There is a real problem in the amount of energy that we use in lighting. Incandescent lighting consumes 22% of all the electricity generated in the U.S., and about 33% of all the electricity generated in the world. But the bulbs produce a lot of heat along with their light. This is often referred to by the anti-incandescent crowd as "waste heat." Funny, but in the middle of winter the two 150-watt spotlights over my desk don't feel like waste heat.

So the government came up with a great idea: Trade your 100-watt incandescent reading light for a compact fluorescent bulb (CFL) that costs 3X as much, takes a while to warm up, provides poor quality light, causes headaches in some users (and stands accused of causing cancer in others), and contains five milligrams of mercury to make environmentally benign disposal difficult enough that we can be sure that lots of mercury will wind up loose in the air, ground and water.

The Great State of California has banned incandescents and has set a 2012 deadline for these lights to be completely phased out. So has Canada. Australia is even more aggressive -- banned by 2010. Europe, not to be outdone -- 2009. Congress is talking about a complete U.S. ban by 2017, although my bet is that it will take them nearly that long to pass any important legislation.

If all those incandescents were converted to compact flourescents, we could cut about a third off of that energy bill. But if you're like me and think that the CFL bulb sounds like a bad idea, how about this: Replace every 100-watt bulb in your house with a white LED (light-emitting diode) light. Unlike the compact fluorescent catastrophe, white LEDs give instant-on, excellent quality light with virtually no wasted heat. And they are on the market today. If we replaced every incandescent light in the country with a white LED light, we could cut the electricity usage of the U.S. in half or said another way, we could slash the country's energy consumption by about $18 billion.

Wow! Now that's a Wall Street concept, well worth running stocks up for and getting excited about. The only problem I see is that the bulbs cost $125 each -- no, there isn't supposed to be a decimal point in there. If the manufacturers can drive costs down 40% a year, which is not an easy curve to stay on, they will get down to 10X as expensive as an incandescent light in about 6 ½ years, and match incandescents in about 11 years. Today, a 100-watt incandescent bulb with a 5600-hour life costs about 72 cents. Because LED bulbs last 50X as long as incandescents, the crossover point for people who change their own bulb comes in about 2012, and the crossover point for commercial users should be just about now.

So, this is a market that will s-l-o-w-l-y build as costs come down and various applications become economic. In the meantime, companies like Cree (CREE) that produce white LEDs will have periodic disappointing earnings reports as their basic business fluctuates. Cree will report higher sales for their June fourth quarter compared to last year, but earnings will be only five or six cents a share compared to last year's 18 cents. They'll start the June 2008 fiscal year with comparably crummy results in the September quarter, and it is hard to imagine them guiding up, given the beating that they are taking on profit margins in the cell phone display backlighting business.

Back in May, I attended the Money Show in Las Vegas. While Vegas is not my favorite place in the world (I think their city motto should be: "Not As Bad As Atlantic City"), I did have a good visit with fellow newsletter advisors and investors. Also, I filmed a short video recording that was posted on MSN Money and the Money Show's main landing page. In the video, I talked with Gary Alexander about two of my favorite biotech companies right now. You can check out the video here.

Guess what? I can fit my body through a hole in a 3x5 index card. And you can, too. Don't believe me? Check out this interesting experiment by Robert Krampf: http://krampf.com/experiments/Science_Experiment11.html.

Today I want to tell you about a company that has unique technology that is perfectly suited for battery-powered, high-volume consumer electronics -- just the sort of device that gets content on demand to the user. QuickLogic (QUIK) and I go way back. The company was founded in 1988 by my old friends at Morgenthaler Ventures, and almost 20 years later, QUIK's founding CEO Tom Hart is still running the show. QUIK is a semiconductor company that began as a competitor to Xilinx and Altera in programmable logic semiconductors -- these chips can be inserted into a hardware programmer that alters them to do different functions based on what the customer wants. For example, if the customer wants to sell a portable device that plays audio and another model that plays both audio and video, the same QuickLogic chip can be programmed for both of these uses.

The demand for programmable logic semiconductors is growing, and many customers, like Dell (DELL) and Apple (AAPL), are turning to companies like Xilinx (XLNX), Altera (ALTR) and QuickLogic for their chips. The reason for this transition is that it is very expensive to make full-custom chips to do each of the desired tasks. So companies like QuickLogic that made chips that can be altered according to the customers' needs are profiting greatly from this change.

What makes QUIK so interesting right now is that on April 25, QuickLogic introduced the new product that I've been waiting for -- the ArcticLink. It is based on the PolarPro, but in addition to low power consumption, it adds important functions like USB 2.0 and Bluetooth support. These added capabilities are going to drive sales for ArcticLink, as USB and Bluetooth markets are expanding at a rapid clip. The market research firm iSupply projects that by 2010 there will be 750 million handheld units shipped that include USB. The second most popular connection will be Bluetooth, which iSupply projects will be included in 680 million handheld units by 2010. So as you can see, ArcticLink is going to be a huge product for QuickLogic.

