When it comes to semiconductors, be wary of bets on PC memory — so-called Dynamic Random Access Memory (DRAM) — and instead bet on Flash, which is used on the iPhone, iPad and iPod. That does not mean that all PC chip bets are bad, though. That’s just the takeaway from a look at semiconductor companies Micron Technology (NASDAQ:MU) and Intel (NASDAQ:INTC).

Late Thursday, Micron reported an unexpected loss. In its report, Micron posted a $135 million loss of 14 cents per share for its fiscal fourth quarter. FactSet-polled analysts expected a penny per share of profit. Micron’s 14% revenue drop to $2.14 billion from was slightly better than the expected $2.11 billion.

The report is a mixture of bad news and good news. The bad news is that the revenues dropped because of Micron’s dependence on DRAM — where prices are dropping along with demand for PCs. The good news is that Micron has shifted its business mix in favor of the more profitable flash memory chips — 41% of total sales. Its DRAM business is down to 36% of the total.

That’s not to say that all bets on PC chips are off. Despite slowing PC sales, corporate PC demand has been stronger than many expected. Its new PC Central Processing Unit chip architecture helped it gain more market share. Dubbed Sandy Bridge, Intel’s CPU architecture combines “high-level graphics and CPUs integrated onto the same piece of silicon,” according to eWeek.

That architecture has given Intel a phenomenal market share of 81.8% in the global processor market, according to market research firm IHS iSuppli. And while the PC DRAM business features dropping prices, the CPU market remains a profit machine for the market king, Intel.

Despite slowing PC sales, corporate demand remains strong while consumer demand wanes. According to Matthew Wilkins, principal analyst for computer platforms research at IHS iSuppli, “Intel in the second quarter benefited from the combination of a recovery in PC demand and strong shipment growth for its new Sandy Bridge line of microprocessors. Strong corporate PC sales were particularly beneficial to Intel, as the enterprise computing segment has been outperforming the consumer market.”

Does this mean you should avoid Micron and buy Intel? I would avoid both but for different reasons:

• Micron: Hard-to-predict earnings. Micron’s latest report is a reversal after a very strong 12 months’ worth of results. During the previous 12 months, Micron’s revenues soared 77% to $9.1 billion while net income of $644 million was up 198% — yielding a modest 7.1% profit margin. The stock is very cheap based on a price/earnings-to-growth ratio of 0.15 (where 1.0 is considered fairly valued). Micron’s P/E is 8.9 on earnings forecast to grow 58% to $0.88 in fiscal 2013. The problem for this stock is that the latest quarterly report is so much worse than expected that it throws into question the sustainability of its longer-term performance and the reliability of its forecasts.
• Intel: Profitable, growing, expensive stock. Intel revenues are up 24% to $48 billion in the past year, during which it posted a $12.3 billion profit, up 162% — a whopping 25.6% net margin. The stock is expensive based on a PEG ratio of 2.5. Intel’s P/E is 10 on earnings forecast to grow 4% to $2.45 in 2012. In its second quarter, Intel profit gained a mere 2% to $2.95 billion, so despite its longer-term profit growth trends, its most recent results suggest it might not be able to accelerate earnings growth.

While Intel clearly is a more solid company, neither looks like a great place to bet your investment chips.

Peter Cohan has no financial interest in the securities mentioned.