Why does it seem that every market opportunity, technical trend, adoption rate, ends up graphing as a hockey stick? I’m sure you’ve been hearing stories about the 5G “slowdown” and some vendors in the space cutting jobs. Well, gosh, 5G is a wireless network evolution, a conversion from a prior standard. Like any conversion, it starts slow, picks up pace as critical support levels in related technologies (handsets in this case) develop, and tapers off once the majority of convertible sites have been, well, converted. Why then don’t we graph the activity along those lines, and expect the technology to eventually plateau and fall off?
One reason is that the hype cycle demands something to focus on. There are some trying earnestly to push 6G into the limelight, or WiFi 7, but for a lot of sites it’s 5G that has to draw interest. Not only that, 5G may be mature as a technology, and may be seeing a falling off, but a lot of those involved with 5G are eager to see stories that suggest that there’s no place to go but up. And guess what? The problem isn’t limited to 5G, as we will see.
A decade ago, I got a call from a reporter, who asked me what I believed the incremental 5G revenues for telcos to be, expressed as a multiple of then-current revenues. A qualifier was added, though. “Famous Analyst Firm says 250%”. For those not tuned into the subtleties of reporter-speak, that means “I have a number from someone else for my story, but if you can beat it, you’ll be in the story instead of them.” Needless to say, I didn’t get in that story, but the tale illustrates the problem; “news” means “novelty,” and that’s one reason we get hockey-stick surprises.
All markets saturate. People buy stuff because it enriches their lives in some way, and they stop when the extent to which they’re enriched by the thing they buy is offset by the extent to which they’re impoverished by the spending. Business buy stuff because it’s going to boost their bottom line in some way, and stop when there’s no longer such a business case. Some say a market is a buyer, but that’s not true. A market is a buy decision, and that’s a whole different thing.
If we want a realistic view of technology progress, we have to start with the purchase-justification piece of the puzzle. Why is a buyer doing this and how long will they want to? Decades ago, I did organized user surveys on technology adoption. One of the things I found was that while buyers were good at predicting what their current technology budgets would be used for, if you went three years out there was only a minimal correlation between what they said they’d do then and what they ended up doing. That’s why I started doing forecasts based on a computer “demand model” rather than based solely on survey opinions.
It’s also true, for both business and personal purchases, that there’s a shift in the value proposition over the life of a product/service. Early on, when something is novel, there’s lower resistance to buying because of the novelty, in the consumer space. In the business space, early purchases are usually targeted at “low apples”, meaning applications that most easily make a business case. Willingness to pay is higher then, but as the lifecycle advances it becomes more difficult to justify a purchase and prices tend to fall. Thus, profit margins for suppliers tend to decline over time unless production economies can offset price shifts.
My model has, over time, shown that the peak spending growth for a technology is roughly 34% above long-term baselines. Based on that, it would be my view that cloud computing has likely hit and passed its peak growth rate, and of course 5G has as well. What would be needed to reignite spending growth in both cases would be some new business cases.
I raise cloud computing here, in what you may have thought was a 5G rant, because cloud computing is at its core a conversion just like 5G is. Businesses have IT needs, and traditionally those needs were satisfied in the data center. As application requirements changed, though, a class of applications emerged that demanded greater geographic ubiquity, greater elasticity in terms of performance, and perhaps specialized hardware or software. These applications drove cloud adoption, part of which took the form of shifting proto-versions of those more demanding apps that had been run (unsatisfactorily) in the data center, and part of which came from the initial implementation of applications that couldn’t make their business case within data center constraints.
Right now, there’s a lot of buzz that AI will somehow push cloud spending upward again, and that’s something that cloud providers, hardware vendors, and the tech media are happy to support. The challenge here is again the challenge of the business case, the justification. There’s a lot you can do with AI. How much of that “a lot” is actually something that pays back on an AI investment is another matter. The fact is that we don’t know whether AI can make any significant business case at this point. Yes, it can write articles or PR, but so can any number of junior writers who don’t get paid all that much. Yes, it can perhaps diagnose medical conditions, but unless that creates a positive outcome from a financial perspective, likely requiring a loss of jobs in the professions, that ability may not pay off. Could it lower insurance rates by improving outcomes? Sure, but how does that trickle down to the level of the people who have to pay for the technology? It’s complicated, which means it’s time consuming.
It’s also true that examples don’t make a market. Just because we can find things we can do with 5G or the cloud or AI doesn’t mean that it would be profitable for us to do them, and profitable for suppliers to offer them. And an example can’t be itself a technology that needs justifying. You can’t justify 5G by network slicing unless you can justify network slicing. You can’t justify AI by image analysis unless you can show why somebody would pay a decent buck for it. A feature is not a market, nor is a technology. A market is a buyer with a reason to buy and a seller who can make a profit on exploiting that reason.
Tech is clearly undergoing a financial reassessment. The reason for that is not that it’s become jaded, or that Silicon Valley has lost its innovative spark, it’s that we’ve lost sight of the concept of a market. That’s bad news. The good news is that there is plenty of market out there, waiting to be found. We have the potential for billions in new cloud revenues, billions in AI revenues, and even billions in mobile network revenues, that could be accessed if we started our search for technologies with the ultimate buyers who have to spend to create it.
