"The difference between stupidity and genius is that genius has its limits." - Albert Einstein
"I'm no good at being noble, but it doesn't take much to see that the problems of three little people don't amount to a hill of beans in this crazy world. Someday you'll understand that." - "Casablanca", 1942 - (screenplay by J. Epstein, P. Epstein, and H. Koch)
Once and awhile it seems like certain things happen that really upset your mental applecart and make it hard to regain your bearings in their aftermath - as if a flash bang grenade went off nearby and you are momentarily robbed of the ability to think straight. Any news associated with the topic of China, Inc. entering the memory business seems to engender this same effect on the stock market. One perfect example of this occurred recently when Mark Durcan of Micron (NASDAQ:MU) responded to a question at the recent Micron-Inotera merger closure ceremony in Taiwan. Asked about the specter of potential over-supply rising from Chinese intentions to build a massive memory industry Durcan said the following:
"It's possible if [China spends] the money that they say they are capable of spending [$150B], that they can create oversupply," Durcan told reporters in Taipei. "Is it a concern or something to think about? Absolutely."
I don't know about you but I am so glad we got that cleared up. I'm sure that we are all four square in agreement that an asteroid the size of the moon - upon Earth contact - would definitely blow the crap out of everything. Next question?
Now, as much as we all might wish that Durcan had answered that question more gracefully, it's obvious that news about the Chinese intent to enter the semiconductor business is not going to go away. Micron and WDC dropped roughly 3% that day, an entirely predictable result of the comment. The stocks have recovered nicely since then as folks regained their senses and focused on the tight NAND and DRAM markets but it's obvious that this issue is on everyone's mind and there's a lot of fear associated with it. Given even the least hint that China will have the capability of producing memory it's freak-out time - sell everything and run for the hills, folks, the asteroid is coming!
So we know the fear is there. Is this simply mass hysteria or is there something we investors need to be concerned about? If not now, when? Even if China is not a big deal in short term, investors are certainly justified in seeing the China threat as a negative when it comes to the longer view and this in turn could depress industry multiples. As I shall outline below, there is reason to be concerned - China actually could be that moon-sized asteroid - but not anytime soon and not in the way that most folks think.
Like most things in this crazy modern world this issue is messy and complicated - there are opportunities and risks - puts and takes - so let's all calm down, take a deep breath and sort out where we are now and where we're headed with China. In so doing there is one thing that is absolutely paramount that we have to understand. The semiconductor memory industry is just one small component of the problem space and there are numerous questions that relate to the timing and tactics of the overall Chinese approach to world trade and the global response to it that will have an outsized impact on the magnitude of the potential impact. Note that I said "global" response. This is political, folks, and entirely beyond the capability of any company, group of companies, or industry to resolve on their own. That is not to say that there is not a vital role for the memory companies or the semiconductor industry as a whole but it does mean that they can't solve the problem on their own. Here's why.
I have written before about the motives, strategy, and outcome of the Chinese entering the memory business and any of you that have read those pieces know full well where I stand on this issue. For those of you that haven't, here's the short course:
The most vital thing to understand is that China's strategy is defined and directed by the Communist Party of China (CPC). It is, in other words, a State initiative, conceived to accomplish State strategic purposes. It is most of all a political initiative designed to enhance and project Chinese power. It is not in any way a product of conventional business strategic thinking. Because this is the case, conventional measures of business success or failure like return on capital or profitability have little or no significance. As any investor or observer of the semiconductor memory industry knows, the whole question is in economic terms completely irrational. But the Party is insistent and has made this one of their top goals. Why?
Originating as one of the elements of the 2006 CPC policy titled "Indigenous Innovation", the roots of the initiative are in a nakedly mercantilistic economic policy that has been at work over the last thirty years. The same people that brought you massive over supply and the attendant global dislocations in aluminum, steel, LED's, and solar panels look at the world and say to themselves - why wouldn't the same approach work with semiconductors?
While the goal of the effort is to build a globally competitive semiconductor industry in China by 2030, the broader implications of China's effort are ominous. As the recent 2016 ITA Semiconductors Top Markets Report states it:
"[…] discriminatory policies implemented to support its industry to become self-sufficient pose a real long-term threat to not only U.S. firms but the entire global semiconductor ecosystem."
What are these discriminatory policies? Here's Rob Atkinson of the Information Technology and Innovation Foundation discussing just a few of the bigger ones:
"China has for years aggressively threatened foreign competitors with serious penalties, up to and including exclusion from the Chinese market, unless they hand over valuable intellectual property (IP). It has pumped up its tech enterprises with massive subsidies. And it has erected a vast array of contorted rules and regulations inside its borders in a quest to skew the terms of competition and gain absolute advantage in key industry sectors ranging from semiconductors to biotechnology."
