Akamai Technologies (AKAM) was clipped after AT&T (T) said that they will build up their Content Delivery Network business. Cowen & Co. downgraded AKAM as a result. Yawn.

Cowen noted that AT&T has such a big network that the average distance between a Content Delivery Network node and the customer could shrink to 100 miles by the end of 2008, compared with Akamai's current 250 miles. Double yawn.

AT&T has been in the Content Delivery Network business for four years, and gone nowhere. They may have a customer for video somewhere, but I've never read about one, and there aren't any customer case studies on their website. Meanwhile, Akamai sold nothing but dumb caching services four years ago, but they now offer a full suite of services around that, including support for streaming video (a la YouTube), fail-safe delivery, content management systems and software, load balancing, network monitoring, real-time reporting on delivery performance, redundancy, security and optimization.

After Akamai started, by 2000 they had lots of big name competitors who said that they were going into content delivery: Quest, MCI, Sprint and even the then-version of AT&T. After a year or so, they were all gone. Meanwhile, Akamai's average revenue per user grows quarter after quarter. Sure, some customers still sign up for dumb content delivery, but they quickly realize that by using Akamai's other services, they can greatly increase the economic value of what they are sending over the Internet. Quicker delivery of physical packets, which is what AT&T is talking about, is a small part of a total Internet distribution service. I watch Limelight Networks, Level3 and others as they try to compete with AKAM, and I know this market is so huge that there will be more than one winner, but right now I think that AKAM is the way to take advantage of the video explosion.

There's an important shift going on at the FDA that is making me more cautious on biotech companies working on cancer drugs. On December 5, Genentech (DNA) got a turndown from the FDA Oncologic Drugs Advisory Committee (ODAC) to expand the label for Avastin to cover breast cancer patients in conjunction with chemotherapy. Avastin is already FDA-approved for colorectal and lung cancer, and it is no small drug, doing $1.7 billion in sales for the first nine months of 2007. This got my attention, because Genentech normally is very good at getting drugs through the FDA. But the head of the Oncology Division of the FDA, Richard Pazdur, has become very strong politically, to the point that industry insiders now refer to some FDA decisions as "being Pazdurized."

On Avastin, he simply constructed the ODAC panel meeting to get a "no" vote, using the always-reliable "let's wait for more data" excuse. Some people think that the FDA will overrule the panel and approve Avastin on February 23 because the vote was only 5 to 4 against approval. But I think they are really wrong, for two reasons. First, as Pogo used to say, truth is a 5 to 4 vote by the Supreme Court. Second, Pazdur has an agenda. In the past, progression-free survival has been a surrogate endpoint for survival, and a drug that showed progression-free survival for a period of time could get approval. But Pazdur seems to want real survival data, which can take much longer to collect at a point that it will be statistically significant. Essentially, he has said: "I don't care if the tumors stop growing if the drug doesn't extend survival, so show me the survival data."

I have mixed feeling about this. None less than the American Cancer Society has said that if you have a healthy immune system, you won't get cancer. In that sense, cancer is a symptom of an unhealthy immune system, not a disease in itself. Treating the tumor without treating the underlying immune system problem is like bringing a fever down without addressing an underlying infection. It doesn't do much good, and may actually do harm. In the case of the fever, your body is trying to burn out the source of the infection. In the case of cancer, chemotherapy and radiation both suppress the immune system, making it more likely that cancer will recur and be fatal.

On the other hand, this is bad news for cancer biotech companies and their stocks, and may reduce R&D in this area. It simply takes a lot more money and time to prove survival than to prove tumors are not progressing. So a new approach that works might not get pursued just because of the daunting financial requirements to get to approval. In all, we must factor these developments into our decisions to invest in biotech cancer companies going forward.

The International Consumer Electronics Shows kicked off yesterday in Las Vegas. This show brings together a variety of technology companies that are showcasing some of their latest products.

