A (0:02)

In 1991, three once fierce enemies, Apple, IBM and Motorola formed an alliance. At the press conference, IBM President Jack Kohler proclaimed that their partnership would launch a renaissance in technological innovation. Apple CEO John Scully added that it would be the bridge over the great divide to a new object frontier. What. Whatever that means. The most visible product of this three sum was PowerPC, a super fast chip architecture made to take on Intel's solidifying monopoly. In today's video, a story about hitting monopolies head on the Original Apple Silicon PowerPC this video is brought to you by the Asianometry Patreon. Apple had been using Motorola architecture silicon for many years, a choice that dates back to Apple co founder Steve Wozniak's decision to use the cheaper MOS technology 6502, a close relative of the Motorola 6800 for the Apple 1 and 2 microcomputers. Then when Apple brought out its Lisa and Macintosh computers, they used Motorola's follow up to that chip, the 68000. But after intel stormed out with the 386 CPU in 1985 and Compaq picked it up for the DeskPro 386, the whole PC clone industry accelerated. Apple's Macintosh failed to grow or even hold its market share. Despite the share losses, the company modestly grew revenues and profits in the second half of the 1980s thanks to a strong position in schools and high end desktop publishing. But the PC clones were coming for that too. The 386 helped enable a new wave of PC superclones with enough compute power and disk storage to challenge what had been defined as high end just a few years before. Apple's 54.7% gross margins in 1990 were a direct result of its high end Macintosh workstations and their 60% margins. With the superclones rising fast, Apple needed to clamber up the compute ladder. But the Motorola 68000 family was aging when that first 68K came out in 1979. It was this magnificent, powerful beast that beat the pants off everything else then available in the market. But semiconductors are a volume game, and the millions of PC chips sold each year gave intel economies of scale. PC profits funded R and D and capital investment for new factories and faster chips. A vicious cycle that Motorola struggled to keep up with. For Motorola, Apple was more than just a big semiconductor customer. They held strategic importance. Most of Motorola's chips were for the telecommunications and automobile sectors smaller things with less processing power and lower average selling prices. Apple on the other hand paid big bucks and high margins for its powerful processors. So Motorola rapidly released 268000 successors to keep up the 6830 and the 6840. But they added so many new features so fast that it complicated the designs, leading to thermal issues, fabrication problems and delays, thus leaving Apple with an aging and underpowered workstation lineup during a critical time. Their lineup seemed especially lacking when it came to the hot thing at the time, the RISC. RISC emerged in the late 1970s. Roughly very roughly speaking, it was a simplification of a processor's capabilities, allowing them to process data way faster, starting with IBM's Project 801 and then the Berkeley RISC project and Stanford's MIPS project. RISC's chips took the microprocessing world by storm in the 1980s. In 1986, Sun Microsystem, which had previously used the 68000 for their workstation computers, announced the SPARC architecture, demonstrating a staggering 10 million instructions per second. Risk seemed real, and Apple spun up two internal teams to explore it. They reviewed options from a startup, spun off the aforementioned MIPS project, IBM, sun, and even amd. But in the end, Apple felt that switching both architecture and chip supplier simultaneously would be too much. So they eventually chose to adopt Motorola's next generation risc architecture, the 88000. Apple engineers began developing an 88110 powered Macintosh with backwards compatibility. But even as that went along, high level Apple executives felt increasingly worse about their chances of keeping up with Intel. In 1990, Apple, Motorola, and a few other random customers would have had about 12% total market share of the personal computing market, not enough to fund a credible leading edge challenger to Intel. It didn't help that Motorola struggled to find another significant customer for the 88,000. And even if Motorola did succeed in bringing out a leading edge node, they will quickly have to do it again to keep up with Moore's Law, doubling Intel's transistor counts every two or three years. The cold hard truth was that Apple needed more teeth. Apple needed a partner in their corner with very advanced semiconductor manufacturing processes. So that is why in 1990 Apple started quietly reaching out to IBM. IBM as a semiconductor player did not have the highest yielding fabs, but they did have the best chip technology, two examples of which included the early implementation of copper interconnects, which helped chips get faster as they shrank, and chemical mechanical polishing, or cmp, which helped improve yields. Yes, IBM created and still sells PCs. And yes, Apple tweaked IBM again and again in its adverts. But that was the past. The real enemies now were Microsoft and Intel. In banding together, Apple and IBM can hopefully go further than if they tried to go it alone. So in July 1991, the two companies signed a