Cisco’s $9B AI Order Boom Reshapes a Networking Giant
3 min readCisco just delivered one of the loudest signals yet that the AI build out is real money flowing through real cables. The networking giant reported $5.3 billion in AI infrastructure orders so far this fiscal year and lifted its full year forecast to roughly $9 billion, more than four times the prior year. Shares jumped about 15 percent after the close, and CEO Chuck Robbins told investors the company will cut almost 4,000 jobs starting May 14 as it concentrates resources on the AI opportunity.
From sleepy hardware vendor to AI plumbing kingpin
For years, Cisco was the slow growing networking veteran in a Silicon Valley obsessed with software. The AI boom flipped that script. Training and serving frontier models means moving staggering volumes of data between GPUs at extremely low latency, and that work falls to switches, routers, and the custom silicon inside them. Cisco’s bet on its Silicon One chip family is suddenly the most strategically interesting thing the company has done in years.
Robbins did not mince words about what that means for competitors. He told analysts on the call that vendors without their own chips will struggle to be relevant to the hyperscalers, framing Cisco as the critical infrastructure player for this AI era. It was a direct shot at networking rivals that lack a serious silicon program of their own.
The numbers behind the AI boom
Total Q3 revenue hit a record $15.8 billion, beating Wall Street estimates, with non GAAP earnings of $1.06 per share. Networking revenue jumped 25 percent to $8.82 billion, and total product orders grew 35 percent year over year. The standout line: hyperscaler AI orders alone totaled $1.9 billion in the quarter, up from $600 million a year earlier, according to CNBC.
Cisco said it secured five new hyperscaler design wins in Q3, including three for systems built around its 51.2 Tbps Silicon One P200 routing processor. Early orders from those wins should land in Q4, with the bigger ramp arriving in fiscal 2027. Looking further out, Robbins said network traffic should triple over the next three years as AI workloads scale across cloud regions.
Why the layoffs land on the same day
The job cuts, which begin May 14 and affect fewer than 5 percent of Cisco’s workforce, are not a cost of doing business trim. Robbins framed them as a deliberate reallocation. Positions tied to legacy product lines are being traded for engineers and sellers who can chase AI infrastructure deals. In a memo to employees, he argued that the companies that win in the AI era will be the ones with focus, urgency, and the discipline to continuously shift investment toward areas of strongest demand.
What this signals for the AI supply chain
Cisco’s update is a useful proxy for everything happening upstream. Hyperscalers are still ordering capacity at a pace that surprises even their own suppliers, and they want vertically integrated stacks where the silicon, the system, and the operating software all come from the same vendor. That favors the small group of companies that can hand customers a full AI fabric and squeezes anyone selling only one slice.
The market read the print as confirmation that the AI capex cycle has further to run. Investors should still watch whether order growth converts to recognized revenue on schedule, since hyperscaler timelines can slip. But after Wednesday night, the bar for the rest of the networking and silicon group just got noticeably higher.