Atkinson's list is bad enough but it is just the beginning of the gauntlet US companies must run to compete in the Chinese market. Other obstacles of note include China's phony patent law, flouting of IP licensing standards up to and including outright IP theft, and parochial Chinese "standards" in contravention of established and more robust standards (Exhibit A being the Chinese "LTE"). Whatever the exact methods it is clear where China's initiative leaves the global memory industry. Here's US Commerce Secretary Penny Pritzker giving her take on China's chip initiative in a November 2 speech.
"[The Chinese plan to invest $150B in semiconductor capabilities…] would be equivalent to half of worldwide semiconductor sales last year and result in market distortions similar to those plaguing the steel, aluminum and green technology industries […] The world has seen the effects of this type of targeted, government-led interference before […] The result has been overcapacity in the global marketplace that has artificially reduced prices, cost jobs in both the United States and around the world, and caused significant damage to those industries globally," […]. Pritzker speech - Center for Strategic Studies, November 2, 2016
But why the memory business, specifically? The question has been asked and answered in the industries Pritzker mentioned in her speech. Steel, aluminum, green tech (solar cells) - what do they have in common? Answer: They are commodity-like products that are foundational and/or transformational. In other words, once you control the worldwide market for those products, you can build massive volume and support-industry scale to drive market domination initiatives across the entire tech-industry value chain. In the new "memory-centric" computing model semiconductor memory is the key pillar that supports much of the upstream value chain. Long-term, Chinese domination of the memory industry would be a direct threat to non-Chinese companies dependent on access to this critical technology. Short term, there can be no doubt about the effect of Chinese success with this policy: China's entry into the memory business would result in catastrophic shareholder losses that are a direct consequence of the massive oversupply condition that would result from China's attempt to dominate the industry. Not only that, but the attendant destruction of shareholders assets would spread far beyond the memory business.
So yeah, this Chinese policy is indeed a moon-sized asteroid. There are existential issues at stake here. But can they really do it? How would it happen? When? These are serious questions that have yet to be answered. It is clear that the memory industry Oligarchs don't believe they can - not anytime within this decade at any rate and certainly not with the current generation of memory - meaning DRAM and NAND. I believe they are right. But don't take my word for it. Here's Siva Sivaram, EVP of Technology of Western Digital (NYSE:WDC), assessing Chinese NAND capabilities in the recent WDC Investor's Day Meeting during the Q&A:
[Their] "stated roadmap is multiple generations behind where we are in the industry in terms of when the technology is planned to be productized… […] They are 2D [..but …] there's a stated goal of getting to 3D NAND [on their roadmaps]. As you can see we are on a cadence of 18 months for new generations of BiCS.[…] As we move further, […] they fall further behind, leaving a large gap where the technologies are productized. […] It's a tough mountain to climb." - [authors transcription and emphasis] - WDC Investor Meeting - SA Transcripts
This statement is completely congruent with past remarks by Micron (Durcan, DeBoer, and Maddock) that place the entry of China late in this decade, if then, and even then with an inferior product that is two to four generations behind the current industry standard. Clearly, the key to the game is IP - the Oligarchs have it - China does not.
If there is to be a legitimate Chinese threat to the incumbents China must get it. But how? If you have been following this issue over than last year or so you can read numerous articles that opine that some sort of "joint venture" is possible with Micron being the likely candidate. This is a fantasy. The oligarchs aren't going to give IP to China or deal it away, Micron included. The memory industry companies, especially Micron, know all too well (from years of bitter experience) what the implications are of irresponsible State capital at work in their business. The other key thing to remember is that even if one of the industry companies was stupid or desperate enough (Toshiba?) to license IP, there is a robust legal framework that would prevent such a transfer. (I'm not going to discuss the details but I have explored this issue thoroughly in this article, and I encourage you to read it.) Finally, above and beyond either the factors of industry self-interest and legal prohibitions regarding IP transfer, there is the most important factor of all, which is political.
Donald J. Trump is two weeks away from being the 45th President of the United States. There is not a snowball's chance in hell that a Trump administration is going to allow strategic IP sale or transfer - via outright M&A or any other mechanism such as a "joint venture" - during the term of his Presidency. (If you want to get a flavor of where things are going read the recent NY Times article on Robert Lighthizer, Trump's nominee to be the next US Trade Representative.)