letter of intent to share technology and collaborate on four overarching initiatives. One was a joint venture called Kaleida, which sought to make multimedia software and other stuff. Another software joint venture was Taligent, which explored the commercialization of an Apple internal operating system with object oriented technology called Pink. There was yet another pact to build a Unix operating system called Power Open Environment. POE melded together Macintosh OS's user interface with AIX, IBM's Unix variant. But the most significant cooperation was for the development of a shared RISC architecture based on IBM's own 32 bit RISC architecture power. IBM previously developed it for the IBM Rs6000, a rather successful workstation. Apple reviewed it before, but passed because it had been IBM internal and and the Power chips wouldn't fit in a Macintosh socket anyway. But here IBM was willing to share and work with Apple to evolve Power into this new thing called PowerPC. Moreover, IBM let Apple bring on Motorola as a second source, a process that included the transfer of their semiconductor, Copper, Interconnect and CMP technologies over to Motorola. Together they created the AIM alliance. The so called deal of the century shocked the industry. Some of the most shocked were the company's own employees. Many at Apple still liked what they were doing with Motorola and the 88,000 architecture. Now they had to drop all that and start on this new IBM derived thing. Media outlets were mostly positive, albeit slightly skeptical. Several analysts said the practical thing, which was that these are two massive companies quite used to getting their way. How can they tolerate working with each other? Others were more colorful. A joke famously circulating in the valley went, what do you get when you cross Apple and IBM IBM? Robert W. Stearns, VP of corporate development at Compaq, succinctly condensed his feelings in saying PowerPC supporters are smoking dope. There's no way it's going to work. When the AIM alliance was established, IBM already had a Power chip design in progress, the IBM 601. And this was a nice head start. But rejiggering was still needed to ensure that the chip can fit Apple's existing setups, which were then half built for the 88k architecture in mind. The result was an interesting hybrid IBM Motorola design. Things had changed so abruptly the that a proper Apple style development environment didn't yet exist. Gary Davidian recalls in his oral history for the Computer History Museum writing his own PowerPC assembler for the Macintosh because he didn't want to develop on IBM RS 6000 workstations. And of course, the companies dealt with cultural clashes, though generally the three meshed well. IBM made good hardware, but wasn't used to dealing with external customers. Motorola's customer skills were far better and they helped bridge Apple and IBM. When the first test PowerPC 601S arrived from IBM in September 1992, it was noted that a bug in IBM's test scan interface meant that the fab couldn't actually test the chips. So Apple engineers pulled all nighters to put the prototype into a machine just to see if the chips actually worked. A month later, the team demoed the first Power Macintosh at an internal sales meeting in Hawaii. There was a harrowing few hours when the demo computer just died for no reason, but otherwise it went beautifully. In fact, the whole execution of this first PowerPC chip went beautifully. Despite the abrupt change in direction and short 12 month timeline, it is remarkable that the alliance still managed to produce a chip then faster than its x86 counterpart. I would like to point out that this was the early 1990s, an interesting times for the PC industry. In 1992, the PC ecosystem exhausted its early easy growth and descended into a ferocious price war led by Compaq and its new CEO Eckhart Pfeiffer. Apple's gross margins fell from 55% in 1990 to just 24%. The financial fallout from that plus the Apple Newton's costly flop cost John Scully his job, though creating Newton did help create arm. So there's that. IBM also suffered immense pain. An American economic recession, corporate restructuring and the market share losses from the PC price war caused profits to plunge. In 1993, new CEO Lou Gerstner cut 45,000 jobs and closed dozens of factories to move away from selling products and towards services, culminating in two of the worst corporate annual losses ever reported up until then. All the while, the Wintel Death machine was in full force. In 1991, intel debuted the legendary Intel Inside marketing campaign that made their brand one of the world's Most valuable. Their 75 to 80% share of the X86 market was grinding AMD and Cyrix into dust. And in 1992, Microsoft finally cracks the gooey puzzle with the incredible new Windows 3.1. It sells millions of copies and its power and user friendliness represent a turning point in Windows adoption. Macintosh market share starts slipping again. Then in 1993, Moore's law ticks again with Intel's successor to the 486, the Pentium. And despite harboring a very infamous bug at 66 MHz and 100 million instructions per second, it closed the PowerPC 601's advantages in speed. But Apple wanted people to know that it was not afraid to build up hype for the forthcoming PowerPC Macintosh launch and maintain their teetering market share. Apple ran an aggressive marketing