Are Memory Chips Commodities? (Part 2)

Photo of an old financial document bearing the title Commodity ExchangeThis post is the second of a two-part series on The Memory Guy explaining commodities, and especially how they relate to memory chips.  Part 1 defines a commodity and explains its attributes.  Part 2 explains the commodity cycle, showing that lessons cam be learned from other commodity markets.

Nobody really likes the commodity cycle.  It causes the market to swing in a way that appears unpredictable.  Yet, the same cycle happens in all commodities, whether mundane or high tech.  Those who understand it and who learn from other commodity markets are better prepared to anticipate these swings and protect their companies from their negative impact while profiting from upturns.

Let’s explore this phenomenon.

The Commodity Cycle

One key attribute of commodities is that they tend to undergo revenue cycles.  The semiconductor industry’s notorious “Boom-Bust” cycle is our industry’s version of this.  It ties to Prof. Kirkwood’s statement in the first part of this series where he said that commodities have higher fixed costs than variable costs.  These fixed costs, the cost of adding capital equipment or of building a fab, typically account for over half the cost of producing a memory wafer.

Also, capital spending must take place well before the new tool or fab reaches its highest production volume.  As a rule of thumb it takes about two years from the construction of a new fab before it reaches full volume.

Now those capital expenditures tend to ramp up when chip makers are profitable, and in a commodity market all vendors are profitable at the same time.  During unprofitable times these same companies slow or stop their investments.  Match this with the 2-year delay, and you get the recipe for the semiconductor cycle:

      • During a shortage prices increase (especially for commodities) and profits grow.  Chip makers increase their capital spending.
      • Two years later that spending creates an overcapacity, and commodity prices collapse, eliminating profits.  Producers respond by slashing capital spending.
      • Two years after the capital spending was slashed a shortage develops, since production was not ramped enough to keep pace with growing demand.  This leads, once again, to a shortage.

Chip makers set their production levels to minimize the cost per unit of output, and remember that fixed costs are high.  This means that they can’t reduce production volume when there’s an oversupply, unless it’s to close down an inefficient facility.  I’ve written two analyses on this: a post on The Memory Guy: A Retrospect of Toshiba’s NAND Production Cut (where Toshiba said that the company would cut production by 30% but then did not), and an Objective Analysis Brief: What a 5% DRAM Wafer Cut Really Means.  An oversupply can’t be avoided in such a market.

Another Memory Guy post called Why DRAMs are Like Steel explains that prices fall until they match the cost of the least-efficient producer, and then usually stop there.  If they keep falling past that point, then that producer drops out of the business, and the price fall stops at the cost of the second-least-efficient one.  In the paragraph above, the 5% wafer cut involved the closing of an inefficient fab.

Chips are Not Unique

Commodity cycles are not at all unique to memory chips.  That’s why that last link above likened DRAMs to steel.  DRAMs share traits with other commodities too.  Let’s look at onions!

Once, when I was visiting Korea, I learned of an onion oversupply.  Prices were so high one year (thanks to a drought in China) that many of Korea’s farmers devoted more of their fields to onions the following year, causing shortages of other crops and an oversupply of onions.  The oversupply caused onion prices to plummet, a situation that could not be resolved for another year.  I found this to be amazingly similar to DRAM’s cycles.

DRAM’s resemblance to other commodities also motivated me to write an article for Forbes some years ago: Why Semiconductors are like Airlines & Shipping.  Both airlines and shipping lines sell commodities, since few consumers have any brand loyalty — they just purchase from the cheapest carrier or shipper.  Both airlines and shipping have high fixed costs and a delay between capital spending on a new vessel and the full production that results from that spending.

What has been painfully learned over centuries can be used to understand commodity chips, making the cycles less surprising, and perhaps less painful.

Make this Work for You

When you put both Part 1 and Part 2 of this series together it becomes relatively easy to understand the more volatile parts of the chip market, and to accurately predict what will happen next.  This is part of the reason that the Objective Analysis forecast has been so consistently accurate for more than a decade.

We can use our understanding and experience to help your company develop a sound semiconductor strategy.  Please contact Objective Analysis to explore ways that we can work together to benefit your company’s strategic plan.