Amazon.com and the fabless semiconductor industry

November 9, 2006 | Baris Karadogan | Comments | |

[VentureBeat criticized Amazon for lack of vision. Baris Karadogan, of ComVentures, begs to differ and commends the company.]

A key theme at the Web 2.0 conference was infrastructure, centered around Amazon’s big announcement to offer start-ups all their infrastructure needs.

Amazon wants to give start-ups bandwidth, storage, server software and even fulfillment of physical goods — all at a reasonable price. This is all the boring stuff you need to go from your great idea to a product. Google and Salesforce.com call this infrastructure “the cloud,” Microsoft calls it “a million pounds of batteries,” and Amazon.com calls it “muck.”

The idea goes like this. Getting your web application to scale globally and serve hundreds of millions of users is expensive. You need to build big data centers and buy software and fat pipes. Google spends more than one billion dollars per year on this (and their ability to do this is a key competitive advantage over others ; see here).

Startups can’t afford that kind of capex. We VC’s don’t see Series A opportunities that say “we are thinking of raising $500K to launch our web site and about $700M to build a global data center.” These startups go to hosting providers and buy computing and storage by the drink. This is what Amazon wants to provide startups, everything; the software, the hardware, the bandwidth and even physical fulfillment.

Sound familiar? This is very much like what happened in the semiconductor industry as it went to the fabless model. The transistor was invented in 1947 and the first IC (integrated circuit) was built by TI in 1958. By 1978, the semiconductor industry reached $10B in sales. As chips got smaller, it became harder and more expensive to build the foundries to build the chips. While Intel and TI continued to own their own fabs (where you build chips), smaller companies could no longer afford the billions of dollars required to build them. As a result, in 1987, TSMC, the world’s first independent semiconductor foundry was established. It is still the largest. Now a startup no longer needed invest the capex to manufacture their chips, they could design it, and the foundry would build it for them. Since the foundries amortized the capex over a number of companies, the model worked. The foundries in the semiconductor world are the data centers of the Internet world.

In the Internet world, a few big players have emerged that have the wherewithal to build their own big data centers. Microsoft, Yahoo, Google and Amazon are what Intel and TI are in the semiconductor business. Now we are seeing that some of them, like Amazon are selling their capacity to startups, and some pure play data centers businesses are emerging. I suspect we’ll see more of them, establishing themselves and rolling up.

Now back to our story. After the large foundries established themselves, fabless semiconductor startups started showing up and growing their business to respectable sizes. They design the chips, TSMC built them. The two big fabless semi companies are Broadcom and Nvidia, which are both around $2.5B in annual revenues. So a healthy ecosystem formed with independent foundries building chips for fabless semiconductor companies which focus on applications and software.

So what can we learn from this analogy? Despite the fact that the fabless semiconductor revenues reached $41B in 2005, it is still merely 18% of the overall semiconductor industry. Intel, who owns its own fabs, has about $40B in sales, nearly the size of the entire fabless semiconductor industry. Those who were early and kept their infrastructure, got really really big. They delivered products faster and cheaper than their fabless competitors. Same will be true for Google and Amazon (and Yahoo if they can get their act together). Amazon could offer to take care of the “muck” for a small startup, but as soon as that startup gets big, Amazon will able to do the same faster and better because they control the critical piece, just like Intel.

So I applaud Amazon’s strategy. It’ll get an early look at the apps, and maintain control of a critical piece of know how. As data center complexity grows, it will be harder and harder for a newcomer to build their own datacenter, and will be more and more reliant on Amazon. What would be a nightmare for them is for an upstart to out-innovate the “muck.” So Amazon’s strategy to dominate early on prevents startups from going there and independent data centers from surviving, since few companies have Amazon’s scale. This has got to be Amazon’s biggest move so far to ensure its sustainability.

Next Story: Forsight Labs launches glaucoma treatment co., Transcend, with $7M
Previous Story: The Charles River Ventures Quickstart Calculator

Bookmark and Share

Photo of Baris Karadogan

About the Author, Baris Karadogan

Baris Karadogan is a partner at venture capital firm ComVentures.

Baris writes a blog titled From Istanbul to Sand Hill Road (http://baris.typepad.com)

  • Brad Twohig
    Very insightful post.
  • JohnW
    Good write up. However, it is important to remember that the ultimate success is not measured by (the growth in) top line revenues, but profits and the investment needed to get those profits (i.e., return). Typically the "cloud" or "infrastructure" business requires - as you said - huge up-front capital investments and on top of that, the profit margins are lower than in the IP/service/media business. Examples from other industries could be MVNOs vs. wireless network operators, AOL.com vs. AOL access business, Skype vs. AT&T carrier etc.
  • Agreed, though this is what debt is for. And I believe Amazon at one point raised $1B in debt.
  • "This has got to be Amazon's biggest move so far to ensure its sustainability."

