RUMOR: IBM way ahead of schedule; G5’s in all Mac models within 12 months

“According to very reliable sources, IBM is planning to push it’s PowerPC production process directly to 0.06um by the end of the year. Actually, it looks like the first tests of production of PowerPC 970 processors at 0.06um are very encouraging and that IBM is thinking about skipping a step on the roadmap of it’s PowerPC processors,” is reporting.

“The fact is that a 0.06um based PowerPC 970 will have a lot of advantages: first of all, it will be noticably cooler than the current PowerPC 970, secondly it will be able to reach even higher clock speeds, and finally, it will be cheaper to produce,” reports. “…the fact IBM wants to jump to 0.06um process for the production of it’s future PowerPC processors is very good news for Apple, as it certainly means the company will be technically and financially able to go 100% G5 for all its Macs, including laptops and consumer Macs within a year.”

Full article here]here]]here[/url][/url]


  1. Faster? cooler? AND cheaper? Nah! Too good to be true. Never EVER trust rumour sites, no matter how “reliable”.

    Looking forward to eating my words though…


  2. I saw this yesterday [at ][url=]]][/url] and noticed no one else was reporting on it. Seems like pretty incredible news. Can anyone [besides David Richards of SmartHouse fame — re: ][url=]]][/url] confirm the source of this tidbit [you’ll notice I was nice and refrained from calling it a baseless rumor]?

  3. When this turns out not to be true, watch all the whining and bitching about how IBM can’t make the grade! Some people think making chips is like making cookies. IBM is barely getting out enough G5 processors at the moment to meet Apple’s demand. Whoa Nelly? No way.

  4. There is a long discussion of this over at — over on their “2nd page” as they consider it a pretty wild and unsubstantiated rumor. The general consensus over there seems to be that it is theoretically possible, but extremely unlikely. A jump over an entire process generation is almost NEVER done — especially not with your bread and butter producst like the 970 and Apple’s custom chips which go with it.

  5. This is the kind of success that IBM had in mind with their new fab. Despite stories of lack of chip orders and startup issues effecting profitability IBM now has the tools necessary to advance in giant steps rather than the normal process advances typified by the standard fab processing.
    Remember, the 980 is due by next summer. The current G5 boxes will have another big jump in performance in a year as well. Now, if the quantity issues are answered we’ll all be happy!

  6. Even if IBM were moving to 0.06 this year with the PPC 970, at that size there’s no guarantee that we would see great improvements in power…heat dissipation. Already at 0.09 manufacturers are finding that the leakage is proving to be a tremendous challenge that is slowing the progress forward to ever-fast and lower-power CPU’s. Prescott is rumored to draw over 103 Watts at 3.2 or 3.4 GHz. This does not represent a huge leap forward by 0.09.

  7. has lots of discussion of this, and people seem to think that the more you shrink the chip, the more it has to be re-designed.

    It sounds like it’s NOT a simple matter to just “skip” 90 nm.

  8. Prescott, like all its pentium brethren, is a power pig. The G5 already uses a lot less power than Prescott. If a die shrink doesn’t give you lower power dissapation, there goes one more reason to do it.

  9. I work in the semiconductor industry, and I can tell you no one is moving to 60nm yet. Not even IBM. Everyone just moved high-end stuff to 130nm. The next step is 90nm, and it is not an easy or cheap one. You might see 60nm in 2005 or 2006, but not this year. Sorry.

Reader Feedback

This site uses Akismet to reduce spam. Learn how your comment data is processed.