Three weeks before the World Cup I find myself on yet another business trip to Beijing. As I peer out the window of my hotel room, the "Bird's Nest" sits bright, loud, and proud in remembrance of Beijing hosting the 2008 Summer Olympics.
Well, that was six years ago, and what I am now looking at is a hollow shell and infrastructure collecting dust. Sure, there are a few visitors here and there, but is this all this magnificent structure aspires to be?
All this got me thinking about the World Cup and the upcoming 10 nm fabs.
Deep in the Amazon, a stadium just finished a $200 million renovation plus the cost of building the surrounding infrastructure to accommodate the expected throng of spectators. Brazil has 11 others either built from the ground up or massively renovated, costing the country roughly $3.7 billion, three times the estimated cost -- not counting the estimated $12 billion spent on the infrastructure such as hotels, airports, roads, etc., to provide public services for the anticipated events.
This is more than double the highest amount ever before spent on the event: the 2006 World Cup in Germany. The Brazilian government is hoping for a $70 billion to $110 billion return over the next 10 years, but even at $1.8 billion spent, Germany is finding that it only saw an increase of €60 million in tourism the year it hosted the World Cup. South Africa still cannot produce a number on the economic impact of that country's hosting the 2010 World Cup.
How is the economics of hosting the World Cup justified when the stadiums will be left empty after the World Cup championship finishes?
The Olympics, Winter Olympics, and World Cup used to be great catalysts to encourage infrastructure investments for cities. But globalization in the past 30 years has seen two opposite forces colluding to skew the economic rationales.
First, globalization has encouraged the IOC and FIFA to award their games to smaller cities in the hope of development (Sochi just before, and the current hot topic is Qatar). Second, globalization has also super-sized these sporting events. As a result, each event draws more attendees and requires grander-scale construction. I am sure every Brazilian soccer stadium will be packed for the next couple weeks -- but what happens after this short but festive period of world sports ends?
There is a similar story with the semiconductor industry.
The mobile industry needs to build smaller, lighter, and thinner phones. This contributes to the escalation of fabs. TSMC, GlobalFoundries, Intel, and Samsung will be the high rollers spending $5 billion on average per fab to manufacture the 10 nm projects.
How long will it take, and how many mobile devices will need to be sold to make a return? Assuming the four giants will build six fabs in total, that'd be $30 billion in capital expenditure. Conservatively, if capex depreciation accounts for 15% of the total cost of goods sold (COGS) over the period of the fab, then it means $200 billion dollars worth of wafers will need to be manufactured, mostly for baseband and application processors.
At an average of $40 per mobile phone sold, my back-of-envelope calculation says we are about to produce 50 billion phones. Is it possible? One has to be an optimist to be in the technology business. But even if we can accept this on faith, the issue remains: What happens after the initial two to three years of smartphone use?
The semiconductor industry has benefited from a series of new market adoptions, from military and government to computing to the Internet and now to the mobile phone.
The commoditization and miniaturization have meant escalation of fab cost in order to meet such requirements. I am highly confident that we will continue to find new ways to use semiconductors to enrich our lives.
Whether enough new markets require the 10 nm fabs is what I'm wondering. The optimistic me says that, for example, because we haven't truly accounted for the cost of energy and environment, the miniaturization movement has focused on convenience and not on conservation. So the cloud side will surely need to catch up, especially with the industrial side of the Internet of Things.
But the worried side of me recalls an old and Euro-centric but still relevant book by Paul Kennedy, The Rise and Fall of the Great Powers. It basically says that great powers fall because they overstretch their economies in unprofitable investments under false assumptions.