Select Page

In the mid- to late- nineties, as we all know, incumbents and competitors built millions of miles of redundant fiber optic cable, thinking that the opportunities to transit data and voice traffic would represent huge growth in the industry. But one thing seized up: demand hit the ceiling. Most of these networks now lie fallow, unlit, and desperately underutilized. Their architects now oscillate in and out of bankruptcy, junk status, and other death sentences.

One camp suggests that the wells ran dry because of the dreaded LAST MILE problem, that being that if consumers and businesses can’t get adequately high-speed connections to their homes and offices, then what’s the point of having a super-high-speed backbone?

Another camp, and here’s where you’ll find me, believes that the real problem is applications. After nearly 10 years of what will become known as the “Business Internet” we are still left with one of its first tools, email, as the most compelling driver in getting people to adopt and adapt. Email, of course, is not particularly bandwidth-intensive. So the need, then, would seem to be one of developing super-high-speed, high-bitrate applications to fill this glut of pipes and resurrect the industry.

Some telcos, notably SaskTel, are launching Video Over DSL. Others are dabbling in streaming audio and video, telephony, and other realtime applications over the public internet. But Broadcast and Realtime are not fundamentally what IP was built to do, and so for now there is an uncomfortable tension between aspirations and the functional capabilities of the medium. And there’s also that last mile thing.

People have fooled around with peer-to-peer file sharing, much to the chagrin of the RIAA. That drove many dialup users to broadband. Could peer-to-peer, born again within some kind of legal copyright-friendly (but costly) framework, resurrect demand? I suspect it will need to. But the bits still cost too much money — the whole pricing structure of the internet industry is designed for small bits, not big bits. Pricing for big bits is way too high relative to their cultural value.

How does this relate to Intel? Well, it’s an allegory. And a common problem.

Processor speed has far outpaced demand. Even the hungry Windoze XP doesn’t make a 3GHz Pentium 4 break a sweat (not even at Gersham’s house). It’s also too expensive. But more importantly, the problem is applications. While XP is great, it’s an operating system — it doesn’t actually DO anything. Applications drive us toward a particular operating system, which in turn drives us to buy the hardware that best supports it.

Processing capacity, or cycles, (see Moore’s Law) is now dramatically outpacing processing requirements in the popular applications we all use. They can only add so many processor-hungry, real-time features (like Mr. Paperclip) to Microsoft Word before it gets annoying. Email, in addition to not requiring lots of bandwidth, doesn’t seem to require lots of horsepower either.

So what applications will motivate my mother, who happily chugs along with a Pentium II/166 with 64MB of RAM, surfing the web and reading her email, to jump on the bandwagon and buy a newer faster better PC?

As a good friend said to me last week, 95% of what the internet will be in 5 years doesn’t even exist today. And that is a message of hope, I think — but I’m torn in trying to critique his point. Does it represent unbridled optimism? Or is it a healthy reflection on where we are at today?

In the interim, though, cycles — like bits, are now officially commodities. And, like the three-year speed bump in the network buildout (which we are presently nearly two years into), there’s very little to sustain the guys who sell us cycles.

Now is the time to innovate; break the mold and forge new models. The good news is that in this down market there’s a bonus: thanks to commodification, bits and cycles are cheap. And this confluence of vacuums in both transport and processing capacity is the opportunity to remedy its own cause. All that’s needed is effort, and new thinking.


———- ap_on_hi_te/earns_intel&printer=1 Intel Stock Plunges Wed Oct 16,10:53 AM ET

By MATTHEW FORDAHL, AP Technology Writer

SAN JOSE, Calif. (AP) – Intel Corp. stock plunged Wednesday after the company missed Wall Street’s third-quarter earnings expectations after a lackluster back-to-school season and weak corporate demand for computers.

The world’s largest chip maker also said Tuesday it does not believe the economy has bounced back enough to support much improvement in the fourth quarter from holiday sales.

“There is some demand out there — but not as strong as you would typically see in the second half of the year,” said Andy Bryant, Intel’s chief financial officer.

Intel released its earnings after the stock market’s close on Tuesday. Shares sank 16.8 percent, or $2.77 each, to $13.75 in early trading Wednesday on the Nasdaq stock market.

For the three months ended Sept. 28, Intel earned $686 million, or 10 cents a share, compared with a profit of $106 million, or 2 cents a share, in the same period last year.

Excluding special items, the company earned $768 million, or 11 cents per share, compared with a profit of $655 million, or 10 cents a share, in the third quarter last year.

Intel reported third-quarter revenues of $6.5 billion, roughly flat with sales of $6.55 billion of last year.

Analysts were expecting a third-quarter profit of 13 cents per share on sales of $6.52 billion, according to a survey by Thomson First Call.

In July, the company estimated third-quarter sales would be between $6.3 billion and $6.9 billion. Last month, it said revenue would fall to the lower end of that range.

Intel is not alone. Earlier this month, rival Advanced Micro Devices Inc. warned that its revenues would fall about $100 million short of expectations and it expected to post a “substantial operating loss for the quarter.”

Intel announced in July it was cutting about 4,000 jobs. On Tuesday, Bryant said Intel had not achieved its savings targets.

Typically, Intel and other semiconductor makers see as much as a 20 percent boost in sales during the holiday-buying season. Intel said it expects fourth-quarter revenues to be between $6.5 billion and $6.9 billion — at best, a 6.2 percent jump from the current quarter.

The semiconductor business is now in its second year of dismal results as demand for personal computers have fallen short and up-and-coming uses for chips haven’t quite taken off.

Microprocessors, which make up the brains of computers and the bulk of Intel’s business, have yet to return to the growth rates seen in the late 1990s.

Analysts say consumers and businesses don’t see a need to pay top dollar for the fastest Pentium 4 when slower models can handle typical tasks just as well.

Moreover, the bulk of any gains in Intel’s processor business — typically at the expense of AMD — are in the low-end products. That tends to squeeze average selling prices and margins.

“There are not a lot of people running out there to buy,” said Eric Ross, an analyst at Investec. “When people are buying things, they’re buying cheaper PCs.”

During the third quarter, Intel introduced 18 microprocessors for desktops, notebooks and servers. Intel’s top Pentium 4 now runs at 2.8 gigahertz. A 3.06 GHz machine is expected in the fourth quarter.

The company also said it expects no material impact in the third quarter from a federal judge’s ruling in a patent dispute that could cost Intel as much as $250 million.

Intel said it has asked the Texas judge to reconsider his ruling in favor of Intergraph Corp., which claimed its patents were infringed by Intel’s Itanium server processor.

More information on the impact of the decision will be released with Intel’s November regulatory filing, the company said.