I couldn’t put it off any longer so here are my technology predictions for 2017. I’ve been reading over my predictions from past years and see a fundamental change in structure over that time, going from an emphasis on products to an emphasis on companies. This goes along, I’d say, with the greater business orientation of this column. That makes sense with a maturing market and mature industries and also with the fact that a fair number of readers are here mainly as investors, something that didn’t used to be so much the case.
Of course we begin with a look at my predictions from a year ago to see how I did. Almost nobody in my line of work does this, pointing out their own mistakes, but then I always have been kind of stupid about my career. So here we go.
Last year I predicted the beginning of the end for engineering workstations. My rationale was that virtualization was going to move a lot of this work into the cloud, technical software companies would be converting their products into services, and that it would all be enabled by the addition of GPUs to cloud servers. This is going a little slower than I had thought but is definitely happening. It’s good, though, that I only claimed it was the beginning of the end. I’ll claim this as correct and I’ll even have more to say about it for this year.
I predicted that both Microsoft and Apple would hit walls in 2016. I’d say there is no doubt that Apple hit something like a wall and if you’ll go back to the actual column you’ll see what I meant about Microsoft was that Windows 10, while a huge success in terms of upgrades, would be an ever smaller component of Microsoft’s success. I’ll claim these both as true.
I predicted my Steve Jobs movie would return to Netflix, which it did though not until the middle of the year. Look for it to stay there until the middle of 2019, by the way.
I predicted that drones and driverless cars would not really hit the market in 2016 despite all the hype. This is certainly true about cars but if you’ll check the actual column you’ll see I was talking specifically about outfits like Amazon using drones for commercial deliveries. Those have happened, sure, but you can count them on one hand so I’m claiming both predictions as correct.
I predicted the Internet of Things would become a security nightmare and that definitely happened when home routers began to turn into zombies. It’s only going to get worse, though I’ll hold off on predicting specific remedies until a year from now because it will take that long (or longer) for the pain to become large enough to help the smoke clear. This will be a major prediction area for me in 2018.
I predicted Apple would not buy Time Warner (AT&T did) but said Apple would buy the Dish Network, which they did not, so I finally got one wrong. My thinking here was that Apple was too paralyzed to buy Time Warner but when AT&T bought DirecTV it seemed like Apple might be inspired to buy the other satellite operator, Dish, though as much for its wireless spectrum as for the satellite operation. I still think Apple should have done it. One right, one wrong.
Finally I predicted that Intel would start to become irrelevant, which is definitely the case. What processor is in your PC? You don’t know, do you? More on this below.
So by being vague I managed to claim 90 percent correct. Maybe I should retire with that? Not a chance!
For 2017 I am taking a Big Picture approach to my predictions. Where last year there were 10, this year there are only six but they are all enormous and interconnected. Read each one and think about the others. There are sub-predictions I could have made by looking at the likely impact of the big predictions on established companies. For example we’ll see a continued decline in the role of the standalone PC. I think PCs are going to disappear completely from government, thanks to Edward Snowden, starting in 2017. I suppose that could have been a prediction. And with PCs declining and the cloud booming we’ll see an accelerated transition to software-as-a-service (SAAS) from all major software companies. While that, too, could have been a prediction, why bother since it is so obvious? Here are the big ones.
Prediction #1 — A Cloud Arms Race If 2016 was the year of the cloud then 2017 will be the year of the CLOUD!!! Amazon, Google and Microsoft have all staked their corporate futures on their cloud services and are throwing investment dollars in that direction. We’re talking about major infrastructure upgrades measured in multiple billions per year per company. Like last year a lot of this has to do with the introduction of high-end GPUs into cloud servers but this year there is a second provider — AMD. As always, competition will push prices down and performance up. Amazon just made a HUGE buy that will be matched or exceeded by both Microsoft and Google. There is no practical limit to what these three companies will pay to get and stay ahead. This can only be good for users of cloud services.
It can only be bad for cloud competitors. IBM, for example, simply doesn’t have the firepower to compete on this level so they’ll try to redefine the playing field as a hybrid cloud where IBM customers are the ones making the big capital investments. This will keep them in the game but not as a major player.
The dark horses in this arms race are Apple and Oracle. While Apple seems content to develop its own cloud services, the company clearly has the deep pockets and data center capacity to do much more. But will they? There’s a prediction about that below. As for Oracle, Larry Ellison has been making Trumpian pronouncements about how Oracle’s success in the cloud is going to be YUGE. If he doesn’t deliver soon, it won’t be.
Prediction #2 — The End of Bufferbloat You may recall a few years ago I included bufferbloat in my predictions and it was probably the first time most readers had even heard of this problem that afflicts most ISPs and home networks. Well the technical work has finally been done to kill bufferbloat with superior Open Source router firmware. Router manufacturers are finally figuring out that Open Source is cheaper and better. And ISPs are almost ready to kill bufferbloat outright as explained in the next prediction. It is going to happen and the result will be noticeably better streaming audio and video services. This is not only a big deal, for most users it won’t even require new network hardware.
Prediction #3 — The Beginning of the End for U.S. Broadcast (and cable!!) Television There are a lot of moving parts to this one and it won’t all come together in 2017 but enough will happen to make the endgame obvious. I laid-out a decade ago where Internet Service Providers were logically headed in the USA. That is toward becoming bit schleppers and landlords for over-the-top (OTT) video services like Netflix. The necessary antecedents for this to happen are: 1) a market for broadcast license holders to sell their spectrum for mobile use; 2) a rise in OTT providers that directly compete with traditional cable TV companies, and; 3) the death of net neutrality.