campaign in late 1993. At the massive Comdex trade show in November 1993, Apple put the yet unreleased PowerPC chip front and center. PowerPC shirts, PowerPC buttons and a PowerPC pavilion for hands on demos and comparisons. The Message was clear. PowerPC was backed by three of America's biggest and most successful companies. A billion dollars had been invested into it. It's a big deal. And above all, its groundbreaking RISC architecture promised incredible speeds with both native apps and and Windows apps. Via simulation. They called out the Pentium for running too hot, using too much energy and costing too much money. Then Apple CEO Michael Spindler predicted that PowerPC would be the mainstay of the desktop. No, Bo Me, then president of IBM's PC division, said, With a PowerPC we can offer Windows users at least the performance that they can get running Windows with the Pentium and a wider range of functionality with native mode applications. It was a Pentium versus PowerPC show. Tom Beaver, VP and Director of PowerPC Programs for Motorola and owner of the greatest surname ever, chuffed, the others are going to be niche players. I think it's going to be a two chip race. Intel responded with typical intel arrogance. Andy Grove said, presumably while sitting on a throne made from the tibias of his vanquished foes, there has not been a day in Intel's life, certainly not in the microprocessor area, where we haven't had major forces trying to take a chunk out of our existence. The PowerPC crowd won't do it. They set up their own Pentium banners and memorabilia and took out ads in Macintosh focused magazines like MacWeek, telling people that a PC can do anything that a Mac can. Paul Odolini, Intel's general manager and future CEO, seemed nonplussed. He told the Guardian a little later, when we come out of this, Apple will be left with only 5 to 10%. If this is a religious war, we've already won. It was amidst this tension that Apple launched the first three power Macintoshes on March 14, 1994. The anticipation and stakes were sky high. The 10 year old Macintosh line now desperately needed a refresh. At the high end, the Power Macintoshes were expected to take 20 to 25% of Mac shipments right off the bat. PowerPC would then gradually take over the rest of the product line. But will people buy it? The cost was significantly more $2,000 to $4,500 while their Pentium powered counterparts range between 2000 to $3,000. And what do you get for that higher price? Yes indeed, powermax can run programs two to four times faster than but only if those packages source codes were recompiled for the RISC architecture. Apple evangelists were deployed as early as 1993 to get developers to recompile. Luckily, the PowerPC Macintosh sold well in the first two weeks after launch. Apple shipped 145,000 units by January 1995, a million. Despite backlogs due to under forecasted demand and component shortages, Power Macintosh sales remained a lifeline during the hard times to come. A reason why the AIM alliance felt like such a big deal at the start was the prospect of Apple, IBM and Motorola establishing shared standards on chips, hardware and software. Apple did indeed get and use the PowerPC chip, but little else. Neither software joint venture made any credible headway. Kolida never shipped anything and IBM exited Taligent in January 1994. The two companies teams also clashed on operating system cooperation. Increasingly frustrated with Apple's tight lipped behavior, IBM considered buying the whole company in late 1994. Such a merger would have created a computer maker far larger than current leader Compaq. But the deal fell through overprice and antitrust concerns, plus other conditions like Michael Spindler demanding a big golden parachute and to report directly to IBM CEO Lou Gerstner. In 1995, Apple's market share lost another percentage point, falling to 8.3%. Spindler spent much of the year trying to cut costs faster than revenues can decline and speed dating other merger partners like sun or Hewlett Packard. Meanwhile, Microsoft brings out the Blockbuster operating system Windows 95, a rare time in tech history when people threw launch parties for a Microsoft product. Apple, on the other hand, struggled to finish their next generation operating system Copland. Spindler was fired in early 1996 after turning a 70 million quarterly loss. He was replaced by Gil Amelio, which with Kaplan acknowledged as a failure. Apple decided in late December 1996 to buy Steve Jobs somewhat failed workstation startup Next Computer and bring him back as an advisor. Little over a year later, the board pushes Amelio out for Steve Jobs. One of the first things that Jobs did was kill the clones. In early 1994, Apple started licensing its technology in an attempt to raise its market share to 20%. A rogues gallery of now forgotten names like Radius Inc. Daystar, Digital, Pioneer, Electric, Toshiba and Olivetti signed on. Some of these got pretty large. Apple's first official licensee in December 1994 was Power Computing Corporation, a startup founded and run by the Korean born American Steve Kang. Power came out of nowhere to rapidly sell 100,000 units, generating between 250 and $300 million of revenue. Jobs recognized that these clone makers not only undercut Apple's own products, but threatened to pry the ecosystem away from them entirely. So he raised a license fee. When