Intel Lends Support to Achronix Against Xilinx and Altera in the FPGA Industry. Excess Competition or the Rule of 3?
In October of 2010, Intel Corp. took an unexpected move by granting Achronix Semiconductor Corp. access to its most advanced foundry to produce field programmable gate array (FPGA) chips. Xilinx Inc. and Altera Corp. currently dominate the FPGA industry. How serious of a threat is Intel’s move to the historic industry cohorts? More specifically, will this move harm industry profits as competition heats up? And, how should executives and investors react?
By comparing the FPGA industry to the industry structure of mature industries, I not only lend my cautiously support to Intel’s and Achronix’s decision, but also lend my hopeful support to the executives and investors at Xilinx and Altera.
The FPGA Industry
FPGAs are high performance chips used in high speed communication, image processing, and encryption applications. According to Gartner research, the FPGA industry is anticipated to grow at 44%, outpacing the estimated 31.5% semiconductor industry growth overall.
The anticipated industry growth alone might indicate that the FPGA industry is growing sufficiently to support more competitors. Yet industry growth alone does not imply that competitors will behave rationally. High growth industries are routinely plagued with damaging hypercompetition and low profits as firms seek to claim the dominant position to reap the rewards of network externalities or economies of scale, scope, or learning. Fortunately, mature managers with level heads are able to foresee the profit damaging consequences of price wars. As such, the risks of excessively competitive actions are somewhat foreseeable and can be mitigated by level headed executives.
Industry Configurations and The Rule of 3
The issue that really gives me pause however is the Rule of 3.
Absent regulatory restraints or anticompetitive practices, the Rule of 3 predicts that industries will evolve towards a stable dynamic equilibrium consisting of three dominant competitors, several niche competitors, and some stragglers that compete either in between or on the periphery.
In Rule of 3 industries, the dominant competitors (greater than 10% market share) tend to be generalists that are volume driven. It has been conjectured that they will tend to split the market with 40%, 20%, 10% market shares.
In Rule of 3 industries, the numerous niche competitors (between 1% and 5% market share) tend to be highly successful, margin driven, specialists.
In Rule of 3 industries, stragglers which lie between the smaller specialists and the larger generalists, (between 5% and 10% market share) are known to lie in the ditch of profitability – too small to reap sufficient economies of scale but too large to reap the benefits of focus. And stragglers with less than 1% market share are struggling to stay alive.
Convergence to this industry configuration can take several decades, but it has been observed across several industries and across time. Four influences are thought to moderate the convergence process: (1) industry cost structure and shared infrastructure, (2) government intervention or deregulation, (3) industry consolidation or globalization, (4) technical process or product standards.
Recent support for the Rule of 3 industry configuration has come from Uslay, Altintig, and Winsor. Using the University of Chicago’s Center for Research in Securities Prices (CRSP), Uslay et. al. examined industries with over 15 years of existence to determine their configuration. From 1369 firms in 152 industries for 1997 and 1108 firms in 165 industries for 2002, the pattern proposed by the Rule of 3 appeared to dominate. (See Exhibit: Frequency Distribution of Generalists.) More industries converged to have thee large generalists than any other configuration. That is, industries with four or more generalists were rarer than industries with exactly three generalists. Similarly, industries were more likely to have exactly three generalists than either two, one, or none. The Rule of 3 is the most common industry configuration.
However, to state that the configuration is common is not the same as stating that it is profitable. To make this second, and more important claim, Uslay et. al. continued their research by looking at the return on assets (ROA).
Industry ROA and The Rule of 3
Pundits have conjectured that overall industry performance, i.e. profits, are higher for industries that converge to the configuration defined by the Rule of 3, than they are for any other industry configuration. To demonstrate that this hypothesis is more than myth and lies in the actual fact category, Uslay et. al. compared the ROA earned by firms that fit within in an industry with three generalists to those with four or more, and those with two or fewer. In both cases, the industries with exactly three generalists outperformed all other industries on ROA, on average.
