With Elpida's petition to re-organize February 27, we are reminded of the harsh reality that the fantastic rumor of Japanese government bailout doesn't come free, nor quick enough to cure a decade of decline. But it's probably safe to assume some sort of consolidation will happen. So the question we ask is, how does the rumored mega-merger affect the electronics industry at large?
First, let's look at why Elpida was left to deal with the $5.5 billion debt on its own. I am sure part of it is the "moral hazard" principle that most governments face. Why spend taxpayer money to bail out companies with years of poor financial performance? But the more important issue is that Elpida has simply run out of maneuvering space, with fierce competition, a strong yen, and poorly positioned commodity products. Simply finding another strategic partner in Taiwan to lose more money is not a viable strategy.
Ironically, Elpida's bankruptcy may actually be the savior to other struggling SoC (system on chip) companies involved in the rumored mega-merger. Fujitsu, Panasonic, and Renesas may return to the negotiating table with more realistic expectations. Innovation Network Corp. of Japan (INCJ), on the other hand, should understand SoC companies are not getting any healthier. Unlike Elpida's more monolithic DRAM business, the SoC business is complex and is probably ill suited for operating by the creditor committee or the court. With many 28-nm SoC designs in these companies already months behind schedule due to lack of funding, INCJ may negotiate for a better term, but will not get a more viable deal as time marches on. Within another 12 months, these companies will be a full generation behind the rest of the world in set-top-box chips, mobile phone chips, and networking chips. So the time to act is now. The likelihood of a mega-giant emerging is indeed quite high.
But if the past is any indication, the combination of the three will only exacerbate their problems, and most likely transfer the financial burden from the semiconductor operations to the Japanese electronics industry. On the first issue, it's easy to see that despite valiant effort, the Hitachi-Mitsubishi-NEC combination is still early in the restructuring process. Adding two more companies with yet more diverse culture will set back the time table two to three years, during which product, manufacturing, and organizational structure will continue be series of polite compromises to accommodate the traditions of five proud giants.
The SoC business of today is a market-facing business, since SoC design is largely automated and manufacturing out-sourced. In this view, the mega-merger's best-case scenario is a focused fabless SoC company dominating the Japanese domestic market, but it doesn't address the glaring needs of an overseas market expansion strategy. This could result in a profitable SoC company in the short-to-medium term. To be sure, the fabs mentioned in the rumored mega-merger are in the last stages of depreciation, so the financial benefit may not be as significant.
What Japanese electronics companies should worry about though is the prospect that the new mega-giant will essentially be given a monopoly of the domestic market. A monopoly saddled with debt as a result of the merger would have little incentive to be price competitive, as it's always easier to "price-inch" up to customers with little choice than to make unpopular personnel cuts and plant shutdowns no one can agree on. What's more, without the oversea business, it's questionable whether domestic consumption alone could justify 28-nm and soon 20-nm investments in SoC development.
So is the Japanese electronics industry destined to suffer the result of the mega-merger? I believe the door isn't shut yet for the leaders to make demand generation the strategic focus, as opposed to the necessary but insufficient conclusion of integrated device manufacturer vs. fabless model. The demand-generation-focused fabless model require smaller and more nimble companies who can maximize Japanese technological advances for their targeted markets. This model will also require strategic partnership with foreign entities, and this hasn't been practiced by the current generation of Japanese semiconductor executives, who've come into power well after the economic bubble burst in the early 1990s.
Therefore, INCJ and other government entities may have to step in and support these smaller fabless companies in their efforts to expand overseas. Japan remains a financially strong nation, due mainly to corporate and individual savings and recent overseas investments into natural resources. But the time is now to make a wise financial decision on how to restructure the SoC business. The easy path of simple roll-up will likely sow the seeds of future disaster in Japanese electronics industry at large. The more difficult path of market-demand-focused restructuring will seem more foreign to the leadership, but it's long over-due, and may be the most critical issue to address. Charlie Cheng is CEO of Kilopass Technology Inc., a provider of embedded non-volatile memory silicon IP for SOCs.
@ NE01. Samsung may not be a good model to learn from neither. Samsung and Hynix alike are using money from the nation to feed the various development over the past 20 years. Samsung's philosophy is pretty much: dominate the market with tons of investment borrowed from the market, hopefully it can breath longer than competition, once the competition are gone, it can demand better payback. In contrast, Japanese semiconductor IC companies are still governed by the old guards who grew up in the manufacturing-driven mentality, not the market-driven mentality. Plus, most Japanese engineers and management could not speak English well, hence they have a lot of problem to understand what the market wants. In contast, I recall that Samsung announced many years ago that all documentation must be published in English first. A good example is: 3 years ago, in one of the visit to a customer in Taiwan, I went to the meeting with 9 colleagues, two from Japan HQ, two from US HQ, two from local office, 3 from the sales agent in Taiwan. Of the 10 people, I was speaking 80% of time, the sales/marketing person from Japan HQ only questioned on schedule and desired pricing with very broken English. I am certain that he didn't understand most of what transpired between customer and the engineer from US HQ and me. The 5 people from Taiwan office and sales agent barely made any comment. Similar situation happened again and again, perhaps in lesser scale in the subsequent customer visits. Just imagine the lost in the operation efficiency and the travel costs for these Japanese managers! How can Japanese semiconductor companies survive with such kind of operation effiicency?
Join our online Radio Show on Friday 11th July starting at 2:00pm Eastern, when EETimes editor of all things fun and interesting, Max Maxfield, and embedded systems expert, Jack Ganssle, will debate as to just what is, and is not, and embedded system.