Remember video? All were going to have profound impact on demand for networking gear. Except none did.
I'd very much disagree with including video here.
9 replies (most recent on top)
Generative AI is where the money is at (systems like ChatGPT) and that ship has already sailed.
Cisco needs to be focused on it's strengths - Security, Observability, and Networking infrastructure for companies that need security like gov/big business - not on the latest buzzword.
The fact Robbins and the ELT fell for this AI nonsense shows how ill-informed they are. Anybody who's done enough research on AI today knows expectations from the network are minimal at best. Very few companies require more than one rack full of high-end DPUs. One rack and that's it. Make it two for fault tolerance. A blip on the radar of network infrastructure vendors. Remember big data? Remember virtual desktops? Remember video? All were going to have profound impact on demand for networking gear. Except none did. But the ELT believed in each of them, every single time. Robbins is desperate to rejuvenate Cisco's brand by beeing seen with the cool kid on the block (nvidia). But when he realizes there's no money to be made there, what's he gonna do? More cost cutting. Again and again.
There is an infinitesimal chance Cisco rules AI and there would still be a 99.999% chance people would be laid off sometime next year. If only the software was that reliable.
23 years of layoffs. How much longer before reality dawns on you?
don't you think it's kind of optimistic to think they will wait til next year for another LR?
“Refocus” will not be the keyword, they are good at rewording but poor at executing.
In reality, AI demands for networking are mostly in the hyperscalers where Cisco already lost. Using pre-trained models for inference isn’t going to increase network demands. If anything, the opposite because the LLM does the job of summarizing a concise answer whereas users would’ve previously visited URLs. The claim that Cisco is already an “AI company” is bogus. The market demand is for GenAI, not some embedded routing algorithm 😂—and Cisco has less than nothing in this space, except a paper “deal” with NVidia, like everyone else.
Most of the routing and switching protocols can be considered as AI. Why not ?. They auto-update routes and discover neighbors and learn topology.
Next year Cisco will refocus on exploring Mars. A limited restructuring will be necessary.