So we have these robust barriers for "legal" transfer but the paranoid among you may well say but what about illegal methods of obtaining the IP? Could China steal it? The answer of course is yes. Much like SK Hynix (OTC:HXSCF) did with Toshiba (OTCPK:TOSBF), China can (and will) hire away managerial, process and engineering talent from the incumbents who may or may not bring proprietary IP with them. In many ways, however, this is only half the battle, and maybe less than that. Using a baseball analogy, you may steal a sign and know that Nolan Ryan is throwing you a fastball. Know that and hitting it are two totally different things. Modern semiconductor manufacturing is an ensemble effort requiring sustained collaborative effort throughout a dauntingly long and subtlety complex recipe chain of numerous steps, all of which must be done exactly right for the final product to be usable. China will, over time - lots of time - develop this process knowledge but we are talking years, not months, for this to happen.
The bottom line for concerned investors is this: China will not have a competitive DRAM or NAND product this decade. Will they in the 2020's? DRAM will be the easiest target because the glacial pace of process advance. Current thinking is that DRAM scaling is coming to an end with maybe only two more nodes that would be usefully produced. So China can eventually catch up in DRAM, but why would they? The new memories on the horizon show the promise of taking big chunks of DRAM market share as the 20's advance. Why would even an economically irrational China want the piece of a potentially rapidly declining market?
NAND on the other hand is a more interesting question. Obviously WDC doesn't think they can catch up, but even if they can't - do they have to? Let's project a scenario that has the other NAND producers 2 - 3 nodes ahead of China in 2022. That should equate a to 30 to 50% cost advantage and with that die density advantage comes reduced power usage and better performance and packaging flexibility. So why would a user buy the Chinese product? Niche markets could always be found for some of their product but China needs massive scale. China Inc. could discount its chips enough to make them attractive but it would take horrendous losses doing so. That being said, when has that ever stopped them? So it is certainly possible we could see a large supply of Chinese NAND on the market. It's also conceivable that China's $150B fund could bring as many wafers on to the market as the rest of the industry combined - by that time that would be roughly 1 million wpm, which because of their presumed density disadvantage, would equate to 500 - 700k wpm for the incumbents in bits. Is this necessarily a negative for the other NAND companies? Maybe not. As I have argued in a past article, by that time the world may well need the increased supply. My analysis shows a 2 ZB deficit (that's 2 Billion TB) in 2020 just to meet the storage needs for analyzing the "target rich" data being produced. Given the process standards likely at the time (a 1536 Gb TLC die of 128 layers), over 40 new fabs costing $350B in CapEx would be needed to meet that magnitude of user demand. Netting it out, China's $150B of CapEx may buy them a lot of fabs but even that level of investment may not be enough to meet demand.
Personally I think that NAND has reached the tipping point and that this is a likely outcome because of exploding demand growth rates. In this scenario China, Inc. is basically subsidizing the storage users of the world by allowing them to buy NAND that they otherwise would not be able to. The other aspect to this scenario is the impact of the new memories. Right now it is clear that DRAM will be affected but what about NAND? It is possible that the "ideal" all-purpose memory comes to market by say, 2025, but even so how much production scale would exist at that time? We know the computing model is changing to utilize the new NVM memories but it seems unlikely to diminish the need for gigantic amounts of storage, and that's the province of NAND.
The larger question in all this rumination is what can China do in the next generation non-volatile memory market? Various flavors of resistive RAM (ReRAM) will be coming to market as the 20's advance, and this would seem to be the ideal point of entry for a truly innovative Chinese producer. If China truly wanted to play by the rules in international technology trade this would be the strategic move that made the most sense in terms of providing a true economic return. This may even be the one area where an M&A strategy for acquiring early-stage memory IP might be allowed.
Boiling it all down from a parochial perspective, Micron, Intel (NASDAQ:INTC), and WDC (the US memory industry) have the will and the tools (legal and political) to thwart the Chinese. If you believe as I do that NAND demand will absorb any extra capacity (even if the product is archaic and heavily subsidized) then I have painted a rosy scenario through at least the mid-2020's. If you doubt NAND demand then maybe you could posit a negative impact on the industry that develops after the turn of the decade. Take your pick. Either way there is absolutely nothing a reasonably prudent investor should worry about for some time to come - well beyond most investing horizons.
So there's nothing to worry about, right? Well, no.
Unfortunately, the real risks to the Micron's of the world have nothing to do with China's semiconductor initiative per se. To see those risks we have to lift our eyes from the memory industry's trees to global trade's forest. It is in that larger view that we will see the true shape and scope of the emerging great power - China - and its threat to global peace and prosperity. WW III? No, but resolution will not come fast or easy and even the largest companies and industries will merely be pawns in the outcome. The real question is this. How can the US lead a global response to China that challenges the Chinese strategy of naked, predatory mercantilism in such a way that needed changes occur?
My next article will explore that larger question. Happy New Year.
Disclosure: I am/we are long MU, WDC.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Editor's Note: This article discusses one or more securities that do not trade on a major U.S. exchange. Please be aware of the risks associated with these stocks.