__________________________________________________

―  THE FRIDAY LETTER  ―

(emailed weekly, from Gilder Publishing,
for friends and subscribers)

__________________________________________________

 | http://www.gilder.com/ | Issue 241.0/April 7, 2006

SIGN-UP A FRIEND FOR FREE!
 

 

HEADLINES:

▪ The Week / The Semi Industry’s Two-edged Sword
▪ Friday Feature / Parrotheads Working For The Spooks
▪ Friday Bonus / Last Mile: Triumphs and Threats
▪ Book of the Month /
▪ Upcoming Events /
▪ Readings /

 

The Week / The Semi Industry’s Two-edged Sword
‾‾‾‾‾‾‾‾‾‾‾‾‾
Excerpted from a recent post to the www.Gildertech.com subscriber-only message board.

George Gilder (04/02/06):  In the semiconductor industry, production volumes rule. This is a two-edge sword for EZchip (LNOP). The problem with application-specific integrated circuits (ASICs) is that their non-recurring expenses (design, engineering, and debugging costs) are steadily rising with each generation, requiring ever-larger volumes, thus diminishing the number of possible markets. This gives EZ its opportunity. In the currently turbulent network scene, programmable network processors have become ascendant. With its NP2, a great product, EZ is now moving to line cards, which give it 10x larger volumes than the central service cards that it previously supplied.

 

As the market settles and the network standardizes, the need for programmability steadily diminishes and customers will turn toward ASICs (as Bill Rossi and Rich Heaton of the well financed Greenfield Networks assume from their extensive experience at Cisco). More specifically, customers turn to application-specific standard products (ASSPs) of the kind that currently dominate the Ethernet enterprise market from Broadcom (BRCM) (its SGX line) and Marvell (MRVL).

 

Then for the high performance and fast changing niches, the market will go to field programmable gate arrays (FPGAs) as Nick Tredennick still predicts and as Xilinx (XNLX) and Altera (ALTR) are now betting heavily. At the Linley Group conference, Xilinx launched an impressive field programmable traffic manager.

 

At present, FPGAs are far too large cumbersome and costly to menace EZchip. (A Xilinx Virtex-based traffic manager costs between $500 and $1000.) But, they are beginning to contend. At a certain point, sometime during the next five years, the capabilities of the FPGAs will enable sufficiently large volume production to bring down the price to a competitive point, wiping out all but the highest performance ASICs and NPUs, and if Nick has his way DSPs (digital signal processors) and CPUs as well.

 

Among the direct rivals, Xelerated is widely dismissed as a goner, but Broadcom-Sandburst is an increasingly potent rival that has yet to make much of a dent but cannot be counted out, and Marvell is closely aligned with EZ.

 

Down-market, with the ever-changing protocol zoo of PONs and DSL variations, programmble NPUs are needed, but they are being supplied by Wintegra (soon to go public), Intel (INTC), Agere (AGR), Applied Micro. (AMCC), and others.

To read more posts by George Gilder and to access to the April issue of the Gilder Technology Report logon at www.Gildertech.com with your GTR subscriber password.  

WHY LOGON TO THE www.Gildertech.com SUBSCRIBER MESSAGE BOARD? GTR Subscriber 04/04/06): I've been able to double the value of my portfolio in the last 90 days… largely in part to all the great research being done by members of today's Gilder Board.  So, I thank all you good folks for sharing information you weren't really required to share.  GTR Subscriber (04/05/06): Anybody get the feeling this message board is about to become a lot more popular? The quality of the content on this board over the last 5 years has been extremely insightful. Please keep the posts coming.

Friday Feature / The Parrotheads Working For The Spooks
‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
Andrew Gillies, Forbes.com (04/05/06): If you like a touch of irreverence in a chief executive, you'll appreciate Essex Chief Executive Leonard Moodispaw. The company's stock symbol, KEYW, pays tribute to his fondness for Key West and the music of singer Jimmy Buffett. According to his official bio, Moodispaw also "enjoys chocolate" and "is growing older but not up."

Jimmy Buffett? Chocolate? Not exactly the sort of stuff you'd expect to hear about an executive in Essex's line of work. Essex uses optical processing and proprietary digital algorithms to analyze signals, images and other big chunks of data for federal defense and intelligence outfits. The Columbia, Md., company can't even reveal the identity of many of its customers, because the contracts are classified.

Speak with Essex brass about strategy, however, and Moodispaw's line about not growing up starts to make more sense. Indeed, company execs see that youthful spirit as one of its most important assets in a tough intelligence-tech services market.

It's a market that's been good to Essex recently. Since 2000, Essex's revenues have grown to $159 million from $3.3 million, a breakneck annualized rate of 117%. For its fiscal year ended last December, the company's net profits nearly quadrupled to $8.6 million.

All that growth has Wall Street panting. Essex stock now trades at 2.9 times sales and 58 times trailing 12-month earnings per share. Among five equity analysts covering the stock, the consensus for three- to five-year earnings growth stands at 37% annualized.

But those numbers look stretched compared with other companies selling information technology services to the federal government. Accenture and SRA International, for example, are now valued at 1 and 1.9 times sales, respectively. And those companies sport higher operating margins--or earnings before interest, taxes and depreciation expressed as a percentage of sales--than Essex.