The FCC is already operating reverse auctions to facilitate the transfer of spectrum in exchange for up to $100 billion. Most of that money will go into the pockets of investors through share buy-backs but some will be used to jumpstart next-generation entertainment services. OTT services are booming and the critical event that happened only a few days ago was when CBS decided to participate in the live OTT service to be introduced this year by Hulu. That will give the Hulu service all four U.S. broadcast networks, the first time that has happened, and it’s a game-changer. CBS’s move would not have happened when it did had not it been clear that the Trump Administration intends to kill net neutrality, which will allow ISPs to charge OTT providers for access.
Explaining this OTT/ISP connection deserves its own paragraph. Let’s use as an example Comcast, though this could apply to almost any large ISP. Comcast is America’s largest ISP by far as well as being America’s largest cable TV company. But if you dig through the Comcast financials you’ll find that not only does the company make more profit from providing Internet service than it does from selling video service — Comcast pretty much makes no money at all selling video. The fees Comcast pays to cable and broadcast networks and independent stations take all the money. If Comcast could just get out of the video carriage business and become strictly an ISP they’d make more money with less hassle as a smaller, simpler organization. But the psychological prerequisite for this happening is finding a new revenue source made possible with the demise of net neutrality, because companies have a hard time selling change that only makes them smaller. With net neutrality out of the way (understand I’m not taking a position one way or the other on this, though I personally tend to favor net neutrality) Comcast can go from paying for video content to being paid for video content — a fundamental change of business model. They’ll still own NBC and all the cable channels, but those channels will eventually be available strictly OTT. This will take 2-3 years to sort out but once net neutrality is gone the result is pretty much inevitable.
Going back to bufferbloat, once the ISPs can charge the OTT networks then the OTT networks can demand better service and that’s the moment when bufferbloat will instantly disappear.
Prediction #4 — Intel spins-off its fab. Intel is in trouble. Yes the company is huge and rich and profitable but it also doesn’t really understand the new world in which it is trying to operate. Intel missed the mobile wave and I don’t think can invest fast enough to catch it. Intel has no advantage in the cloud. Intel graphics can’t compete with either nVIDIA nor AMD. Intel literally threw away its position in ARM processors and ARM is now Intel’s greatest threat — a threat I don’t believe it can beat now that the hyper-aggressive Mr. Son of Softbank owns ARM Holdings. So I predict Intel will this year spin-off its fab, following the example of AMD’s old fab, Global Foundries. This will appeal to investors and if Intel doesn’t think to do it we’ll shortly see activist investors do what they need to to make it happen. Better for Intel itself to do the dirty work, loading-up the fab with debt to finance the surviving company’s next stage of life. And I’m not saying this is a bad move for the fab, either, since it will be able to more freely pursue profitable merchant business. But for Intel it will be a jarring change and their path to success isn’t at all clear.
Prediction #5 — Apple makes a huge (for Apple) acquisition. Think for a moment about the Amazon Echo and Dot speakers and their very similar competitors from Google, Asus and others. Why didn’t Apple invent those products? Apple already had Siri, which defined the voice assistant and led the segment for years. What is Alexa but Siri’s sister? There is no reason at all why Apple shouldn’t have invented that product, yet they didn’t. This isn’t a tablet or a phone where Apple could enter the market late and snatch leadership through superior design: Apple had the lead and let it slip away. Apple is in trouble.
There are two things enormous companies can do in this situation — they can fire the CEO or the CEO can do a huge acquisition that will be spun as changing the whole character of the business. The best example of this behavior is actually Apple, itself, circa 1996, when Gil Amelio, attempting to keep his job, acquired NeXT for $425 million, which was a lot of money for Apple at the time. That it led to Gil’s own demise at the hands of Steve Jobs doesn’t matter: the point of the acquisition was to change the character of Apple, which it obviously did.
Tim Cook isn’t going anywhere so a big acquisition it will be. Apple doesn’t make really big acquisitions, but it has to be bigger than Beats, Apple’s largest acquisition to date. The acquisition could be almost anything (I’m not sure it really matters) but a high margin business is better and Apple’s recent success in services suggests it will be something in that direction. Apple’s lack of leadership in the cloud, mentioned in an earlier prediction, may mean they’ll use this one to bolster their cloud cred. I can think of a couple very bold possibilities in the $5-10 billion range, but don’t feel comfortable naming names at this point. The point is to either change the world (again) or at least appear to be changing it. I just hope they don’t buy Tesla.
Prediction #6 — Come-to-Jesus time for IBM. The most important 2017 event for IBM will be the retirement at 60 of CEO Ginni Rometty. The future of Big Blue absolutely depends on the actions this year of her successor. If she follows the example of the two previous IBM CEOs Rometty will stick around for a year as chairman before flying-off to paradise in her jet.
The reason I call this a Come-to-Jesus time for IBM is because the next CEO will have a chance to do something different with the company. If he or she decides to break with the past IBM has a chance. If the new CEO takes whatever game plan Rometty hands over and runs with that, then IBM is doomed.
Rometty was handed a bad plan by Sam Palmisano, loyally tried to make it work for two years and then has been improvising poorly ever since. She should have used her mandate and changed the plan immediately. Mandates don’t last.
This is very tough for the incoming CEO. IBM’s current plan isn’t working and isn’t sustainable. You can only eat your seed corn for a short time before future harvests are compromised. IBM can only sell divisions and borrow to buy back shares and raise the dividend in the absence of real revenue for so long. Interest rates are rising and the dollar is strong, both of which are bad for this strategy. The right CEO will realize this and will break with Rometty’s plan even as she remains as chairman. If they wait a year for Rometty to depart, it will be too late.
So it could mean a bloodbath in the executive suite at IBM. Except it probably won’t. I believe the company has lost its heart. That’s still a Come-to-Jesus moment except in this case Jesus won’t be there.