The Gilder Friday Letter of April 4, 2008
gilder.com
--
Select abstracts from today's letter:
The Week / Meeting the Exaflood Challenge
George Gilder, Gilder Telecosm Forum: A report last month from AT&T (T) on the traffic from the some one billion devices attached to its networks affirms that exabytes are already mounting, with consumer traffic up 145% over the past two years, business traffic up 60%, and wireless data up 300% last year over 2006 even before the launch of the iPhone. Processing the exaflood is estimated to require some $137 billion in new capital spending on global network infrastructure over the next three years.
I explained to the OFC audience (in late-February) that nearly all the scores of optics component players are treating the technology as mature and are targeting their incremental advances on the few giant systems vendors, chiefly Cisco (CSCO) and Alcatel-Lucent (ALU), who in turn supply a relatively small number of carriers, such as the aforementioned AT&T and Verizon (VZ). Multiple suppliers of incremental advances cannot win substantial margins from oligopsonies or lucrative multiples from stock markets. We still approve of Finisar (FNSR) and Opnext (OPXT), among others. But we are not touting them with a target of tenfold gains.
As our old friend Clayton Christensen explains, in a settled industry it makes sense to go modular as these companies are doing. Modularity reigns, with every interface standardized for multiple suppliers of commodity components. But in a dynamic industry where the existing systems will have to struggle to contain the impending exafloods of traffic, it makes sense to optimize every interface in an integrated system designed top down to achieve the highest possible performance.
Optics is embroiled in a tempest of change, with no settled network or computer architecture and no systems standard. The industry has not decided on whether to adopt ever denser wavelength division multiplexing of scores of wavelengths on every fiber thread or to pile up more and more bits in a single terabit wavelength stream. It has not decided on whether to switch packets in nanoseconds or wavelengths in milliseconds. It has not decided how to amplify the streams or how to add and drop them. It has not resolved on whether to modulate its signals inside the laser or outside it. It has not chosen between loop architectures and mesh architectures. We could continue. But the point is already clear. This industry is anything but settled.
In a dynamic industry, the profits go to the systems innovators. Systems innovators exploit their component inventions to enable new and radically superior systemic capabilities. Infinera (INFN) is a company that has followed this path with photonic integrated circuits (PICs) fabricated on indium phosphide wafers and developed into new digital systems. But we believe that their claimed pace of advance (doubling capacity every three years) is about half of Moore's law and inadequate to meet the challenge of the exaflood or even the pace of advance of wavelength division muxing optical technology. They are an impressive company with brilliant leadership and a good systems strategy, but they are unlikely to achieve a tenfold appreciation because of their tricky indium phosphide substrates and their dependence on continually converting their optical wavelengths to digital packets and back at every node across the network.
Learn which optics stock George Gilder most strongly supports. Register to become a Gilder Telecosm Forum member today: gildertech.com
--
Friday Blogger Bonus /
John Rutledge, www.Rutledgeblog.com: This is important. India has crossed the 250 million mark with the addition of 8.53 million mobile phone subscribers in February. India will become the second largest wireless network in the world after China in the first half of April 2008. (Hint: That makes the U.S. #3.)
Read the full article in the Economic Times:
telecomasia.net
Future growth of income, productivity and jobs will depend on who has the best information and communications technology, because high-speed communications allows the economy to perform as a massive parallel-processing information network. Hats off to China and India for making R&D and investment in new networks a priority. Wouldn't hurt if the U.S. government were a positive force for investment here too.
Continue Reading:
rutledgeblog.com
--
FRIDAY LETTER BOOK OF THE MONTH
George Gilder: David Berlinski's new book, The Devil's Delusion, says April Fools to the barbarians of specialization who currently dominate the philosophy of science. A sophisticated scientist and mathematician himself, he shows that not only is atheism totally unsupported by any scientific evidence, but that atheism undermines the pursuit of scientific truth itself by inducing physicists to pursue reductionist goals (particles and strings) that yield no wisdom or truth about the universe.
Scoffing at the infiniverses of Richard Dawkins and the alien visitations of Francis Crick and Fred Hoyle, he shows that effective science benefits from a belief in a monotheistic God.
amazon.com
--
Readings /
15th Anniversary: Why the Future Still Needs Us a While Longer
ladycmog.multiply.com
Global High Performers
forbes.com
Lightwave Logic's Terry Turpin to Appear on Panel at Gilder/Forbes Conference
pr-inside.com
With Bleak Jobs Data In Hand, Investors Await Earnings Season
online.wsj.com
What Killed Bear?
forbes.com
Graphene Transistors
technologyreview.com
__________________________________________
The Friday Letter is published weekly for subscribers and friends of Gilder Publishing. If someone you know would enjoy it, please feel free to forward a copy.
To SUBSCRIBE please visit gilder.com
------ |