    I am not sure your article addresses the value of this new effort to Amazon's business, so the last sentence of your article seems unjustified to me. Can you elaborate on why you believe Amazon's web services effort will have such a large impact on its business, big enough to ensure Amazon's sustainability?
  • How easy do you think it is for Intel to be unseated by a processor vendor who does not own its own fab? Unless somebody invests $3-5B to build their own facilities, and be on the cutting edge of performance, it will be hard to compete with them. Nobody has succeeded in doing it so far.

    So if Amazon and the big Internet players effectively become the only ones with big high performance infrastructures, can others threaten them in the long term?
  • Great, I'm glad to see the finance people beginning to understand the displacement implied by AWS.

    I'm curious about the point of competitive intelligence: "So I applaud Amazon’s strategy. It’ll get an early look at the apps, and maintain control of a critical piece of know how."

    On one hand, as AWS grows in popularity, they will become compelled to provide compliance for SarbBox, SAS-70, Basel II, etc., so it is not in their interest to have too close of a "look" into what their customers are doing with AWS - other than from a business perspective. The risk of data privacy issues loom too large for AMZN otherwise.

    On the other hand, AMZN core business is built on a principle of leveraging the "long tail" - making a nickle off each transaction from a myriad of small vendors... while displacing the big fish - notably Wal-Mart, in the early history of AMZN. In that context, AWS promises to squeeze better profit by displacing core business lines of IBM, Oracle, HP, Sun, Dell, BMC, etc., than they would by squeezing the startups which take hold.

    In fact, already this week we're seeing the high end ISPs hurry to launch their own grid/utility packages - too little, too late IMHO. Those ISPs will fall first, but afterwards we'll see the bite taken out of Oracle and IBM.
  • I should have made that clear. Early look, means, following their growth and traffic metrics. They'll know this since they manage bandwidth, storage etc. They'll know which services are more profitable than others.
  • R Clark
    I find it hard to believe that Amazon, who after 10 years is only marginally profitable, will be able to provide a valuable service to the potential customers. If an online market like Amazon makes only a 3% profit margin why would anyone consider them a credible model to follow?
  • Great analogy. BTW clark - read jeff's comment on how he believes in high volume low margin businesses since they dont have a underbelly
  • Rod Recker
    Seems to me the analogy isn't quite right. Is Intel, like Amazon, leasing its fabs to fabless startups? Would a fabless startup want to allow the gorilla in the room to control their operations? Won't a startup jump to a more independent "cloud" vendor as soon as they start to compete with Amazon?

    Seems like the real opportunity is for an independent company to provide the services Amazon is providing. Then you get rid of this basic conflict of interest.
  • I did mention in the article that we'll see data center/hosting providers roll up. But will they every reach the scale of Amazon, or Google? Can they offer the service as cheap as Amazon can?

    If they can't, will Amazon beat them out?
  • R Clark
    I would liken Amazon's new new thing to the old line "those who can't do teach" and add "those who can't teach rely on marketing", which is what Bezos is good at. What other company do you know that would have any crediblity after years of failed promises.

    Notice in this article that a Consumer Reports analysis place Amazon behind Cosco.com, Crutchfield.com, and buydig.com in satisfaction.

    http://www.informationweek.com/news/showArticle...
  • tomo
    I like the anology of the datacenter market to semiconductors.

    This strategy makes sense if they're able to keep scale which will be a challenge should this be successful. Amzn has deployed a bunch of assets, most have a fixed cost and aren't being used to maximum efficiency. So they take the excess capacity, for lack of a better term, and sell it off. The question becomes how ineffcient is their current use or how much capacity do they have to sell. How many youtubes, facebooks, myspaces or whoever can they support and at what point does selling off your excess capacity stop or does it? if it doesn't and they must continue to scale to support their customers requirements and build datacenters and everything else that this initiative could easily become and this is extremely expensive.
  • Pranav
    Baris,
    First of all, great post. I do need some clarification towards the last paragraph though.

