BusinessWeek has a good roundup of this recent story here. We know that people shopping for an iPhone online using NYC zip codes were unable to buy one for a day or so. This is useful synecdoche – a big story in a small vignette, a part referring to a whole.
Explanations:
1. “New York wasn’t ready for the iPhone,” because AT&T “lacks a sufficient number of cell towers to meet the heavy data demands imposed upon the network by iPhone users.” (Customer service rep and BusinessWeek.)
2. Fraudulent activity – heavy ordering of phones to be sent to particular addresses, so AT&T halted sales to all relevant zipcodes until it could fix the situation. Phones could be sent to other countries after having been unlocked.
3. Credit card fraud, not necessarily associated with unlocking phones.
It’s well-known that AT&T’s cell capacity isn’t able adequately to handle use patterns in urban areas. Calls get dropped, data moves slowly, and subscribers are frustrated.
What’s the answer?
One move is “more spectrum,” and FCC/NTIA are on the hunt for commercial and federal spectrum that can be repurposed for wireless high-speed Internet access. (They’re also working to encourage sharing of both commercial and federal spectrum.) This won’t be easy – there are hold-ups and relocation costs at every turn. And it’s not clear that the two big players (AT&T and Verizon) should be allowed to buy up all the spectrum that is cleared, because their holdings are so large already that they’re able to stomp on competitors.
But another key move has to be “more towers,” “more fiber to more towers,” and “more investment in infrastructure generally.”
AT&T does its level best to keep its dividend up (always, year after year) and its available cash numbers high. But that has to mean that the company isn’t investing as much as it could in improving its high-speed Internet access networks, both wired and wireless.
AT&T itself has remarkable margins for its wireless service. According to the most recent SEC documents I could locate, its EBITDA margin (operating cash flow divided by total revenue, so the percentage of a company’s revenue that is remaining after operating expenses) for wireless services is about 40%.
It’s hard to tell how much AT&T is actually spending on towers and equipment – they lump all of their capital expenditures for all of their services together (not just high-speed Internet access), and say that they spent 11B over the first three quarters of this year. FreePress has done an analyis that notes that capital expenditures by the wireless industry as a percentage of revenue are down over the last few years. This disinvestment is happening at the same time that profits are going up, costs are going down, and there isn’t very much regulation.
So if we assume CAPEX is going down and ARPU is going up for AT&T, where does that leave high-speed Internet access for the rest of us?
In Japan, NTT doesn’t pay dividends and they’re wiring the country for 50-100mb. I’m confident that the FCC is looking at international comparisons.