According to UK-based consultancy Coda, US mobile ad revenues will reach $4.2 billion by 2015. From the top-level data released for PR purposes:

• SMS will make up just over half of all mobile ad revenues in 2010, these will fall to around 3% in 2015.
• Revenues from search will rise to 70% of all mobile ad revenues in 2015, which is up from a third of mobile ad revenues in 2010.Of this 70%, a third will be derived from local search.
• Among other findings is that users accessing video via mobile phones will total 74m in 2015, generating $530m in subscriptions, and $420m in advertising.

The notion that 70% (or 73% per Kelsey) of all mobile ad revenues will be search is simply incorrect. Most of the forecasts coming out are derivative of one another and employ basic or too-high-level thinking about user behavior, advertiser positioning and revenue models. There does appear to be a good deal of work that went into the Coda forecast; however I disagree with its conclusions. 

Display (broadly construed) will be a big part of the mobile ad mix and SMS will continue to be used in a significant way, although perhaps in a "marketing" context more than as "advertising" per se. Search, to be sure, will be significant and a meanginful source of revenues. But unless you mean "search" broadly as "directional intent" lookup, it's not going to have that kind of share of mobile ad revenue. Search is "diversifying" in many ways that I've discussed before. And much of "local search" advertising will be a directional/contextual hybrid in the future. 

Here are some screens reflecting the Coda revenue projections:

 Source: Coda/Fierce Wireless