rebuffed, he canceled agreements. And with regards to Power Computing, Apple bought the assets of their Macintosh division for $100 million in stock worth about $200 billion today. Motorola also made and marketed their own small line of Macintosh clones called Starmacs. They were actually the second largest Mac clone maker after Power Computing, and for them, clones were strategically important in expanding the PowerPC architecture's market share enough to make its chip production economically viable. Phil Pompa, director of marketing for Motorola's RISC microprocessor division, told BusinessWeek in mid-1996, we'd need 20% market share of the computer business just to get to a stable position. So Jobs crackdown did not please them. The Wall Street Journal reported an explosive phone call between Jobs and Motorola CEO Christopher Galvin about the issue. Frustrated, Motorola canceled their clone line in September 1997, taking a $95 million charge. The stock kerplunked 11%. Joe Gulielmi, corporate vice president and general manager of Motorola Computer Group, said, Apple has made its decision. Now we have made ours. We will be phasing out our investments. He added in another interview, this is the end of a long and arduous negotiation with Apple. We are disappointed. The practical effect is that Apple is going to go alone. They will have to provide all the new innovation themselves. The former special relationship between the two indeed seemed quite damaged. An anonymous Motorola executive told Wall Street Journal, there will be just another customer. The late Katie Cotton, Apple's legendary spokeswoman, replied, we respect their decision. Apple and Motorola have a strong business relationship. She then mentioned that Apple buys $300 million of chips from Motorola each year. Sassy snark aside, the issue was that $300 million a year was not enough for Motorola's Semiconductor Products sector division. Motorola SBS revenues declined from $8.5 billion to just 8 billion between 1995 and 1997, presumably because the clones but operating profit crashed from 1.2 billion to just 332 million. Outside of Apple, the PowerPC had little significant desktop or laptop business. Despite attempts to woo canon, intel and x86 had pretty much sewn up the market, either through competence or monopoly arm twisting. The parent company was also struggling. Motorola's bread and butter were in telecommunications, but their choice to back the CDMA cellular standard over GSM hurt them in the United States and abroad. How long can they afford to keep investing in a sector unrelated to their core business? As early as 1997, analysts like Dataquest's Nathan Brookwood predicted that Apple would eventually move to Intel. Brookwood pointed out that an Apple operating system experiment called Rhapsody had been built to run on both PowerPC and x86 processors. The technology capacity had always been there. But Apple denied such rumors and went on to produce a line of beautiful high performance workstations and laptops using the PowerPC 700 series, more famously known as the G3. The G3 powered Macintoshes were fast, particularly the PowerBook G3 laptop computer. But reviewers noted that it was also expensive compared to Pentium powered versions. And it did not sell enough to justify the rising costs of leading edge semiconductor production. Motorola's semiconductor division spent $1.1 billion on capital expenditure in 1997. Intel spent nearly five times as much. In the wake of the Apple clone fallout, Motorola and IBM turned to a new sector. In October 1997 they announced a collaboration designed PowerPC processors for the embedded sector. Back in the day the term referred to special less powerful chips for things like printers, networking equipment and video game consoles. These chips cost far less than Intel's high end computer chips, anywhere between $13 and 100 as compared to the $200 to $500 that desktop computer chips are priced at. But the competition is less intense. IBM and Motorola can play into existing strengths. IBM then did a good job notching up several PowerPC design wins in the video game console space, which had already been using RISC CPUs for a while. In May 1999, Nintendo announced at the E3 conference that it chose an IBM designed PowerPC CPU for its N64 follow up the GameCube in a $1 billion contract. Then in 2000, IBM struck a three way collaboration with Sony and Toshiba, an alliance that unfortunately shortens to STI to produce a special PowerPC based design called the cell broadband engine. Cell ended up costing between $400 million to $3 billion and four years to complete. Sony used it for their PlayStation 3, a pricey console known for its supercomputerish power and fiendish difficulty to program. Sony intended to use it for all their computing products, but their divisions pushed back on the basis of cost and Power Cell didn't quite live up to expectations. But the PS3 nevertheless sold quite well, cementing PowerPC as a player in embedded devices. Meanwhile, back in the world of Apple, PowerPC was struggling to keep up with the surging microprocessor industry I mentioned in a few prior videos the intense rivalry between intel and AMD. Throughout the 1990s, AMD lingered as a distant second player before suddenly re entering the fray with the Athlon cpu. This competition, as well as an acceleration in semiconductor manufacturing technology