MTS

Wiki/Primary documents

Elpida bankruptcy — February 27, 2012

Filing date: February 27, 2012. Court: Tokyo District Court, under Japan's Corporate Reorganization Act. Significance: Largest manufacturing bankruptcy in Japanese history at the time. Resolution: Micron sponsorship agreement July 2, 2012 (Micron 8-K). Acquisition closed July 31, 2013, ~$2.5B net of cash and assumed obligations.

Why it matters: Elpida is the second of the two consolidation events that established the current Big 3 structure of the DRAM industry. Together with Qimonda, it taught the surviving producers that even a major DRAM maker with strategic-national support (Japanese government backing, multiple-year subsidy talks) can fail when the cycle turns. After Elpida, the merchant DRAM market consolidated to Samsung / SK Hynix / Micron — a structure that has held for 13 years and that is the precondition for the modern discipline doctrine.

Primary sources:

  • Elpida Memory press release, February 27, 2012, announcing the Corporate Reorganization Act filing
  • Micron 8-K, July 2, 2012 (sponsorship agreement)
  • Micron 8-K, July 31, 2013 (acquisition close)
  • Micron 10-K FY2013 (~$2.5B purchase price disclosure)
  • Contemporaneous Reuters, Nikkei, FT coverage

What happened

  • 1999: NEC and Hitachi merged their DRAM businesses to form Elpida Memory.
  • 2003: Mitsubishi Electric folded its DRAM business into Elpida.
  • 2007–2009: Suffered through the 2008–09 cycle alongside Qimonda; survived where Qimonda did not, partially due to Japanese government support.
  • 2011: Trading losses deepened; multiple restructuring talks; Japan Bank for International Cooperation involvement.
  • February 27, 2012: Filed for Corporate Reorganization Act protection in Tokyo District Court. ¥448 billion in debt.
  • July 2, 2012: Micron announced sponsorship agreement. Micron 8-K.
  • July 31, 2013: Acquisition closed. ~$2.5B net of cash and assumed obligations. Micron 10-K FY2013 disclosed the price.
  • Aftermath: Elpida's Rexchip joint venture (with Powerchip in Taiwan) was acquired by Micron in the same transaction.

The shape of the failure

Elpida was the textbook case of "strategically supported but structurally underwater":

  • Strategic significance: Japan's last merchant DRAM maker. Government had multiple political reasons to keep it alive.
  • Technology position: credible 25nm DRAM development; not a clear technology gap.
  • Financial weakness: deep cumulative losses through 2008–2009 left the balance sheet impaired even after the 2010 recovery.
  • Cycle exposure: the 2011 demand softness after the post-iPhone-era smartphone-DRAM cycle pushed Elpida from "wounded but operating" to "insolvent."

What Elpida demonstrates that Qimonda did not — Elpida had years of government engagement, restructuring talks, and credible turnaround plans. None of it was enough. The structural reality of being the marginal producer in a cyclical industry overwhelmed the political and strategic interest in keeping the company alive.

What the consolidation produced

After Elpida's exit and Micron's acquisition:

  • Merchant DRAM industry: Samsung, SK Hynix, Micron (Big 3, >95% bit share), plus a Taiwanese tail (Nanya, Winbond, PSMC).
  • The three-firm equilibrium: The smallest viable oligopoly structure that can sustain disciplined capex without triggering antitrust scrutiny. Two firms would be a duopoly; four firms would be unstable. Three is the equilibrium that has held since 2013.
  • CXMT (ChangXin Memory Technologies, China): The only meaningful capacity entrant since Elpida exited. As of mid-2026, CXMT has not qualified HBM3e with Nvidia. It is the wildcard the wiki returns to repeatedly — the entity that could break the three-firm equilibrium if it succeeds.

Why this sits in the wiki

Two reasons:

  1. The Big 3 structure is the precondition for the discipline doctrine. A three-firm oligopoly with deep institutional memory of consolidations can sustain unilateral capex discipline without bilateral coordination. A four- or five-firm market could not. Elpida's exit produced the structure that makes today's "sold out through 2026" credible.

  2. The bar for "we will not let this happen again" is institutional, not merely strategic. Elpida had strategic support and still failed. Today's producers have no equivalent strategic backstop — Micron is American but operates without a sovereign guarantee; Samsung and SK Hynix are Korean but operate in a less interventionist regime than 2010s Japan. The lesson the survivors took from Elpida is that no one is coming to save you. The capex discipline today is the rational response to that lesson.

What to read next