I’ve talked about the profit problems of network operators for years now, simply because they’ve talked to me about those problems for years. In rough terms, revenue per bit has declined by 50% per year for five and a half years running, and this decline means that profits on infrastructure have dropped sharply even though operators have put price pressure on vendors. All of the challenges of networking that we hear about, including Ciena’s miss yesterday, can be directly attributed to the simple fact that you can’t pay more to produce something that keeps selling for less and less.
Today, Morgan Stanley bucked a shaky consensus that wireless profits too were under pressure but still above ground, saying that their model shows that post-pay ARPU for the US giants will likely trend negative by the end of 2012. With seven-plus-year financial inertia to contend with, operators can draw the curves themselves based on their inside numbers, and it’s clear that they can see where they end up. The AT&T drive to consolidate via T-Mobile is, as I’ve said, an indication that economies of scale are now required in wireless, which isn’t the sign of a market with profit growth still baked in.
Another interesting data point is Verizon’s deal with the cable guys. The latest speculation on the Street is that what Verizon will do next will be a shocker; they’ll start reselling cable broadband and TV even inside their own region! Why? Because the profit on DSL is simply not there, and FiOS can’t be made profitable for any more of their footprint than the current 20-million-homes-passed target, at least under the current price scenarios. I’ve noted for some time that the real competition between carrier and cableco wasn’t fiber versus cable but cable versus DSL, and pressure from the FCC to upspeed the baseline broadband rates is only going to make DSL harder to justify.
Some of the Street speculates that Verizon might then shift to using their extra spectrum to replace DSL with wireless, but I think that’s a misunderstanding of intent. It would only hasten the price commoditization of wireless. What I think Verizon is doing is replacing “wire” wireline, meaning copper and DSL with…nothing. They’d love all those customers to shift to cable because they can’t service them profitably. After all, hasn’t Verizon been selling off areas where there was limited upside for copper loop already?
From an equipment vendor perspective this is alarming, not that equipment vendors haven’t been reading the tea leaves on this for half a decade already. The drive to exploit marginally free bandwidth has created enormous profits for a few VCs and early investors, but it’s destabilized the underlying bit industry. Operators have seen this coming and worked to gain traction in those new higher-layer services but without any real vendor support. Now, our survey says, it’s right on the edge of too late, and the Street’s indicators seem to back this up. We are going to see a major shift in capex next year, a shift toward cloud and IT equipment and away from transport and connection. What will be hit the most, says our model with striking irony, is routing. Most valuable Internet traffic today is metro, which is optics and switching. Yet Ciena shows that’s no great business either. So where do operators invest? Where it makes them money, which is up above and not in the network. Our model says this will be visible even next year and a juggernaut trend in 2013.