advancement, kicked off a race between the two for the fastest clock speeds, presuming the number to stand for raw performance. We remember this as the megahertz wars, and it seemingly left Apple's fastest PowerPC chips in the dust with serious repercussions for its high end Power Macintosh line of workstations. In 1999, Motorola's latest chip, the G4, produced with a 180nm node, had a maximum clock speed of 500 MHz. With intel and AMD shipping chips with clock speeds of 700 MHz, this seemed to consumers underpowered. Of course, the reality is that clock speeds are not directly comparable, especially between RISC and CISC chips and and it does not tell you much about real life performance. Steve Jobs was screaming that from the rooftops in 2001. I find this whole thing a classic example of Goodhart's Law, that is when a measure becomes a target, it ceases to be a good measure. Nevertheless, when Motorola released that 500 MHz chip, Steve Jobs and Apple moved to announce the G4 power Macintosh with it in late August 1999, displaying the 500 MHz number front and center, perhaps feeling the pressure to deliver something to match at a critical time. Unfortunately, they moved too fast. In October 1999, Apple had to admit that Motorola couldn't produce the 500 MHz G4 chip in high enough volume due to a data cache corruption glitch. This forced the G4 Macintosh to take what Netizens called a speed dump. Apple downgraded the three computers in that lineup to 350, 400 and 450 MHz. In response, IBM was brought in as a second source for that G4. IBM had co designed that chip with Motorola in Austin, but at first declined to fab it. Rumor had it was because the IBM team did not like Motorola's addition of a vector processing unit called Altavec to the design. But at Apple's behest, IBM stepped in and in February 2000 Apple brought back the 500 MHz G4 Macintosh again. The 1999 G4 speed dump fiasco was partly Apple's fault. They shouldn't have put out a computer that fast. But it made clear the major problem Motorola was losing its edge in the semiconductor business and in fact behind the scenes they had been thinking about exiting the since at least 1997 when the Mac clones were extinguished. Their management wanted to focus on the increasingly few things they believed their company was good at and they did not believe it can afford the semiconductor industry cyclical volatility and deep capital requirements. Former division president Hector Ruiz recalled in an oral history for the Computer History Museum. I think it became pretty clear that Motorola couldn't support two major thrusts at the size the company was. It became clear that a divestiture of the semiconductor business had to be considered in 1999, when Motorola sold the semiconductor components division to the private equity group TPG for $1.6 billion. They refused to reinvest any of the proceeds into the division, instead keeping it for themselves. Then in 2000, the end of the dot com and telecom bubbles in inflicted sharp declines onto Motorola's telecom businesses. The company turned $6 billion in losses in 2001 and 2002, laying off thousands of jobs and closing factories all over. In 2003 they announced that they will spin off the entire semiconductor division as its own company, Freescale Semiconductors. Freescale would later be acquired by NXP in 2015. The separation was essentially Motorola's exit from the PC desktop industry. This new smaller spinoff entity can keep producing legacy G3 and G4 microprocessors but did not have the resources to invest like intel or even AMD for the next generation. IBM had to step up. The problem with that was that Apple was a difficult customer. A 2009 CNET article by Brook Crothers said cited a former IBM executive pointing out that Apple's demands for high performance meant that IBM had to invest a great deal of money not just for new semiconductor equipment, but also supporting technologies like chipsets and compilers. Apple also demanded a lot of extra stuff that had to be customized for them, pulling worker resources. Yet at the same time Apple had so little market share and and demanded split volumes with second sourcers whenever possible. And IBM could not achieve economies of scale, so they had to charge Apple a premium, which made Apple very unhappy. In 2002, IBM delivered the PowerPC 970, a new 2 GHz 64bit processor derived from the company's Power4 chips inside their Regatta servers. Apple branded it as the G5 and launched the power Macintosh G5 in June 2003. Steve Jobs gushed at the time, the PowerPC G5 changes all the rules. This 64 bit race car is the heart of our new power Mac G5, now the world's fastest desktop computer. Jobs kept on gushing. During his June 2003 WWDC keynote, he promised that this PowerPC architecture has had so much more room to run and that by June 2004 Apple's Power Max would hit 3 GHz clock speeds. Now, before I move on, I highly recommend this helpful post by the blog 512 pixels, written by Stephen Hackett. It discusses this sordid piece of history. So go check it out. The post details Jobs promises and double down promises on the pledge after first saying it at the June keynote. He doubles down on it again in Paris a few months later in September, though hedging that it would take until the end of next summer. Turns out that never happened and Apple had to retrack the promise. The next year, Apple