Podcast: Things Past and Things to Come January 17 2023
Today our podcasters cover the greatly improved stock situation, the steady improvement in fundamentals, and the specific impact of both things on tech and tech companies. They also cover the need for a transformational model of work and experiences to drive growth in tech overall, how that model might look, and how we might get to it.
Remember that we’ll be doing our Mastodon chat on this podcast on January 18th at 11 AM Eastern US Time.
01-Nigel: Hello and welcome to Things Past and Things to Come, a TMT Advisor podcast. I’m Nigel Hardwicke and I’m joined as usual by Maven and Tom. If you want to learn a bit about us, you can read the introduction to the podcasters on this website.
Last week was notable in the markets, with increasing signs that inflation was being tamed, a growing consensus that there wasn’t likely to be a severe recession in the US, or in fact any recession at all, and a positive swing in stocks.
At the same time, we’re confronting some indicators that the whole economy is entering a new phase, and that some new model of financial policy may be needed to deal with it. None of this is a surprise to us; we’ve been saying it all along.
For tech, the biggest question may well be how that sector will respond to these broader shifts. What people want seems to be trending more toward experiences than to goods, but tech is the primary vehicle for experience delivery. How will that impact both the consumer side and the production side, and all the infrastructure pieces? A lot of questions for sure.
Maven, why don’t you kick things off with stock comments.
02-Maven: Thanks Nigel. I’m Maven Lightheart and I want to welcome you too. Last week was a rare one, where stocks had managed to put together a bit of a rally early in the week. On Thursday, the CPI data came out, and it showed December inflation right where the consensus of economists had predicted. Markets took that in stride and closed up again, but on Friday some mild-recession comments from banks (who reported results that beat estimates) sent futures down sharply, and the market opened down as well. But in an hour most of the losses had been erased. )(*&)(*]
What I believe we’re seeing here is the tail end of classic short selling. Short interest is running almost a trillion dollars and profits on short positions are over three hundred billion dollars. None of the hedge funds want to walk away from what’s been their biggest success, but short selling hits the rocks when buying conviction develops in the market. We’re seeing buying conviction now. Short-sellers are trying to step on rallies by selling on anything that even looks like bad news, but the buyers are coming back in. If there’s a decisive upturn, the short-sellers could lose three hundred billion covering their shorts when prices are higher than when they borrowed the shares.
We’re not out of the woods yet, of course, but what’s developing is what we said would develop, and we remain confident in our projections for the balance of 2023.
Tom you plan to talk a bit about tech, right?
03-Tom: I do, Maven. I’m Tom Nolle, president of CIMI Corporation, and let me start by reminding our audience that we’ve said we believed that tech was oversold, and that stories about the collapse of tech were nonsense. Well check numbers end of day Friday for the week, the NASDAQ closed up almost three and a half percent, the S&P 500 was up a bit over 2% and the Dow a bit less than 2%. That doesn’t sound like a sector in collapse to me.
At the fundamentals level, we’ve said all along that tech, meaning the NASDAQ, fell at least double what fundamentals would justify, so if it lost roughly 40% then it has a 20% upside even if there was no real improvement in company performance, meaning earnings. And here’s the thing. Tech is oversold, so expectations for the sector are low. That means it will be easier for companies to beat estimates. Not only that, tech layoffs and other cost-cutting measures are going to improve earnings in the first and second quarters of 2023. Put those two things together and guess what you have? The potential for a short trap. Maven, comments here?
04-Maven: You know I want to! You’re absolutely right, and we’re starting to see stories that are forecasting problems for the short-sellers. Of course we said that every earnings season was a problem in an oversold market, and that’s been true. This time could be worse for them because of growing buyer conviction. As long as hedge funds that were into “long” investing, meaning betting things would go up, knew short-sellers would drive prices down, why buy in? Wait and make more when stocks go up. But if it starts to look like short sellers could have to start covering, then stocks will go up and those long buyers need to lock in their positions. That’s what creates short traps.
But lets get back to technology, and look at what we can expect. Go ahead Tom.
05-Tom: Sure. I still don’t see any signs of a collapse in consumer tech. Yes, there are indications of some conservatism, but not a collapse, and as we’ve been saying, it looks like even that conservatism will be over in the May timeframe. In business tech, I see exactly the same thing. Budgets for tech in 2023 aren’t up as much as they were last year, but they’re still up. Yes, there will likely be some pushing of projects toward the second half, and that’s likely to make Q1 the most problematic for business spending, but again I think it’s clear that will ease in the May period. The big imponderable for 2023 is the political deadlock in Washington. An impasse on budgets or debt ceiling would derail progress.
Ad sponsored services for consumers are going to be under pressure in the first half because advertisers are holding back until economic issues show signs of resolution. However, those signs could emerge even a bit before May, and no company will take a risk of losing market share because their competitors are advertising and they are not. That doesn’t mean that ad sponsorship is out of the woods; I’m hearing that advertisers want better deals and better proof of effectiveness in advertising.
The cloud is in a similar situation. As long as consumer spending is under pressure, there’s less need to grow cloud services. Not only that, companies now believe they’ve overspent on the cloud, which they have, and so there’s counter-pressure to reduce cloud costs. I think multi-cloud is an area that’s particularly likely to see problems, because the truth is that companies have to spend a lot more for multi-cloud, and there’s growing consensus that it’s not really buying them anything. I think the cloud will hit the bottom of its growth rate in roughly March, but I think that 2023 will still be off the pace of even 2022 in terms of cloud spending growth.
