A persistent rumor states that NVIDIA could offer to run an x86 processor. Rumor has several reasons: NVIDIA launches into the calculation, the company has recovered engineers at Transmeta and the Company's reports with the major processor manufacturers (Intel and AMD) are conflicting. Of course, rumors indicate that the company would not move to compete with Core i7 processors and other Phenom X4 II NVIDIA but rather would build on the entry level.

However, this rumor has a problem: the company has the capacity to produce a powerful processor? Yes and no. Produce a GPU, or rather a chip dedicated to calculations such as Fermi, is one thing. Produce a chip x86 - instruction set notoriously complex - which is competitive is another. Many have tried to compete with Intel and AMD, but only to a lesser extent, Via has managed to compete with Intel, while many manufacturers have broken teeth on the market (Rise & T, IBM, NEC, etc..). The idea that NVIDIA could use the skills of engineers Transmeta is seductive, but there is a but: if Transmeta has disappeared, not without reason. The company had its little success at a time, because it offered processors that consume little but with very low performance. As the gain in consumption was significant, the weak performance does not pose any problems, but now, with the Atom and Nano, this market exists and there is little chance that NVIDIA comes to impose them. In fact, the concept itself of "binary translation" proposed at the time by Transmeta (and some would chip in NVIDIA) is the basis doomed to failure: for all that happens to compete with a processor that uses a native instruction set would require its raw performance is much higher than its target. And be able to beat Intel in this field while in grade priced content, it's almost a mission impossible.

That NVIDIA continues to offer processors based on ARM core (which also is not developed by NVIDIA) is one thing. That the company from competing in a frontal AMD and Intel in the x86 field is another. Is a solution, often suggested: buy Via. A Nano coupled with NVIDIA chip, all engraved with the newer technologies (and not the 65nm Fujitsu) and some optimizations could really get footholds in netbooks and the load is not too important for NVIDIA, unlike to achieve a x86 CPU "from scratch".