Gordon Thomson has watched customers ask Cisco to simplify its management stack for the better part of 15 years. At Cisco Live 2026 in Las Vegas this week, where the company launched Cloud Control, a unified agentic AI platform for managing infrastructure across networking, security, and compute, Thomson told me the wait is finally over and that AI made it possible.
It’s a fair point. For enterprise IT teams who’ve spent the better part of a decade managing fragmented tooling across Catalyst, Meraki, Umbrella and a dozen other Cisco products, a single operational view is genuinely useful, not just a slide-deck aspiration. The launch spans autonomous network operations, data centre and campus networking, security against AI-accelerated threats, multicloud connectivity and long-term resilience services.
The competitive context matters. Juniper’s Mist AI platform, now within HPE’s portfolio, has been making similar claims about AI-driven operations for several years and has a strong track record in campus environments. Nokia has pursued comparable ambitions on the service provider side. The difference Cisco is pressing is cross-domain breadth, spanning campus, data centre, security and cloud under one agent-powered management plane, which none of those competitors can match from a single vendor. That’s a real differentiator, though it depends entirely on how deeply a customer is already inside the Cisco ecosystem.
Thomson framed the Cloud Control launch as evidence of a cultural shift inside Cisco: shipping a product in US-only controlled availability rather than waiting for global readiness. “We need to move fast,” he said. “There’s a better decision to make the wrong decision and adjust than to wait for the perfect decision.” That’s a credible posture for a company of Cisco’s size, which has historically moved at the pace its installed base required rather than the pace the market wanted. Cloud Control enters controlled availability in the United States now, with global availability to follow at an unspecified date, and it’s on that timeline that the Africa conversation becomes more complicated.
Thomson is optimistic about the continent, and that optimism isn’t hollow. His argument is that AI has reduced the capital cost of infrastructure capability to the point where the levelling opportunity is real. “AI is the leveller,” he said, “and it’s about the pace you move at.” The on-premises angle matters here: if AI workloads can run on a machine next to your desk rather than exclusively in a hyperscaler’s region, African organisations that can’t access sufficient cloud infrastructure regain some ground. Cisco’s Sovereign Critical Infrastructure portfolio, which formally arrived in South Africa in April, is the product expression of exactly that logic.
The part of the conversation Thomson navigated most carefully was the question of digital colonialism. When I raised the concern that African organisations are repeatedly told to trust international companies with their data and their skills development, he didn’t deflect. He described a South African customer who’d been advised to build a dual-vendor strategy, one US and one non-US technology provider, as a hedge against geopolitical risk. Cisco’s response was a legally binding document guaranteeing no back doors and no remote access, removing the need for a second network. “I’m not going to tell a customer this is the right way to do it,” he said. “I’m going to give them choice.” That’s more considered than the standard vendor pitch, and the SCI portfolio’s air-gapped configuration option reflects it.
Where Thomson’s four-part framework, time, technology, trust and talent, gets genuinely difficult is the fourth element. He acknowledged that technological literacy constraints on the continent are structural, not cosmetic. South African enterprise teams have demonstrated they can work around infrastructure constraints that would stop most IT departments cold, but Cisco’s own research found that 78% of South African companies don’t have enough skilled cybersecurity staff to manage current AI-driven threats, with over half carrying more than 10 unfilled roles. That’s the talent gap in practical terms, and it’s not a new finding.
According to SAP Africa’s June 2025 report, the continent’s AI sector is valued at around $4.92 billion, representing roughly 2.5% of the global market. Cisco’s Networking Academy has trained 1.6 million learners across the continent, which is meaningful at scale, but skilling initiatives tend to run several years behind the deployment curve. The infrastructure African organisations are being asked to run, agentic AI platforms, post-quantum security frameworks, sovereign on-premises deployments, requires a level of specialisation that the regional talent market is still building toward.
Thomson’s point isn’t that everything is ready. It’s that the window to move is open now, and that constraints have historically produced more African innovation. That argument has earned its place in the conversation. The question African CIOs will be sitting with is whether Cisco’s most significant new products will be available to them while that window is still open, or whether “global availability to follow” means watching another product cycle from the outside.