7 Comments

After the intel and apple story for iPhone, the rise of Qualcomm and Snapdragon SoC would be a good read too.

Expand full comment
author

Hi Vikram, that's a great idea. I think there was an acquired podcast on Qualcomm a few months ago but that was more on the business side. I'll add it to the schedule!

Expand full comment

I really enjoyed this read, it give a complete picture what actually happend in those fascinating times.

Expand full comment

Such a good read thank you so much (took me like a week to finish)

Expand full comment

I am not convinced that Intel did anything wrong here. They were delivering great shareholder value. They did not really stumble until AMD came out with Ryzen and Epic. I would much rather read about their stumblings in fab creation. That is where they fell down, not in avoiding the worthless ARM market. No one made anymoney there save for Apple. Putting capital there would have been a bad decision.

Intel stumbles now because of AMD's innovations, not Apple's. The mobile phone market loses money for everyone but Apple. Nvidia could not even make money their with their Tegra gaming ARM chip. Intel was wise to avoid that commodity business. Let the Asian firms kill each other for T-bill returns. The real money is in the x86 CPU market, and that has been their primary failing for the last 6-8 years.

The other problem relates to graphics. AMD has perfected the integrated graphics gaming CPU via console sales. That could bankrupt Intel. Intel has survived on server companies that cannot get an Epic allocation and laptop sales. AMD has fixed both. Where intel massively lags behind on graphics will likely determine where they end up. Nvidia clearly does not care about gaming any more (a real shame), but AMD is putting together a real game changer with CPU's that have game-ready graphics built in. Nvidia has so shamelessly abused its gamer base, the men who literally provided 90%+ of their revenue over the decades, this foolish, short-sighted set of policies has given AMD the ultimate wind, offering affordable all-in-one gaming chips just like those in playstations and xboxes. That is what could kill Intel. They have wisely invested in GPU technology, but that is a far more interesting topic than intel making the CORRECT decision not to waste time with low-margin commodity businesses such as mobile phone chips. If anything, they wasted TOO MUCH time and money into the entire field.

Expand full comment
author

Hi Garry, That's a fair challenge but we need to distinguish between the attractiveness of the business of making phones (not great) and designing and building phone SoCs (pretty good). In the latter you've got Qualcomm, TSMC and even now Arm whose market values are all bigger than Intel's. Intel could have had a chunk of that and slowed down TSMC's investment in fabs - which are now being used by AMD to compete with Intel in PCs and in the datacenter.

Expand full comment

I see what you are saying, but Intel made systems on a chip long ago. The chiplet design that Apple and AMD are using is fascinating, and Intel would be is far better shape if it had gambled on this, but I do not see the connection between mobile ARM SOC and the chiplet designs that are dominating today in Intel's actual space. Furthermore, Qualcom and ARM are legitimate comps, but TSMC is not. They pioneered the pure fab industry and have a real niche. I did some work with them when I lived in Taiwan, and that organization is full of very cool guys. That said, I would compare raw financial metrics (cash being thrown off metrics like EBIT/price or EBITDA/price) to compare entities, not a constantly changing, overly emotional share price/enterprise value.

Intel has a good business, and they definitely dropped the ball on manufacturing as well as the importance of GPU's. That said, everyone said AMD were run by morons when they bought ATI and hacked off their foundry business. They eventually found their way with CPu's, but they still struggle with GPU's (but now the CPU sales can subsidize their answer to CUDA and other Nvidia goodies). Intel is far from out. They stumbled, and that happens, but if anything they are a great deal right now. Compute is not going anywhere. I think their careless disposal of their Optane business was a larger own goal than not tying up tons of cash into far lower margin ARM sales. Intel is still making decisions on a cost of capital that is rapidly falling. Note how they sold that "low-margin" business to Hynix, a firm not so different from Qualcom chasing crumbs in commodity RAM and solid state storage silicon. Intel following the MBA playbook is what is getting them into trouble now.

In general, it is really hard to take a business as insanely lucrative as intel in 2014 and steer it right. No one knows what matters. Intel was shamelessly milking their customers for so long and they underinvested, assuming that the competition could not hurt them. Now they are paying for that, but I strongly suspect that there was nothing they could have done to avoid their current situation. They still would have been screwed even if they had a massive, low-margin SOC business. If anything, they would have been better off never messing with mobile in the first place, but how could they know? No one knows which technology will come to dominate.

I would not count intel out, but I cannot honestly Monday morning quarterback them either. They will catch up with AMD, and they will experience what AMD did, plus have a massive loss-making GPU unit that they cannot afford to close. I cannot in good conscience say that they truly made any mistakes that anyone other than an omniscient god could have avoided.

This is why capitalism works so much better than state-run development at the productivity frontier. No one knows what will work best, so the system with lot's of unrest and turnover and competition will out-innovate. Some big firms must die so the capital can be re-directed to more useful endeavors. Developing economies can exploit a sophisticated government to catch up, but once you are at that edge of innovation, no one knows who will come up with the best ideas. No one knew that chiplets would be so good. No one knew TSMC would be so much better than everyone else. We all wish we bought AMD at $1.50, but few of us did.

That said, great work. I really enjoyed reading this.

Expand full comment