Extreme co-design is becoming the default
Cisco landed 5 new hyperscaler design wins in Q3 — 2 in optics, 3 in systems — including the first wins for its Silicon One P200 chip in scale-across workloads.
Robbins, Cisco’s CEO, said hyperscalers want “silicon diversity,” meaning they want multiple trusted options at very deep technical layers.
The lesson is not only about hardware.
It shows where strategic partnerships are heading when AI demand gets large enough: customers and suppliers work together much earlier, much deeper, and much closer to the architecture.
The same dynamic is starting to show up in software.
Joint engineering, co-built reference architectures, and deep technical integration are becoming the price of entry for any serious AI partnership across the stack.
The AI buildout is spreading from hyperscalers into the enterprise
Cisco raised FY26 hyperscaler AI infrastructure orders to $9B, up from $5B at the start of the year. The bigger signal for ISVs is what’s happening downstream:
Enterprise data center switching orders up >40% YoY
Nexus switches tagged for AI deployments up ~50% sequentially
Cisco reported ~$900M of AI orders YTD from neocloud, sovereign, and enterprise customers, separate from hyperscalers
Robbins called it the “AI acceleration we’re starting to see in the enterprise.” This is no longer a 2027 story.
AI demand is starting to pull forward modernization across networking, security, observability, data, and cloud operating models.
A single security narrative reshaped customer roadmaps in 90 days
Robbins said Anthropic’s Claude Mythos was not even on the radar 90 days ago. Now it’s “the number one topic that we talk about with almost every customer on the planet.”
The effect is concrete: refresh and security projects that customers had been delaying are now getting rapidly pulled forward.
Meanwhile, Splunk shows continued shift from on-prem to cloud subscriptions, with 1,000+ new logos expected in FY26.
Three lessons for alliance leaders:
Co-design is becoming the norm. The deepest AI partnerships now begin in engineering, often before commercial terms are on the table
Enterprise AI demand is showing up in pipelines already
Customer priorities can flip in just 90 days. Alliance teams need to be ready to change just as fast.
Cisco showed a broader hyperscaler ecosystem story: when AI demand accelerates, the impact spreads across the stack.