    [i]So I applaud Amazon’s strategy. It’ll get an early look at the apps, and maintain control of a critical piece of know how. This has got to be Amazon’s biggest move so far to ensure its sustainability.
    [/i]

    do you mean that Amazon is trying to thwart competition by these efforts -- by getting an insight into their apps and a pseudo-lock-in ?(once my app gets integrated into Amazon and reaches critical mass, it will be difficult to rearchitect the app and decouple amazon from it)...

    hmm...If my understanding above is correct, then I dont see the point you make...Why would Amazon invest so many resources if its goal was to thwart competition - wouldnt it be simpler to acquire the startup ?
  • What I meant is that Amazon will know what's working and what's not because they'll by definition know the storage, bandwidth and infrastructure requirements. If say one app is getting a lot of traffic and not using much in the way of resources, it may be a clue to them to build it themselves. That's the early look.

    Amazon could decouple them, instead of the other way. Think of Microsoft. They allow people to write apps to their OS, but every release they take some of the key apps and make it part of the OS, thereby capturing the value themselves.

    Amazon could give webapps the cheapest and simplest way to go live, but the tradeoff is that they can replace you easier than you can replace them.
  • Pranav
    Baris,
    Microsoft does it -- but alongwith it comes all the detest and the notoriety. If Amazon chooses to do it, and the details become public (courtesy Digg / blogging), its gonna hurt their reputation in the long run. That one scar is going to act as a barrier from getting any future customers for its AWS.
    Besides, building it themselves would make sense if it fits into their biz model.

    This doesnt mean that I totally disagree with your point -- only that I'm not 100% convinced :-)
  • That could be the case but they frankly, most companies who are very successful do these kinds of things. Some do it more carefully than others though :-)
  • Quite interesting comments. One thing that seems to be missed in the discussion about Amazon AWS is the "cleantech" aspect. In a sense, they are leveraging intellectual capital (know-how for providing infrastructure) to refine natural resources (cheap, plentiful, sustainable energy) into a global export (data services). I believe that trend will continue, whether AWS succeeds or not.

    A key limiting factor for AMZN or any other competitor in this space is utility costs. The biggest cost as server farms scale is electricity. Secondary limiting factors are network latency (how fast can customers in major population centers access results?) and access to technical talent (quality service, responsive to the changing needs of high-tech entrepreneurs, English speaking?).

    The apotheosis of services like AWS may tend to follow a geographic distribution like aluminum production - based on requirements for cheap energy. Perhaps even more closely, it may tend to follow aircraft production (energy + tech talent). As an interesting coincidence note that AMZN and Boeing both emerged from the same region.

    Proximity to the northern polar circle would balance cheap energy with network access for population centers. Pacific Northwest seems ideal, but other likely locations include Quebec (PQ Hydro + NorAm population), Iceland (geothermal + both NE NorAm and NW Euro population), Norway (tidal energy + Europe), and Russian Far East (Siberian hydro + Far East Asian population). Any of those five regions could certainly supply the tech talent as well. Following the smart money, I wouldn't be surprised to see either Reykjavík or Oslo take the initiative.
  • tomo
    Paco,

    WRT to PacNW, Amazon's actual datacenters are in the northern VA area. They do have a small datacenter in Seattle but nothing when compared to VA.

    I am still questioning how much extra capacity do they really have and how on earth will they scale it once they reach 70% or 80% utilization. It's not like they can just add datacenters and other infrastructure with the snap of a finger. The notion of selling off excess capacity is a great one and one that has proven succesful for telcos and carriers but multiplexing wavelengths is a much more scaleable proposition than creating a new paradigm in the outsourcing of infrastructure/datacenter/computing services. The marginal utility to Amazon of adding customers will at some point hit diminishing return because there can't be infinite scale to datacenters based on todays environment....there is a sweet spot to actually benefitting from economies of scale when building datacenters and it is finite.

    One more thing to consider, the demands of supporting hundreds or thousands of customers on a service like this require a totally different support and service organization than what running an ecommerce site does and at some point, the two businesses may be equal or the AWS may dwarf the retail which is why I wouldn't be surprised to see Amazon create a new service company that is essentially "AWS" and whose first customer is a retail ecommerce site called Amazon.
  • Jon
    yes i fulfully agree with Paco Nathan on what he has brought to the table in this discussion.
  • Paul
    Interesting post. There are crucial differences however. Pure play data centres lower the barrier of entry for smaller companies that can't afford the upfront capex. Makes sense. However, data centres should benefit from Moore's Law and become cheaper over time. Therefore, scale is important but you are riding technology wave. Semiconductors get more expensive to design and manufacture as you are driving the technology wave. Intel's sales might be growing but look at their capex. Asset turns and their ability to turn that investment into profit is declining. And they are the biggest in the industry!.