This finding might be explained by considering the challenges and benefits of competition. On the negative side, industries with more than three generalists may suffer from lower profitability due to excessive competition which prevents mutual forbearance in entering into price wars. On the positive side, industries with fewer than three generalists suffer from lower profitability due to complacency, decreased innovation, and decreased quality which leads to an insufficient expansion of the potential market. As such, three and exactly three generalists may lead to an optimal tradeoff at the industry level between the challenges of excessive competition and the benefits of competition induced enhanced productivity.
Firm ROA and The Rule of 3
Enhanced industry performance may bode well for investors seeking make portfolio allocations between industries, yet individual executives cannot diversify the inherent risk of depending upon a single firm for their paycheck. Hence, Uslay et. al. also examined the ROA of firms within industries conforming to the Rule of 3 in comparison to their market share.
Pundits have conjectured that, for generalist firms within industries conforming to the Rule of 3, increasing market share is correlated with increasing profits, but only up to a point. Uslay et. al. demonstrated the validity of this hypothesis by examining the ROA of firms with 10% to 40% market share in comparison to those in the ditch (between 5% and 10% in market share) and those with excessive market share (more than 40% market share). They found that the statistics indicated that these 10%-40% market share generalists outperformed the others.
Similarly, pundits have conjectured that, for focused specialist firms within industries conforming to the Rule of 3, firms with market shares between 1% and 5% tend to earn higher profits than those in the ditch or those struggling to survive. Again, Uslay et. al. demonstrated the validity of this hypothesis by examining the ROA of firms with 1% to 5% market share in comparison to those in the ditch and those with insufficient market share. They found that the statistics indicated that these 1%-5% market share specialists outperformed the other specialists.
Together, these results have led to the following understanding of ROA and market share. Highly profitable small specialists with market shares between 1% and 5% are the norm. And, generalists with 10% to 40% market share are, on average, more profitable than other generalists. (See Exhibit: Return on Assets on Market Share.)
Rule of 3 and the FPGA Industry
Applying the insights earned from examining the Rule of 3 to the FPGA industry, we come to a very tempered conclusion that Intel’s support of Achonix might improve the health of the FPGA industry, and moreover the health of specific competitors within the industry. If it leads to a somewhat stable dynamic equilibrium with three competitors, rather than just the two of Xilinx and Altera, industry level profits can be anticipated to improve. Moreover, if we contrast the ROA of firms with a healthy 10% to 40% market share to those with an excessive (>40%) market share, we may furthermore determine that these three firms will actually reap higher returns on capital deployed. Perhaps instead of lamenting Intel’s move, Xilinx and Altera should instead take a suggestion from Uslay et. al. “For market leaders approaching the upper limit of the economies of scale, it might make sense to … expand the pie and protect share rather than maximize share.” 44% market growth in a year implies a lot of pie growth. Perhaps there is some room for a third, stabilizing, cohort in the FPGA industry.
Note: At the time of writing, the author is not currently a direct consultant to nor investor in any of the firms listed in this article.
Don Clark, “Intel Grants Rare Access to Start-Up: Chip Giant Will Manufacture Achronix’s Semiconductors Using Most Advanced Production Process,” The Wall Street Journal, Eastern Ed. (31 October 2010): B.4
Can Uslay, Z. Ayca Altintig, & Robert D. Winsor, “An Empirical Examination of the ‘Rule of Three’: Strategy Implications for Top Management, Marketer, and Investors,” Journal of Marketing 74, no. 2 (March 2010): 20-39.
Jagdish N. Sheth and Rajendra S. Sisodia, “Only the Big Three Will Survive,” The Wall Street Journal, Eastern Ed. (5 November 1998): A22.
BCG, “The Rule of Three and Four,” Perspectives, no. 187, Boston, Boston Consulting Group (1976).