Since its inception, Semiconductor Manufacturing International Corp., better known as SMIC, has pursued an aggressive plan to join the elite chip foundry companies. During the Chinese firm's chase of the big two--TSMC and UMC--it passed Chartered to ascend to the number-three slot a couple of years ago, although it has been nip and tuck with the Singaporean concern since then, alternating between third and fourth place. But the one cornerstone of SMIC's gameplan where it has failed miserably is one of the most basic to any business model--profitability.
Every quarter of every year, the story has been depressingly similar. SMIC posts net loss, SMIC execs initiate spin control and say they expect to break even or be profitable soon, so don't worry Mr. and Ms. Shareholder. Been that way every year this decade, except for 2004, when the company turned a small profit, which turned out to be as fleeting as a blue-sky day in Beijing.
At the end of its last fiscal year, the song remained the same and SMIC again ended up in the red--and the results of the first two quarters in the books so far for FY2008 have revealed even greater financial hemorrhaging, to the tune of a net loss of more than $272 million, although some of the bleeding has to do with the company's efforts to get out of commodity DRAM and focus more heavily in the logic area.
So when I read the brief DigiTimes report making the rounds this week about the company's "process node migration" plans, I had to shake my head and chortle.
The story cites a source at SMIC who repeats the company's claim that it will make the transition to 65 nm "by the end of this year" (the last of the Big Four foundries to do so), but adds that "32-nm development is also expected to follow next year," supposedly in 3Q09. The source goes on to say that the foundry has "also started development for cutting-edge process nodes, including 45 and 40 nm."
Can the industry really count on a company that has been a loss leader throughout its entire existence and must include its dubious IP infringement track record with every financial report to be in business when it's time to not just develop 32 nm process but ramp it to volume production? And in the longer term, does SMIC have the staying power to make it successfully to the ever-more-daunting 22-nm process regimes?
I will stipulate that SMIC has many apparently satisfied customers, a decent family of volume fabs, its finger on the pulse of the nascent Chinese fabless industry, technology partnerships with the likes of IBM and Spansion, that it does indeed have some positive things going for it. But the question remains: Will SMIC's well run dry before it has a chance to complete its quixotic quest for technological parity and, gasp, profitability?
Perhaps SMIC's future lies in the diversification sector du jour--solar. In the company's most recent conference call in late July, president/CEO/founder/face-of-SMIC Richard Chang gleefully noted how pleased the company was "with our profitable solar business and approved expansion plan. We will expand our solar capacity five to six times our current production output in the first half of 2009."
At least there's one part of SMIC that's showing a profit.