06-Nigel: Unless there’s a new paradigm for the cloud as you pointed out last week. But it seems to me that there are a lot of areas where a new paradigm is needed. Do you agree?
07-Tom: I sure do, Nigel. In fact there are two areas.
The cloud is the first. You can improve application quality of experience and effectiveness using hybrid cloud technology to modernize your user interfaces to existing applications, but all this kind of stuff tends to deliver most of the benefits early on. We’ve recognized that a cloud paradigm modernizes the user interface by removing data center limitations and inertia from GUI development. The cloud is a different compute model. Now we have to ask whether that different compute model could deliver a different kind of application altogether, one that links IT to workers in a different and more intimate way.
The second area is network services. The model we use for network-building today draws way too heavily on the old days of wireline and analog voice, a world where connection was the service. Today experience delivery is the service, and increasingly experiences that are both valuable to the user and profitable for the provider are hosted less than twenty miles from where they’re consumed. Today’s networks are focused on hierarchical traffic aggregation. Tomorrow’s networks have to be focused on experience hosting and delivery. That’s why I believe that the “metro” space is the sweet spot for the transformation to a future network model.
Metro is important because of my third area, contextual services and point-of-activity empowerment. Right now, both workers and consumers have to build their activity around information access, experiences. If you need something, you have to ask for it. That limits the value of information because it’s harder to integrate that pull model into our real world. What we’d like to have is delivery of relevant information to us when we actually need it, without asking, which means we need services that understand what we’re trying to do, what we need to do it.
08-Maven: Isn’t that what Meta wants to do, with what you’ve been calling their “social metaverse”? I thought you believed that wasn’t a viable model, or at least wasn’t a transformational one.
09-Tom: I don’t think a social metaverse is transformational, and I’m not sure it’s even viable. The problem, if you go back to my contextual services comment, is that we’re creating experience relevance by taking what we experience out of the real world. I think we have to go the other way, meaning bringing experiences based on information technology into the real world.
09-Nigel: How would your metro and contextual focus change networks overall? Isn’t a lot of the network’s architecture set by the need to handle user connection and traffic efficiently?
10-Tom: A lot of it is, all of has been, but maybe it shouldn’t be. If we assume that the Internet is the network, and that its goal is to deliver experiences, then users have to be connected to experience hosts. That’s what I think metro points represent. There’s anywhere from a dozen to perhaps four hundred ideal metro sites per country, depending on population and economic density. We can’t mesh users, but we could mesh metro sites, particularly if we allow perhaps two transit hops. In the US, the ideal metro structure would tie 250 metro sites back to 9 regional points then mesh those regional points.
The same thing is true on the access side. All we want is to get to a metro site with the minimum of latency. So as soon as you hop onto fiber your strategy should be to use the maximum capacity the fiber can handle, and aggregate with minimal latency. It’s as close to all-optical as you can make it.
11-Maven: OK, that’s interesting, but is it something that could really change network architecture? Networks are high-inertia structures, after all. You can’t rip out the old at every turn.
12-Tom: They are high-inertia and you can’t turn them on a dime; more like tens of billions of dollars, in fact, because you’d have to take a write-off of any residual value in the old stuff! The pace you can realistically expect to implement changes depends on how fast you’re already budgeting for modernization and what benefits you can draw on to reduce the impact of those write-downs. If optical capacity is indeed cheaper than capacity management and metro-level meshing with few hops is indeed practical (both of which I think are true) then we may have reached the point where a metro-centric re-architecting of networks could be justified. If that’s true, it might offer an on-ramp to edge computing based on metro hosting. The problem with edge computing is that you need a lot of pre-investment, which operators call “first cost”, before you can start to sell services. In fact, before you can expect anyone to even think about those services. Who plans to use what isn’t there yet?
13-Nigel: Would your metro model necessarily support the growth of edge computing, though? It seems to me that you could create that kind of metro-forward-metro-backward mesh without any edge computing? If there’s no hosting in the metro area, would there be any real impact on the edge computing space?
14-Tom: Good question. I think there are two forces here. One force is that if future services are to be, like present services are, fundamentally delivered experiences, then those experiences have to be created somewhere, and they need to be created through hosting software elements. One way or the other, I think we can expect to see the currently disconnected content cashes and feature hosting points get organized. Where? Where access latency is low and where cloud and edge interconnectivity can be high. Metro fits the bill. It’s not a slam dunk, but I think it’s a likely outcome.
Metro and the edge, I believe, are the foundations of contextual services. If there’s any reality to the metaverse beyond IoT and contextual services, then it’s the foundation of that too. Getting it in place could transform business IT and consumer behavior, making it the most transformational thing since the Internet. But we need to work hard to get it, and right now we don’t have a clear path to making that first-cost thing bearable. That could happen if we redid the networking model, justified by the reduction in opex that substituting capacity for capacity management could give us.
15-Maven: So metro is critical to the edge in two ways. First, it’s the logical place to do what we’re calling “edge hosting” versus the real network edge, which would be too distributed to be efficient. Second, it’s a logical focus for network change, a change to use capacity rather than capacity management to lower opex, and even potentially transform capex. That all makes sense, and I hope we can make it into a reality soon. Otherwise we may see subsidies of telcos by OTTs become a global reality.
That concludes our podcast. Thank you for listening and we hope you’ll join us next week.