One company, QuickLogic (QUIK) announced last week that it will be showcasing new products and platforms, based on the company's Customer Service Specific Standard Products (CSSPs), at the Consumer Electronics Show. Some of the products and platforms that QUIK will have at the show include:

• A pocket-sized wireless hard disk drive that can stream content to and from a Bluetooth-enabled device, like a laptop or music player;
• An enterprise personal digital assistant that provides high speed data synchronization for external devices, like a keyboard or mouse, and it also uses CompactFlash interface to support other devices, like printers and scanners;
• Two portable navigation devices: One enables MP3 playback and high speed map downloads, while the other is an important part of QUIK's storage solution portfolio;
• A 3G wireless data card;
• Ultra Mobile PC (UMPC)/Ultra Mobile Device (UMD)/Mobile Internet Device (MID), which enables the speed and conversion necessary to connect to external devices, like a television;
• 3G handset development platform that provides the high speed connectivity for data synchronization with devices like laptops or PCs;
• Two portable consumer electronics development systems that show various solutions to QUIK's proven memory system;
• A portable multimedia player TV-out reference design that enables portable devices, like a cell phone or PDA, to have their content displayed on a TV.

For more details on other companies attending the Consumer Electronics Show, check out the website here.

Sequenom (SQNM) is a stock that came public at $26 a share in March 2000, right at the peak of the tech and biotech bubble, based on the excitement around the possibilities from sequencing the human genome. It shot straight up to $171. A year later it was under $10, and six years after that it as trading around at $4.50.

But unlike many of their peers, it looks like Sequenom will make genetic sequencing pay off. Their prenatal Down syndrome product is non-invasive and would replace amniocentesis. There is an alternative method of looking for Down syndrome that we did earlier this year for our little girl who was born last Thursday, which combines an ultrasound scan to look for an abnormally thin neck on the fetus plus a blood test. While it is safer than amnio, it is expensive because it requires an experienced radiologist to interpret the scans.

Sequenom's prenatal screening tests are designed to be more accurate than other tests as well as noninvasive -- about 275,000 women a year are falsely diagnosed as being at risk for carrying a child with Down syndrome, and then put through more invasive tests. The company will do a cystic fibrosis test next, and they can grow to a substantial size just in these two indications.

How is it that Sequenom has been so successful in the aftermath of the tech and biotech crash? Sequenom went into the contract research business. They developed their own tools for more accurate and quicker genetic analytical work, and then marketed the tool as their MassARRAY product. They now have 200 customers, including 14 new ones that were added in the September quarter. Sales grew from $19.4 million in 2005 to $28.5 million in 2006, and I am looking for $40 million in 2007. I expect them to beat $55 million in 2008. So while the company is still losing money as they research and develop their prenatal diagnostic tests, they have trimmed the losses from $26.5 million in 2005 to $17.6 million in 2006 and around $13 million in 2007. They should be down in single digits, around $7 million to $8 million, in 2008.

Sequenom also did a $30 million private placement in the September quarter, and that cash should carry them through to profitability. Once the first two prenatal tests are approved, they can use the cash flow to address a huge array of other disorders with non-invasive tests, both within and beyond the prenatal markets. And they have finished early work on a test for genetic material from malignant cells to diagnose cancer before any tumor can be detected. Because we all have some cancer cells circulating in our bodies all the time, this test will have to be sensitive but not too sensitive.

So Sequenom came back from the near-dead after the tech crash with the help of good management that focused the new genetic science on practical diagnostic opportunities that have a shorter path to approval than drugs. I'm expecting good news from their Down syndrome program to drive the stock higher.

iRobot (IRBT) had an eventful December. But before we get to all the details, you need to understand what occurred in the fall of 2007. Back in August, iRobot had sued Robotic FX on patent infringement and trade secret misappropriation. Then the smaller competitor snatched up a military contract from iRobot in September.

