* A cluster of smaller GPUs, which will be comprised of modular designed units and connected through an interconnect such as HyperTransport or a FSB [1][3], will replace the current single large GPU design used in current high-end graphics cards [1][3]. Therefore, the R700 will be more modular and more scalable than previous GPUs. [1]
* The R700 may have much better energy efficiency, maybe derived from Phenom processors' split power-planes[4], and may also enable power saving in idle mode. [citation needed]
* According to Phil Hester and Dave Orton, the Radeon R700 will be targeted to launch on a 55 nm fabrication process with 6-month product cycle for enthusiast and performance segments, with small updates (presumably a die shrink, minor architectural changes, and improvements to performance) to the architecture every 6 months after original architectural release. For mainstream and value segments, the product cycle will instead be 12 months without architectural alterations. The planned successor, presumably codenamed the Radeon R800 aimed at a 2008 launch, will be targeted to launch on a 45 nm fabrication process [5], with the expectation that the Radeon R800 will be compatible with the next major version of DirectX API which is aimed at a 2009 launch.