Think of this as Volume 18, Number 7 of the newsletter I have written weekly since March, 1997. Enjoy.
Last year I wrote a lot about the “barbell” developing in computing.
The computing middle class of PCs and servers were being replaced on the one hand by cheap devices and on the other hand by clouds delivering software and services.
Recent events have shown that, from the U.S. perspective, this barbell has only one side.
That side is the cloud.
Google’s decision to get rid of Motorola, along with IBM’s decision to send its server unit to the same company – China’s Lenovo – pretty much ends America’s reign in computer hardware. The marketing and branding of PCs, servers, and devices are now all going to be done in China, which is good for consumers, however bad it may be for the national interest. (Don’t cheer for the Chinese. Their workers will face new threats from lower-cost labor markets from Bangladesh to Africa. In a race to the bottom, there are no real winners.)
Apple is thought to be an exception to all this, but is it really? No, it’s not. Apple markets devices, and it designs devices, but all those devices are built in China, mainly with China-made parts. Investors no longer value Apple’s device profits highly, and there is a limit to how many, and what type, of devices Apple can produce from here. There will be some American-made Apple kit, mainly produced for public relations purposes, but the truth here was written by Google, which was recently selling a U.S.-made Moto X phone for $400 without a contract, and virtually the same Chinese-made model for $200.
When it comes to mass manufacturing in electronics, we can’t compete. Anything that requires heavy use of manual labor is going to the place with the lowest hand labor costs. We need to move on, and move up the stack of skills.
All this has created, even accelerated, the most important business trend of the year, which I call the cloud gold rush.
During the last quarter Google put over $2.25 billion into new server farms. Amazon spend nearly the same amount. Microsoft, IBM, Facebook and Apple are all under pressure to respond. They will, because they face a choice between either spending billions on their own server networks or finding themselves, two years from now, using someone else’s, and paying that competitor for the privilege of survival.
The appointment of Satya Nadella as Microsoft CEO, following by a week Microsoft’s appearance at Facebook’s Open Compute project meeting, where it showed (and basically open sourced) its system for server management, puts all this into perspective, and shows where we’re going. The companies left out of the Google-Amazon axis are getting together, sharing design costs, and working to build compatible server farms around the world, using the Google-Amazon cost structure as their template.
Over the next two years this is all great news. It means that prices for cloud infrastructure will continue coming down. It means ad-based Web services built on cloud infrastructure can plan for enormous growth without seeing an explosion in costs. All of this means that, despite falling ad rates, these services should continue to make money.
It also means that, despite the end of net neutrality rules, there will be enormous pressure to keep costs in the last mile down, in order to deliver this abundance to consumers. Cable and phone companies are treating Internet as another form of pay TV, hoping to avoid cost pressure from popular content by charging other providers, but the Internet is not TV. TV is just one file type.
The investments being made in core delivery technology, both in the form of new server farms and content delivery networks (like the one Netflix already has and Apple is building) are going to make TV appear as a relatively slow file type over time, just like telephony is now. The pressure of the cloud service providers, their need and desire to reach customers, will not allow the last-mile providers to get the monopoly rents they think they’re in for. Those companies will either yield through the market, or they will yield politically. But they will yield.
The problem with the cloud gold rush is that every boom is followed by a bust, and this one will be no different. At some point cloud server capacity will dramatically exceed demand, to the point where no matter how little it costs, building new capacity becomes uneconomic.
Right now, it’s beginning to look like this abundance will crest right alongside the coming energy abundance, in 2017. Energy prices are due for a crash as soon as renewable costs fall below the costs of fossil fuel energy – they’ve already passed the price of such energy in many places. The cloud gold rush ends when there aren’t enough customers to pay the cost of building new centers, and this will happen as inevitably as the present boom does.
What’s funny about the next recession is that it will be the product of abundance, not scarcity. Energy abundance and Internet abundance, when combined, are going to remake our world. But there will be some economic pain along the way, an immense bulge of deflation that policymakers will have to answer. The likelihood of leaders bumbling through it, rather than acting decisively, is high because they’ve seldom acted decisively before, and because this problem will be materially different from those that came before.
What follows this recession of abundance, of course, will be a new form of scarcity – labor scarcity. Kids who are having such a hard time getting their lives started now are going to be in high demand once the next recession lifts, because intelligent labor will be in short supply when compared with other economic inputs. New types of jobs will be created, new types of companies will be created to organize that work, and your kids – along with my kids – will be all right.
But as usual, it’s going to be a bumpy ride. And the cloud gold rush is just one indication of how bumpy it will be. At least one company, and probably more, that now looks impregnable is going to go completely under some time in the next three years. Will it be IBM? Hewlett-Packard? Dell? All three? And who else?
I can’t answer that. I see trends, not specifics. What happens to any company depends heavily on what its leaders do in response to constant, accelerating change. Can Satya Nadella ride the current changes to new prosperity for Microsoft? I would say the odds are in his favor, but there’s no certainty, anywhere. Not really.
That’s what gives life its flavor.