Setting a New Standard for Unpaired Spectrum Pricing: AT&T’s Purchase of Qualcomm’s 700 MHz Spectrum

By | December 28, 2011

On December 22, the FCC approved of AT&T’s purchase of the unpaired D and E blocks from Qualcomm for a total of $1.925 Billion. What went missing in most media reports that the D and E bands are unpaired 6 MHz bands in what’s commonly known as the “lower 700 MHz band.”

US 700 MHz Spectrum Band Plan

US 700 MHz Spectrum Band Plan

The D band (716-722 MHz) is nationwide, while the E Band (722-728 MHz) covers five Economic Areas (EA) with a population of over 70 million: New York City, Los Angeles, San Francisco, Philadelphia, and Boston.  This puts the price of the licenses at $0.85 per MHz-PoP. By comparison, Qualcomm paid just over $38 million for the D band in Auction 49 (June 2003) and $554.6 million for the E-band licenses in Auction 73 (March 2008) for a total $592.6 million.

So, in short, AT&T paid 3.25x the price Qualcomm paid to acquire these bands. Is it worth it?

The short answer in my opinion is “No.” I will base this on two factors: 1- the use case of these bands, and 2- what others have paid for similar spectrum.

AT&T says they will use the D and E blocks for downlink carrier aggregation which would pair with their AWS spectrum (although it is possible to pair with other bands they hold). Carrier aggregation across bands requires more complex RF front end and in this case, it means even more specific equipment to AT&T’s network. So a handset would have to support the duplex AWS band plus the downlink 700 MHz D and E bands, in addition to other bands that would be required for roaming. We have already seen reports of how the bifurcated frequency band for LTE is slowing down handset availability and deployments. The second issue for carrier aggregation across band is additional network planning complexity. 700 MHz and 2100 MHz have different propagation characteristics, especially when it comes to in-building penetration. This means different coverage area for each of the two channels which is bound to add complexity in terms of planning. Should they place the cells to close, the 700 MHz carrier will experience higher interference, but should they separate them to accommodate the 700 MHz signal quality requirements, coverage gaps on 2100 MHz will open up.

Of course, they can pair these bands with their 700 MHz B band holdings. This makes things easier especially when it comes to network design but still means dedicated equipment. Carrier aggregation is not going to happen in the near term – this is more an issue of economics rather than technology.

On pricing, Verizon paid a $4.74 billion for the paired C band in the upper 700 MHz band in 2008. This is about $0.7 per MHz-PoP for 2×11 MHz paired spectrum. Overseas, recent spectrum auctions in Western Europe have prices paired spectrum in 800 MHz between €0.5 and €0.8 per MHz-PoP for paired spectrum. It seems that unpaired spectrum at $0.85/MHz-PoP is excessive, especially as to use it requires customization on equipment (both handsets and base stations), or another technology altogether (e.g. TD-LTE). Of course, it goes without saying that Qualcomm’s acquisition price is 69% lower than what AT&T paid!

So, for the reasons above, I think that $1.925 billion is a high price to pay for a couple of 6 MHz unpaired channels. Drop me a note a let me know what you think – look forward to hearing the different opinions!