I don't think most people realize the power of China. Chinese exports hit $103 billion in June, their first month over $100 billion. Their official foreign exchange reserves now stand at $1.333 trillion dollars. But what is really amazing is that their quarterly inflow is accelerating. Just like the 1800s, when the UK was the manufacturing floor to the world or the 1900s, when the U.S. held that title, the cash is pouring in the door to the manufacturing superpower of the 2000s -- China. Here is the recent quarterly growth in their foreign reserves:

Q4:05 +$49.9 billion
Q1:06 +$56.2 billion
Q2:06 +$66.0 billion
Q3:06 +$46.8 billion
Q4:06 +$78.4 billion
Q1:07 +$135.7 billion
Q2:07 +$130.6 billion

During the first half of the year, they have accumulated $266.3 billion in U.S. dollars, more than all of last year ($247.3 billion). They are banking half a trillion dollars a year, which they have to put somewhere.

So, it was with a mixture of sadness and amusement that I read that our Housing and Urban Development (HUD) Secretary is in Beijing, asking (begging?) them to buy U.S. mortgage-backed securities. So, let's see, we want them to buy securities in a collapsing sector, and then the U.S. Congress is demanding that they revalue the yuan by 20%, guaranteeing China a 20% purchasing power loss on whatever they buy, on top of whatever the credit losses are. If the HUD Secretary can have this conversation with a straight face, it certainly explains why he is qualified for his job and I am not.

The worldwide personal-computer market grew at a better-than-expected-by-Wall-Street rate in the June quarter, with shipments up about 12%, according to market researcher IDC. As you know, I have been estimating PC sales above the Street for this year, with the strength coming mostly in the second half. It is a pleasant surprise to see it starting this early.

Intel (INTC) reported an interesting quarter after the close on July 17. The stock hit a 52-week high at $26.33 before the announcement. Sales rose 8% from last year to $8.68 billion, which was just above the consensus estimate for $8.54 billion. Earnings per share excluding a one-time tax gain hit 19 cents, right on the consensus and up 27% from last year.

The reason earnings only hit the consensus, even though revenues came in better than expected, was that product gross margins fell to 46.9% of revenues, at the low end of guidance. The biggest problem was weak demand for NOR flash memory chips, which are mainly used in cell phones. But Intel is in the process of fixing this problem by spinning off this money-losing division to a joint venture. The company also announced that microprocessor selling prices were a little lower than expected. The stock dropped $1.27 yesterday in reaction to the report.

But the really important points that Wall Street is missing are:

• The NOR flash memory division is irrelevant, even though Intel can't account for it as a discontinued line of business quite yet.


• The company guided for a much better gross profit margin in the September quarter, 52% plus or minus two points.


• The company guided for $9.0 to $9.6 billion in sales, in line with the consensus.


• The last time Intel reported a gross margin this low, in the third quarter of 2002, their operating profit margin was 16.5%. This time it was 19.3%. That means they have delivered on their promise to get leaner and more productive, holding down expenses. Also, Intel reduced its inventory by 5.5% (almost $250 million) from the March quarter. That is the first sequential quarterly reduction in inventory since 2003, and inventory reductions almost always have a negative impact on gross margins.


• Intel's manufacturing efficiency is running far enough above expectations that the company will reduce its 2007 capital spending budget.

Amgen (AMGN) reported after the close on July 27. They did $3.73 billion in sales and $1.12 a share, above Wall Street's expectations for $3.68 billion and $1.06. Of course, the main issues overhanging the stock are Medicare reimbursement and any possible further FDA action against Epogen and Aranesp, but I don't expect them to be able to say much about either of those yet.

In the Sicko department, on July 2 Blue Shield of California changed its reimbursement policy for Aranesp and Epogen from coverage for patients with hemoglobin levels of 11 grams per deciliter (gm/dl) of blood or less, down to 9 gm/dl or less. (Patients with heart conditions would be covered at 10 gm/dl or less.) Essentially, they tried to get doctors to wait to give the medication until an anemic cancer patient's hemoglobin level fell to levels closer to the cutoff for needing a blood transfusion. Doctors protested loudly, and yesterday Blue Shield reversed part of the policy by raising the level to 10 gm/dl for everybody. Blue Shield has three million members in California, and I expect doctors to continue to prescribe Aranesp and Epogen as needed. These two drugs were the top selling biotech drugs in 2006, with sales of $7.2 billion, and Johnson & Johnson's Procrit (Epogen by another name) was fifth with $2.9 billion in sales. This is all about money, not good medicine.