John Chambers, Tech History, and AI and Cloud
John Chambers, former Cisco CEO and current venture capitalist, was interviewed recently by Investor’s Business Daily. He offered some perceptive observations, including one that I will explore in detail, but first allow me to take you down memory lane for a short diversionary stroll.
When I joined Cisco Systems in the mid 90s, John Chambers was the CEO. At the time, I was one of many recent hires who were involved in a training regimen that included an executive audience with Chambers. Dozens of new hires were put through the corporate induction process in a nondescript room within one of the many nearly identical buildings on Cisco’s sprawling San Jose campus. Chambers appeared before us to talk about the company culture and to provide his perspective on Cisco’s market opportunity (which was, as the record attests, of considerable size in the mid 90s).
Chambers worked the room as only he could – he always worked the room superbly, whether he was at a convention centre, a customer site, or in a training lab – and he held our attention for the entire session. With no need for a microphone in such a small setting, he had leeway to walk up and down the aisles, looking each of us in the eye as he emphasized various points.
I was impressed by his seemingly natural and effortless ability to communicate cogently and to connect personally with an audience. It’s a rare skill; few can do it truly well. As I took in the spectacle, Chambers seemed to me an uncanny blend of superb salesman – he was steeped in a sales background – charismatic motivational speaker, and ardent evangelist. His religion was Cisco, and his mission was bringing Cisco’s networking products to customers everywhere. In speaking to us, he was proselytizing to potential acolytes.
John Chambers is also an astute historian of the technology industry. He’s lived a lot of that history, starting his career at IBM and Wang Laboratories – you kids might have to look up that last one – before finding fame and fortune at Cisco. His timing in making that career move was impeccable, joining Cisco just as the Internet era was about to catapult routing and switching, cornerstones of computer networking, into valuable elements of business connectivity.
In his recent interview with Investor’s Business Daily, Chambers delved into the history of the information-technology industry; he did so not in a reverie of sepia-tinged nostalgia, sentimentally reliving the glory days of yore, but to understand where we might go next.
I don’t mean to get mystical or New Age on you, but an examination of the historical record makes sense. Time is a flow: the past was once present, and the future is an unrealized, latent part of the flow. If you can understand the rhyming cadences of the flow, you can get some inkling of where we might find ourselves. Nobody knows exactly what’s around the corner, but a careful analysis of the historical record can help us gain insights into potential probabilities and outcomes.
Breaking the Generational Cycle
You can read the entire Investor’s Business Daily Interview here, but I would like to call your attention to the part of the discussion where Chambers expounds on potential winners and losers in the forthcoming battle royale for AI riches. Here’s the salient quote from Mr. Chambers:
Traditionally, the incumbents who led in one phase have done OK in the next, but they never led it. IBM mainframes, DEC, Wang, minicomputers, PC players, Compaq, the HPs, the Dells. Then the cloud: Amazon, Microsoft, Google were stronger.
Every time, it's been the new players who were the dominant players in the next generation. It was rare and nonexistent where the leader in the prior one was one of the top three leaders in the next one, five or 10 years out.
This might be the one exception. If you watch the nimbleness of Microsoft and Google and Meta and Amazon, they are all focused on this in a big way. With OpenAI changing the market, they realize they could get disrupted and they are the most economically powerful tech companies we've ever had. So I think one and maybe two of them will end up being one of the top three or four here.
Normally, the leaders of the prior generations weren't the leaders of the transition. This time we might actually see one or two make it through that.
Chambers has a point: the leaders in cloud are well placed to reap further prosperity as AI accelerates into full commercial stride. That said, I’m not sure that the cloud giants’ AI success will be predicated entirely on their nimbleness, flexibility, or innovation. Those ingredients are necessarily part of any complete AI recipe, but they alone will not ensure that the globe-girding cloud hyperscalers thrive as mainstays in an era redefined by increasingly sophisticated AI.
False Dichotomy
Unfortunately, a false dichotomy has arisen that posits cloud and AI as somewhat mutually exclusive. The truth is, cloud and AI go together like chocolate and peanut butter (allergies notwithstanding).
In key respects, AI represents the next wave of killer applications that will run in cloud datacenters and cloud edge environments. We have discussed this theme before, including this recent post. I see no reason why AI should trigger a counter-revolutionary recrudescence of on-premises enterprise datacenters. Given costs, complexities, learning curves, and a multiplicity of data requirements, the case for doing AI in on-premises enterprise datacenters is questionable.
There is no contradiction between cloud and AI. Clouds are not only capable of running AI, but clouds are probably the most logical environments in which to run most types of AI. The largest clouds possess the capital, expertise, resources, and scale that are prerequisites for long-term AI prosperity.
Yes, cloud giants – Google, Microsoft, Amazon (AWS), Alibaba, and Meta on the B2C side of the divide – must continue to innovate and stay ahead of the adoption curve, but, as the steady drumbeat of news confirms on a near-daily basis, that’s exactly what they’re doing. They’re not slackening the pace of in-house development or external investment. If anything, they’re upping the AI ante, making it difficult, though not impossible, for upstarts to achieve meaningful displacement in the cloud firmament.
A New Chapter, Not a New Book
Even recent history is instructive. Edge computing neither killed nor materially wounded cloud computing, despite the fevered predictions of some pundits; nor did an enormous wave of workload repatriation materialize to turn the enterprise tide against cloud migration. The telcos – who struggle mightily even to make the transition to cloud-native application environments and infrastructure – haven’t done much to weaken cloud’s grip on enterprise patronage. It would take a leap of faith and an omission of logic to believe that telcos will suddenly achieve primacy in AI.
Cloud market leaders have been at or near the forefront of AI research initiatives for a long time, and there’s no reason to think they’ll suddenly lose the plot or get derailed by a regressive shift back to on-premises workloads. I don’t see how AI takes us back to the future.
That’s not to say that some AI workloads, for reasons of industry compliance or company-specific requirements, won’t run in on-premises enterprise datacenters, edges, or co-location environments. No market is entirely uniform in its demand dynamics or disposition, and there are always exceptions to every rule.
Nonetheless, the applications and data already in clouds, and the need to combine disparate data to generate desired outcomes, suggests that cloud, with its global reach in nearly every major center of international business, will remain at least as relevant in the future as it is today.
The cloud giants have gained proficiency and scale in the global arena applications and data. They possess the knowledge and means to make data useful and actionable, and they have a significant competitive advantage over those who possess only piecemeal capabilities.
If the next wave of industry giants will be largely the same as those who dominate the technology universe today, it will be because of the affinities and mutual reinforcing tendencies of AI and cloud.