If being wrong about the value of fiber-optic cable wasn't bad enough, Gilder predicts Wi-Fi will be killed by 3G: Gilder's latest predictions have proven right in the short term, it turns out, according to Forbes, which notes that a service tracking his model portfolio has had high returns. (Let's remember that great phrase, though, which is that historical returns don't predict future performance.) Watch the sketchy video linked from the article--shot outdoors with little editing, so it doesn't do him justice, honestly. I've never agreed with parts of what he's said more.
Gilder specifically says that 1xEvDO will kill Wi-Fi, which can't be the case because 1xEvDO is too asymmetrical: it has a typical upload speed of 50 kilobits per second (Kbps) and download speeds of 200 to 400 Kbps on average. Compare this with the service being rolled out across the downtown in Salem, Massachusettes, which is deploying three 6 Mbps downstream, 1 Mbps upstream ADSL connections as part of SalemOpen.net for a cost of about $180 per month total. There's a lot more bandwidth potential in the conventional wireline side using plain copper.
Gilder has a lot to say that's correct, which is that other countries--he especially notes Korea--have massively more bandwidth available to homes and businesses in the wireline and wireless markets. This is due to something he doesn't mention: cost. South Korean and Japanese firms particularly have managed to roll out extremely high-speed DSL and cell services often at the same price we pay for tiny pipes in the U.S. I'm still not entirely sure how they pull this off.
Regulation, as he points out, might be one issue, but it's also the market climate: a lot of money was spent by U.S. companies pursuing Gilder's last vision, and thus the resources that could have been devoted to pursuing inexpensive DSL were spent in overbuilding competing fiber-optic networks and in messy acquisitions at fraudulent valuations.
Gilder's observations on the Wi-Fi side show that as a holder of Qualcomm's stock, he's not looking closely at the current generation of Wi-Fi and related wireless deployments. He notes in the video for this story, that there is "small coverage and not all that much bandwidth [in] the average Wi-Fi hotspot." As noted above, DSL service has become cheap enough to offer enormous increases in speed, and a number of cities are deploying partial or complete Wi-Fi service using mesh-based or smart antenna technology. This trend seems to have legs rather than slowing down, implying that the deployments actually work in cities that try it.
Because Wi-Fi service can have many high-speed links and has such an overall local area network capacity, it's possible to have an enormous number of users of a larger network with injections of bandwidth all over the place rather than relying on the limited spectrum--there's that regulation again--that cellular operators have to dole out for 2.5G and 3G networks.
Verizon Wireless told me in an interview recently that their rollout isn't quite an experiment, but they're chary about offering any exotic services, like a network sharing offering such as the Junxion Box (which I wrote about yesterday). They and other carriers are investing a lot in networks that they're not entirely sure what the capacity is once users start to sign up in droves. They want a clean network to see what sticks, and it's clear with so much near-term competition prices will drop.
The capacity is surely quite high, but we get back to the issue of how much bandwidth is needed in different scenarios. 3G will be ubiquitous, which is useful for all kinds of purposes, but most people work only in certain kinds of spaces in which Wi-Fi will be available with higher upload and download speeds than can be relied on in the cellular network over the next two to three years domestically.
Just remember that 50 to maybe 100 Kbps upload limit on an unloaded network on the fastest cell data trunks that will roll out in the next 12 months, and a 512K symmetrical DSL line or, better, a 6 Mbps down/1 Mbps up line doesn't seem quite so slow. Gilder says that 1xEV-DO offers "up to 3.1 Mbps connection." Sure, if there's no one else in your cell and you're on top of the tower.
Cell operators may be paying Qualcomm large patent royalties, but they're also building Wi-Fi networks and reselling existing networks. It's not an either/or: it's a both. [link via TechDirt]