Since its GeForce 6 graphics card series in 2004, Nvidia has managed to execute almost flawlessly against the competition. However, The Inquirer says the company may have a rougher time with its next-generation GT200-based cards.
According to The Inq's report, the GT200 graphics processor will have a 576mm² die size, or almost 1.8 times bigger than the G92 in Nvidia's current speediest single-GPU products. Unsurprisingly, such a massively increased die area will lead to considerable thermal envelopes: 236W for the top-of-the-line GeForce GTX 280 and 182W for the cut-down GeForce GTX 260.
The report adds that Nvidia will price the GTX 280 at $649 and the GTX 260 at $449. If there's any truth to rumors we've heard about next-gen AMD cards, Radeon HD 4000 offerings with single GPUs should all be priced under $350, and they should be competitive with current Nvidia offerings. That means Nvidia may have to rely on aggressive G92 pricing rather than its next architecture to compete.
To curb the damage, the report says, Nvidia has already taped out a die-shrunk GT200 based on a 55nm process, which should materialize into slightly cheaper GeForce GTX cards in the late summer or early fall. The graphics company might not have substantially cheaper derivatives out until much later, though.
The Inq goes so far as to compare the GT200 to the ill-fated GeForce FX 5800 Ultra, but such comparisons may be a little hasty. Nvidia most likely won't face the same architectural obstacles this time around, and it has already succeeded in producing great high-end cards with a huge, power-hungry GPU (the G80). The site also claims Nvidia will attempt to deceive reviewers by announcing lower-than-actual prices, "[taking] advantage of sites that don't bother checking up on the 'facts' fed to them" since "no reviewer ever updates their pricing or points out that the price performance ratio was just blown out of the water." That's clearly not the case around here, though, and we take exception to this assertion.