On the most recent earnings call with investors, AT&T’s EVP and CFO John Stevens reported that AT&T has only 800,000 customers nationwide remaining on traditional DSL. That’s down from 4.5 million DSL customers just four years ago. The company has been working hard to work its way out of the older technology.

The company overall has 15.8 million total broadband customers including a net gain of 82,000 customers in the first quarter. This compares to overall net growth for the year of 2017 of only 114,000 customers. The company has obviously turned the corner and after years of stagnant growth is adding broadband customers again. The overall number of AT&T broadband customers has been stagnant for many years, and if you go nearly a decade the company had 15 million broadband customers, with 14 million on traditional DSL.

The 15 million customers not served by traditional DSL are served directly by fiber-to-the-premises (FTTP) or fiber-to-the-node (FTTN) – the company doesn’t disclose the number on each technology. The FTTN customers in AT&T are served with newer DSL technologies that bond two copper pairs. This technology generally has relatively short copper drops of less than 3,000 feet and can deliver broadband download speeds above 40 Mbps download. AT&T still has a goal to pass 12.5 million possible customers with fiber by the end of 2019, with an eventual goal to pass around 14 million customers.

The AT&T fiber buildout differs drastically from that done by Verizon FiOS. Verizon built to serve large contiguous neighborhoods to enable mass marketing. AT&T instead is concentrating on three different customer segments to reach the desired passings. They are building fiber to business corridors, building fiber to apartment complexes and finally, offering fiber to homes and businesses that are close to their many existing fiber nodes. Homes close enough to one of these nodes can get fiber while those only a block away probably can’t. It’s an interesting strategy that doesn’t lend itself to mass marketing, which is probably why the press has not been flooded with stories of the company’s fiber expansion. With this buildout strategy I assume the company has a highly targeted marketing effort that reaches out only to locations it can easily reach with fiber.

To a large degree AT&T’s entire fiber strategy is one of cherry picking. They are staying disciplined and are extending fiber to locations that are near to their huge existing fiber networks that were built to reach large businesses, cell sites, schools, etc. I work across the country and I’ve encountered small pockets of AT&T fiber customers in towns of all sizes. The cherry picking strategy makes it impossible to map their fiber footprint since it consists of an apartment complex here and a small cluster of homes there. Interestingly, when AT&T reports these various pockets they end up distorting the FCC’s broadband maps, since those maps count a whole census block as having gigabit fiber speeds if even only one customer can actually get fiber.

Another part of AT&T’s strategy for eliminating traditional DSL is to tear down rural copper and replace DSL with cellular broadband. That effort is being funded to a large extent by the FCC’s CAF II program. The company took $427 million in federal funding to bring broadband to over 1.1 million rural homes and businesses. The CAF II program only requires AT&T and the other telcos to deliver speeds of 10/1 Mbps. Many of these 1.1 million customers had slow DSL with typical speeds in the range of 1 Mbps or even less.

AT&T recently said that they are not pursuing 5G wireless local loops. They’ve looked at the technology that uses 5G wireless links to reach from poles to nearby homes and said that they can’t make a reasonable business case for the technology. They say that it’s just as affordable in their expansion model to build fiber directly to customers. They also know that fiber provides a quality connection but are unsure of the quality of a 5G wireless connection. That announcement takes some of the wind out of the sails for the FCC and legislators who are pressing hard to mandate cheap pole connections for 5G. There are only a few companies that have the capital dollars and footprint to pursue widespread 5G, and if AT&T isn’t pursuing this technology then the whole argument that 5G is the future of residential broadband is suspect.

This is one of the first times that AT&T has clearly described their fiber strategy. Over the last few years I wrote blogs that wondered where AT&T was building fiber, because outside of a few markets where they are competing with companies like Google Fiber it was hard to find any evidence of fiber construction. Instead of large fiber roll-outs across whole markets it turns out that the company has been quietly building a fiber network that adds pockets of fiber customer across their whole footprint. One interesting aspect of this strategy is that those who don’t live close to an AT&T fiber node are not likely to ever get their fiber.