On Capital Allocation

One of the paradoxes of the “post-industrial” era is the aversion to application of capital to growth opportunities. Generally speaking, capital has become trapped in bank accounts as opposed to equipment which could be used to produce value. This aversion is rooted in many dysfunctions, chief among them being the misunderstanding of the purpose of the firm.

But there are exceptions. Illustrated below are the patterns of spending in property plant and equipment (capital expenditures) by companies that still recognize that there are opportunities to be obtained by investment in the means of production.

Screen Shot 2014-08-13 at 11.36.03 AM

The highest spenders in this cohort are Samsung and Intel, both of whom spend mainly on semiconductor production equipment. Apple has been a big spender on production equipment and, to a lesser degree, on data center infrastructure. As the graph shows, Apple’s spending has been nearly at the same level of Intel but with recent volatility, suggesting shifts in production processes[1].

Companies such as Microsoft and Amazon have been increasing their spending but since their assets are mostly data centers[2] they have been considerably below the “manufacturing” investment level.

The interesting story emerging is Google. Its spending must include data centers primarily but the level is now reaching that of Intel and Apple. It actually spent more than Apple for the last four quarters. There might be a lot more need for infrastructure to support its users, but the revenues have not risen as rapidly as this line item

The question of what Google spends on is interesting. There are many projects which require capital including announced fiber networks but nearly $10 billion/yr. in capex is a staggering amount and I’m curious if its explainable as sustaining their current strategy or something emergent.

  1. I’m assuming that the 5S/5 required a surge in spending followed by a throttling back for the 5C with the iPad processes wild cards []
  2. Amazon also accounts for its distribution warehouses and web site development costs as capex []
  • KirkBurgess

    From the linked story:

    “Google’s remarkable capex increase over the last year has raised concerns among investors,” Bernstein Research analyst Carlos Kirjner wrote in a recent note. Kirjner speculated that Google may be building data centers even before the company needs them.

    That view essentially was confirmed by Google’s Chief Financial Officer, Patrick Pichette, on the company’s April 16 conference call. “In the case of data center construction, we have found that the option value of having more capacity on standby and available to us to grow versus not having it is actually a real strategic issue for the company,” Pichette said.

  • normm

    I don’t understand the use of the word “paradox” here. We had a world wide collapse of financial “innovations” that led to a collapse in demand, and companies stopped investing in new plant and equipment as a result. Governments have mostly responded with austerity, worsening the situation. A few companies that still see rapid growth in demand, such as Apple and Samsung in mobile, have continued to invest.

    We can anticipate a lot of pain as automation eliminates traditional jobs with even relatively high skill levels. But that’s a longer term issue, not the current crisis.

    • JKL

      Most governments have responded by talking about austerity, not implementing it, most have simply reduced the leval of increases.

  • highr0llerr

    “There are many projects which require capital including announced fiber
    networks but nearly $10 billion/yr. in capex is a staggering amount and
    I’m curious if its explainable as sustaining their current strategy
    or something emergent.”

    It could be both. The amount of data created around the world is rising exponentially. So, it makes sense that Google needs a lot more data centers that it needed just a few years ago. At the same time, as you mentioned, there are many capital intensive projects and a lot more “special projects” that we don’t even know about. I think, that’s been Brin’s focus for some time.

  • charly

    Google’s fiber broadband is something that would need a lot of capex. It is also something with a health return for Google even if it is a failure.

  • Innterrnett

    How certain are we that Google is buying physical infrastructure vs a simple accounting trick to capitalize and depreciate software development costs?

  • stefnagel

    Neuro fiber?

  • Walter Milliken

    Perhaps Google is investing in manufacturing automation for building data centers and/or server hardware?

    Also barely possible they might be going into the silicon fab business, if they think custom silicon would make their centers cheaper/lower power, but there wouldn’t be much reason to go beyond the fabless stage (which doesn’t need CapEx), unless they think their production partners might not be reliable enough to produce stuff on Google’s schedule. But there would probably be rumors about that, and I haven’t seen any.

    Or it may just be that Google’s push into the telecom business is getting bigger. Telecoms have a *lot* of CapEx. AT&T is around $20B, for example, or about $5B/quarter, well above Google’s current level. And AT&T’s network is pretty mature, so someone expanding into the telecom market should be burning CapEx at a much higher rate, relative to footprint.