to monitor or manage a network depend on that same infrastructure being responsive. Treating the network as background plumbing and AI as a separate project can leave the two plans working at cross-purposes.
Mahant’s argument is that infrastructure decisions should account for both workloads and operations. That is useful guidance for organisations investing in AI, though the article is an executive’s perspective, not evidence that every company needs the same organisational chart or operating model.
Our read
The strongest point is refreshingly unglamorous: coordinate the people, budgets and measurements before the next expensive cluster arrives. AI networking may sound like a new technical frontier; some of the first gains could come from getting the planning meeting right.
Mahant writes from a Cisco role, so readers should treat this as an industry argument, not neutral research. Still, his proposed measures give teams something more concrete to debate than another promise that infrastructure will simply keep up.
What to watch
- Whether companies combine AI capacity and network upgrade planning in their budgets.
- Whether operators report job completion time and network-response measures together.
- Whether AI-led network operations reduce delays without making teams less able to understand or challenge automated decisions.
Discussion spark: Should AI and network teams share budgets and accountability, or would that blur responsibilities better kept separate?
Sources and evidence
- Networking For AI And AI For Networking: A Co-Design Loop For Intelligent Infrastructure – Forbes (9 October 2026, 12:30 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.