The News
Microsoft and Alibaba Run AI Networks on Cisco Chips
On October 12, 2026, Cisco announced at the Open Compute Project Global Summit in San Jose that its Silicon One P200 routing chip has entered production at Microsoft and Alibaba Cloud.
The P200 handles 51.2 terabits per second of routing capacity, designed for the bursty traffic of AI workloads. Alibaba Cloud uses it inside its in-house Quantum Cat routers running the open-source SONiC operating system, buying only Cisco's chip rather than a complete system.
In plain terms, Cisco built its business selling complete networking systems, but these deployments show it selling the chip alone to companies that design their own hardware. That opens a path into AI data center spending that Cisco's traditional product line could not reach.
Cisco also highlighted the G300, a 102.4 terabit-per-second switching chip now shipping, and the G302, a top-of-rack variant expected before year-end. Separately, Cisco and Nvidia introduced USFP, an open optical module specification with four times the bandwidth density.
Your Download Link Will Expire
If you still haven't downloaded my free "Safe Trade Options Formula" guide...
...please take a few seconds and download it right now before your new temporary download link expires.
I eventually plan to charge money for this training, so do yourself a favor and download it now...
That way, no matter what it costs in the future, you'll have a free copy on your computer.
Make sense?
The Company Behind It
Cisco Built Its Name on Systems, Not Chips
Cisco is a public technology company best known for networking equipment, security products, and enterprise infrastructure software. Silicon One, the chip architecture behind the P200, handles both routing and switching on one design, giving Cisco a single product line to sell into both networking markets.
Nvidia designs the GPUs and networking equipment used across most AI data centers and co-developed the USFP optical module specification with Cisco at the same event. The specification is designed as an open standard, with connector manufacturers TE Connectivity and Amphenol building the physical modules.
Cisco has traditionally sold complete networking systems, including the hardware, the operating system, and the support. The P200 deployments prove Silicon One can compete as standalone silicon in networks where the customer controls everything else.
Why This Matters Financially
The Revenue That Routers Cannot Reach
A hyperscaler that builds its own networking hardware will never buy a complete Cisco router. It will buy the Cisco chip inside, and that is the revenue path the P200 deployments create, giving Cisco access to AI networking spend that its traditional systems business could not capture.
Once a hyperscaler designs its network around Silicon One, the architecture creates switching costs that compound with every deployment. Alibaba deployed P200 silicon at massive scale across its eCORE backbone.
If the USFP specification becomes an industry standard, it shapes the next generation of AI optics around modules compatible with Cisco and Nvidia hardware. Selling the chip without the complete system strips out manufacturing and assembly costs while keeping the customer on Cisco's architecture.
A federal board just voted unanimously to lend roughly $3 billion to one American mining company.
That loan is bigger than the company's entire market cap. Not half of it. Bigger than the whole thing.
One of those two valuations is wrong, and the papers get signed in the second half of this year.
Limits and Uncertainty
Two Customers Are a Start, Not a Market
Broadcom's Tomahawk and Jericho chip families supply most of the switching silicon inside cloud data centers. Arista Networks, one of the largest AI networking switch vendors, builds its platforms on Broadcom chips.
Two production deployments are a proof point. Displacing a supplier that sits inside most cloud data center switches takes a longer list of wins, and neither the USFP specification nor the G302 switching chip has begun shipping in volume.
Hyperscalers are also building custom chips of their own, which could reduce demand for outside silicon from any supplier. The production deployments matter because they prove Cisco can compete at the chip level, not just the systems level; the real impact depends on whether more cloud builders adopt Silicon One and whether USFP gains enough industry backing to set the next optics standard.
Disclosure: This content is for educational and informational purposes only and does not constitute investment advice or recommendations. You should always conduct your own research or consult a qualified financial advisor before making investment decisions.