released a G5 computer lineup that topped out at 2.5 GHz. Why? Jobs got up at that year's WWDC and said, now I want to talk about 2.5 GHz because I stood up here a year ago and said we'd have 3 gigahertz within a year. What happened? The G5, as you know, is a very complex chip and in the semiconductor industry, to make things run faster, they traditionally shrank the geometries. The PowerPC was being made in 130 nanometer geometries and in the last year the the semiconductor industry has gone from 130nm to 90nm, expecting everything would just get faster. No problem. It hit the wall. The whole industry hit the wall at 90nm and it's been a lot harder than people thought. I've covered before the issues relating to the end of Dennard scaling, but let me try to add some context to Steve's remarks. When transistors get smaller and they require less voltage to switch and can also switch faster because the channels between the source and drain get shorter. So you can put More transistors onto the same size chip still use the same amount of power and get faster performance. This is denard scaling and people thought it would last forever. But the transistor eventually shrank so much that the oxide layer between the transistor's gate and channel got to just nanometers thickness. When that happens, the gate no longer works the way it should. A current is left running even while the gate is closed, causing power consumption and heat generation to surge. These chips became hotplates. When IBM and Intel hit Steve's aforementioned wall at the 90 nanometer node, there was no choice but to stop trying so hard to get single core clock speeds to 3 GHz and beyond and instead parallelize. Ergo, Apple put two G5 chips into the rig and intel and AMD went multicore. Now, at the 2004 WWDC keynote, jobs said that IBM was doing really well relative to the rest of the industry. But behind the scenes, Apple was seeing things that it did not like. IBM's nanometer node struggled with yields, perhaps due to to the integration of exotic technologies like strained silicon to improve the gate's control over the channel. This caused substantial delays for the iMac. G5 and then two G5s together in the same rig force Apple to do wild things like include a liquid cooling system. First major commercial PC was something like that, but the bigger problem was mobile. Not mobile as in cell phones, but laptops. IBM's Power 4, as I mentioned, was originally made for servers. Big heavy things with little concerns about power and heat. But for Apple, sales of laptops like the PowerBook and iBook were becoming increasingly important. In 2003, 42% of all Macintosh sales were portable laptops, up from 2001 where it was just 31%. If the desktop G5 needed liquid cooling like a freaking cray, how can you put that into a laptop with its fewer fans and vents? During a 2005 earnings call, Tim Cook, then EVP of Worldwide Sales and Operations, dubbed it the mother of all thermal challenges. And with the Apple chip business minting more prestige than profit, IBM was not really interested in changing their long term roadmap to fill that gap. They had recently won a significant deal with Microsoft to provide PowerPC technology to create the combined CPU GPU unit for the second generation Xbox gaming console, the Xbox 360. The aforementioned former IBM executive said they had hoped Apple would adopt the cell architecture. I suspect this strategic Disinterest is why PowerPC didn't get used for Apple's mobile devices. Like iPods. ARM's architecture also just worked better. Intel on the other hand, was aligned with Apple. After playing the megahertz clock speed game, they too hit the same thermal and power problems for PC laptops. So their roadmap did indeed address the performance per watt issues that Apple cared about. Analyst Charles Wolf at Needham company put it pretty succinctly for the New York that's a huge looming problem for Apple if it can't keep up with intel notebooks in performance. And that's been an IBM problem. IBM hasn't delivered a cool running G5. And so it goes. In 2005 Apple announced that momentous processor transition to intel, the venerable PowerPC chip architecture left behind for video game consoles and IBM's then still sizable mainframe business. PowerPC is still around though its place and embedded is challenged by Army In 2013, IBM started the Open Power Consortium that granted IBM's Power ICA patents license free. They later merged with the Linux Foundation. In the end, the rise and fall of PowerPC as a desktop challenger to intel had little to do with technology. For what it is worth, it seems like the technology mostly kept up with whatever intel had. Whatever problems that PowerPC suffered, such as thermal issues intel had too. The issue was more with Motorola and IBM as Silicon Fab technology partners. And you can't blame them because Apple's business alone was then not big enough to support the sizable investments needed to accommodate their demanding requests. Back then Apple was just a niche operator with a historic name, charismatic salesman, CEO and a minor music player business. Fortunately, Apple found Intel and the two worked together on everything and became great partners and lived happily ever after. The end. Alright everyone, that's it for tonight. Thanks for watching. Subscribe to the Channel, sign up for the Patreon and I'll see you guys next time.