Making things hairier for Essex, its business mix isn't terribly diverse. In 2005, the company drew 69% of its sales from one intelligence customer; a third of revenues came from one contract.

So what does Moodispaw's belief in "growing older but not up" have to do with all this? For one, company executives insist that an entrepreneurial spirit and technological innovation are key to the company's success. "We can't just focus on big programs," says Edwin Jaehne, Essex's chief strategy officer. "Part of keeping that innovative edge is looking at the toughest problems."

Example: Jaehne says his chief scientist, Terry M. Turpin, is now tinkering with ways to apply Essex's expertise in optical processing, or the use of light waves to process information, with its work in cognitive computing or artificial intelligence. The combination, Jaehne suggests, could bump up the processing capability of artificial intelligence by a factor of 50,000 or more.

Other new initiatives are more mundane but could broaden the revenue mix in the shorter term … [Read On]

Read the Complete Article:
http://www.forbes.com/2006/04/04/essex-intelligence-security-cx_ag_0405beltway_print.html

 

Friday Letter Book of the Month Houses of the Berkshires, 1870-1930 (The Architecture of Leisure) by Richard S. Jackson and Cornelia Brooke Gilder Houses of the Berkshires surveys 35 of the most renowned resort estates of Lenox and Stockbridge, Massachusetts, designed by nationally-known firms, including McKim, Mead & White, Carrere and Hastings, and Frederick Law Olmsted. The client list included lawyer and Ambassador Joseph Choate, inventor George Westinghouse, and novelist-social observer Edith Wharton. Illustrated with over 300 archival photographs and floor plans, the book uniquely chronicles a distinctive social and literary colony and now vanished way of life. (314 pages, Hardcover) Order Your Copy Today on Amazon.com: http://www.amazon.com/exec/obidos/ASIN/092649435X/gilderpublish-20

Friday Bonus / Last Mile: Triumphs and Threats
‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
A Superspeed Connection Covering That Vexing Last Mile, New York Times (04/05/06):
IT'S the bottleneck that bedevils businesses everywhere: the "last mile" of wire connecting offices to a phone company's switching station. In most of America, one carrier — typically a Bell company — maintains the high-capacity data and calls over that route, which means little competition and high prices.

But Lou Slaughter, the chief executive of GigaBeam, a technology company in Herndon, Va., has come up with an alternative: millimeter-wave technology, which transmits data over wireless connections at one gigabit per second — 1,000 times as fast as a DSL connection.

The idea of commercial high-speed wireless connections is not new. Companies like Sprint, Winstar and Teligent made stabs at developing similar line-of-sight technology a few years ago, but the projects were stopped for various reasons.

GigaBeam's proprietary technology, though, may be a more elegant — and potentially more durable — solution. Unlike many other high-speed wireless technologies, including Wi-Fi and WiMax, GigaBeam's pencil-thin signals travel over licensed spectrum, which is regulated and thus requires paying fees and much paperwork. Using this spectrum also means that GigaBeam's signals do not interfere with other signals and are thus more secure … [Read On]

Read the complete New York Times Article (PDF file):
http://www.gigabeam.com/admin/InTheNewsManagement/PDFs/Telecommunications2.pdf
__________________________________________

Re-regulation Threatens Last Mile

George Gilder (04/05/06): Separating content and conduit by force – “net neutrality” – unnecessarily exalts regulation and elevates bureaucrats over market forces. The million-word re-regulation of the industry that was the Telecom Act of 1996 resulted in the Great Telecom and Technology Crash of 2000-2003. Net neutrality risks a replay of this carnival of lawyers, micro-mis-management by apparatchik, price controls, the socialization of infrastructure and the screeching halt of innovation and investment in the "last-mile" local loop.

For years the doomsayers have said telecom will contrive content-conduit plays like the cable industry, that they will thereby reap profits from broadband content and that it will be the end of civilization as we know it. They forget that content and conduit are naturally separate. If you have the best content, you want it on everyone’s conduit. If you have the best conduit, you want everyone’s content on it. There are absolutely no synergies between creating attractive and original content and building powerful and available broadband networks. By far the most profitable product in cable is not their pathetic TV content with its endless clutter of ads and spam but their open Internet service. The market will continue to push telecom and cable to provide consumers with more choice not less.

Now with the worst regulatory excesses finally behind us, last-mile telecom investment finally poised take off, and an expectation of a deregulatory path to the future, is Congress seriously contemplating re-regulating the industry again?

Gilder Technology Report subscribers are winning big in ‘06: The 4 companies added to the Gilder Technology Report’s “Telecosm Technologies” list thus far in 2006 have gained an average of 56.5%, including a whopping 139.8% run by Finisar! Get Gilder’s latest pick by subscribing today and downloading the March 2006 issue of the Gilder Technology Report. Click Here Now!!

Upcoming Events /
‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
April 11: Portland
Telecom's Last Frontier
Featuring U.S. Sen. Gordon Smith & Futurist George Gilder
http://www.discovery.org/

APRIL 13: SEATTLE
Is the Blogosphere the Death of the Mainstream Media?
Featuring George Gilder