Keep pulling the thread on Chuck Robbins.
Countries building sovereign AI infrastructure will need to partner with trusted U.S. technology companies because they lack domestic firms with comparable capabilities.
There are no U.S. companies that manufacture 5G radios because foreign government-subsidized competitors have made the market unprofitable.
At Cisco, the use of AI for code generation is shifting the bottleneck in software development from writing code to validating and testing the output.
The technology from Cisco's acquisition of Liba, now known as Silicon One, underpins GPU training models and runs enterprise AI stacks.
Cisco was one of the first 11 companies to gain access to an unreleased Anthropic AI model, referred to as "Mythos," for vulnerability detection in its code.
AI models like Anthropic's Mythos are expected to massively increase the volume of discovered software vulnerabilities.
Cisco has built and demonstrated a prototype of quantum networking silicon.
Cisco is exploring redesigning some of its products to require less memory as a short-term solution to the ongoing memory shortage.
Cisco is now concerned about the challenge of "code sprawl," the accumulation of unmanaged code, as a consequence of using AI code generation tools.
The primary driver for countries pursuing sovereign AI is geopolitical uncertainty and a desire for self-reliance, rather than a lack of trust in American technology companies.
The European Union is developing a "trusted technology vendors" framework as part of its approach to sovereign AI and technology procurement.
Cisco spends approximately $9.5 billion on research and development annually.