Well, iRobot didn't take any of this sitting down. In mid-December, the company was awarded the $286 million contract with the U.S. Army for 3,000 new PackBots, a bomb disarming robot -- the Army backed out of its deal with Robotic FX. iRobot has already delivered more than 1,200 PackBots to the United States military, so this was another big win for the company.

Then only a few short months after iRobot sued Robotic FX, the results of the suit not only came back in iRobot's favor, but the courts actually disbanded Robotic FX on December 21. I've never seen anything like that! This not only takes away one of iRobot's biggest competitors, it also puts the company as one of the go-to names for defense contractors -- any defense contractors and law enforcement agencies that were customers of Robotic FX will now be knocking on iRobot's door.

Energy Conversion Devices (ENER) has a subsidiary, United Solar Ovonic, that makes the highest efficiency thin-film photovoltaic (PV) cells for solar-generated electricity. Solar modules traditionally are made of polysilicon cells that are mounted in a metal frame, covered with protective glass and wired together for connection to each other in PV arrays. It's an expensive, mechanical assembly process. Combined with a shortage of polysilicon, it has kept the cost of generating solar power about five times as high as the petroleum-based alternatives.

In the mid-1990s, AstroPower developed a continuous film coating process to put the polysilicon on films that could be attached to a prewired frame and shipped as a module. I had these modules in an array on my ranch in Half Moon Bay, and they worked great. But they were still about triple the cost of conventional electricity, and if it hadn't been for a 70% co-payment by Pacific Gas & Electric and the taxpayers of the State of California, I wouldn't have done it.

United Solar Ovonic takes that idea a step further to deposit a thin film of amorphous silicon, not the hard-to-get polysilicon, in a continuous process that produces a roofing material usable in new or retrofit construction. It has about 16% efficiency these days, and it is cost-effective under the current Energy Bill tax incentive of 30% of cost.

The reason I brings this up is because recently, I've had a couple of questions about nanosolar, applying nanotechnology to greatly reduce or even eliminate the need for silicon. The other week, a technology breakthrough at an Israeli nanotechnology institute caught my eye. They used conductive glass -- no silicon -- and nanodots of platinum to produce PV cells 100 times as large as current cells. They reduced the amount of expensive platinum needed to make a solar cell function by 97.5%.

The institute is working with a private Israeli company, Orion Solar, to develop the technology. This is just one of hundreds of similar innovative programs at research institutes, universities and private companies happening today, all around the world.

The China Miracle looks to be in deep trouble in the stock market, just as I forecast, but I do expect them to rally back with the U.S. market to new highs after the Beijing Summer Olympics. Look at these recent IPO catastrophes:

Xinyuan Real Estate (XIN): IPO date was December 11, with the price at $14. As of January 16, the price was at $9.90 -- -29% loss.

ChinaEdu (CEDU): IPO date was December 10, with price at $10. As of January 16, the price was at $6.85 -- a -32% loss.

China Nepstar Chain Drug (NPD): IPO date was November 8, with the price at $16.20. As of January 16, the prce was at $13.94 -- a -14% loss.

Giant Interactive (GA): IPO date was October 31, with a price at $15.50. As of January 16, the price was at $11.44 -- a -26% loss.

Earlier Chinese IPOs, like Xinhua Finance Media (XFML) and Noah Education (NED), that went public earlier in 2007 have seen their stocks cut in half.

But China keeps growing, with 11.2% growth just reported for the fourth quarter. That's the fastest growth in 13 years, and the government is putting in price controls to stop inflation, which is also at the fastest rate in 13 years. Yet they continue to print money like crazy, trying to keep up with Helicopter Ben so that the yuan doesn't rise too quickly against the dollar. It's a very volatile situation, and if there is a serious recession in the U.S., the Chinese market will blow up. But I don't see a recession in the U.S. yet, just very slow growth for a couple of quarters, and in that environment the country can hold it together through the Olympics. But I would not be surprised to see the U.S. market substantially outperform China from the opening of the Olympics through an April 2009 top.