 From theCUBE Studios in Palo Alto in Boston, bringing you data-driven insights from theCUBE and ETR. This is Breaking Analysis with Dave Vellante. Much of the discussion around Intel's current challenges is focused on manufacturing issues and its ongoing market share skirmish with AMD. Of course, that's very understandable. But the core issue Intel faces is that it has lost the volume game forever. And in Silicon, volume is king. As such, incoming CEO, Pat Gelsinger, faces some difficult decisions. But on the one hand, he could take some logical steps to shore up the company's execution, maybe outsource a portion of its manufacturing, make some incremental changes that would unquestionably please Wall Street and probably drive shareholder value when combined with the usual stock buybacks and dividends. On the other hand, Gelsinger could make much more dramatic moves. Shedding its vertically integrated heritage in transforming Intel into a leading designer of chips for the emerging multi-trillion dollar markets that are highly fragmented and generally referred to as the edge. We believe Intel has no choice. It must create a deep partnership in our view with a semiconductor manufacturer with aspirations to manufacture on US soil and focus Intel's resources on design. Hello, everyone and welcome to this week's Wikibon Cube Insights powered by ETR. In this breaking analysis, we'll put forth our prognosis for what Intel's future looks like and lay out what we think the company needs to do, not only to maintain its relevance, but to regain the position it once held as perhaps the most revered company in tech. Let's start by looking at some of the fundamental factors that we've been tracking and that have shaped and are shaping Intel and are thinking around Intel today. First, it's really important to point out that new CEO, Gelsinger, he's walking into a really difficult situation. Intel's ascendancy and its dominance it was created by PC volumes and its development of an ecosystem that the company created around the x86 instruction set. In semiconductors, volume is everything. The player with the highest volumes has the lowest manufacturing costs. And the math around learning curves is very clear and it's compelling. It's based on Wright's law named after Theodore Wright, TP Wright. He was an aeronautical engineer and he discovered that for every cumulative doubling of units manufactured, costs are going to fall by a constant percentage. Now in semiconductor wafer manufacturing that cost is roughly around 22% declines. And when you consider the economics of manufacturing and next generation technology, for example, going from 10 nanometers to seven nanometers, this becomes huge because the cost of making seven nanometer tech, for example, is much higher relative to 10 nanometer. But if you can fit more circuits on a chip, your wafer costs can drop by 30% or even more. Now this learning curve benefit is why volume is so important. If the time it takes to double volume is elongated then the learning curve benefit, they get elongated as well and you become less competitive from a cost standpoint and that's exactly what is happening to Intel. You see x86 PC volumes they peaked in 2011 and that marked the beginning of the end of Intel's dominance from manufacturing and cost standpoint. Now ironically, HDD hard disk drive volumes peaked around the same time and you're seeing a similar fundamental shift in that market relative to flash. Now because Intel has a vertically integrated model, its designers are limited by the constraints in the manufacturing process. What used to be Intel's ace in the hole is process manufacturing has become a hindrance, frustrating Intel's chip designers and really seeding advantage to a number of competitors including AMD, ARM and NVIDIA. Now during this time, we've seen high profile innovators adapting alternative processors, companies like Apple, which chose its own design based on ARM for the M1. Tesla is a fascinating case study where Intel was really not in the running. AWS probably Intel's largest customer is developing its own chips. You know, through Intel a little bone at the recent reinvent and announced its use of Intel Sabana chips. And practically the same sentence had talked about how it was developing a similar chip that would provide even better price performance. And just last month, it was reported that Microsoft, Microsoft Intel's monopoly partner in the PC era was developing its own ARM based chips for the surface PCs and for its servers. Intel's Zenith was marked by those peak PC volumes that we talked about. Now to stress this point, this chart shows x86 PC volumes over time. That red highlighted area shows the peak years. Now volumes actually grew in 2020 in part due to COVID, which is not really reflected in this chart, but the volume game was lost for Intel when, and has, it's been widely reported that in 2005, Steve Jobs approached Intel as it was replacing IBM microprocessors with Intel processors for the Mac and asked Intel to develop the chip for the iPhone, Intel passed and the die was cast. Now to the earlier point, PC markets are actually quite good if you're Dell. Here's some ETR data that shows Dell's laptop net score. Net score is a measure of spending momentum for 2020 and into 2021. Dell's client business has been very good and profitable. And frankly, it's been a pleasant surprise. You know, PCs, they're doing well. And as you can see in this chart, Dell has momentum. There's approximately 275 million, maybe as high as 300 million PC units shipped worldwide in 2020, you know, up double digits by some estimates. However, arm chip units shipped exceeded 20 billion units last year worldwide. And that's not apples to apples. You know, we're comparing x86 based PCs to arm chips. So this excludes x86 servers, but the wafer volume for arm dwarfs that of x86, probably by a factor of 10 times. Now back to Wright's law, how long is it going to take Intel to double wafer volumes? It's not going to happen. And trust me, Pat Gelsinger understands this dynamic probably better than anyone in the world. It's certainly better than I do. And as you look out to the future, the story for Intel and its vertically integrated approach, it's even tougher. This chart shows Wikibon's 2020 forecast for arm based compared to x86 based PCs. It also includes some other devices, but as you can see, what happens by the end of the decade is arm really starts to eat in to x86. As we've seen with the M1 at Apple, arm is competing in PCs and much better position for these emerging devices that support things like video and virtual reality systems. And we think even we'll start to eat into the enterprise. So again, the volume game is over for Intel, period. They're never going to win it back. Well, you might ask, what about revenue? Intel still dominates in the data center, right? Well, yes, and that is much higher revenue per unit. But we still believe that revenue from arm based systems are going to surpass that of x86 by the end of the decade. Arm compute revenue is shown in the orange area in this chart with x86 in the blue. This means to us that Intel's last moat is going to be its position in the data center. It has to protect that at all costs. Now, the market knows this, that knows something's wrong with Intel and you can see that it's reflected in the valuations of semiconductor companies. This chart compares the trailing 12 month revenue in the market valuations for Intel, NVIDIA, AMD and Qualcomm. And you can see at a trailing 12 month multiple revenue of 3x compared to about 22x for NVIDIA, about 10x for AMD and Qualcomm, Intel is lagging behind in the street's view. And Intel, as you can see here, it's now considered a cheap stock by many. Here's a graph that shows the performance over the past 12 months compared to the NASDAQ, which you can see that major divergence. NASDAQ being powered in part by COVID and all the new tech and the work from home. Stock reacted very well to the appointment of Gelsinger. That's no surprise. The question people are asking is, what's an X for Intel? How will Pat turn the company's fortunes around? How long is it going to take? What moves can he and should he make? How will they be received by the market and internally, very importantly within Intel's culture? These are big chewy questions and people are split on what should be done. I've heard everything from patches, clean up the execution issues. This is very, very workable and not make any major strategic moves. All the way to Intel should do a hybrid outsourced model to Intel should aggressively move out of manufacturing. Let me read some things from Barons and some other media. Intel has fallen behind rivals and the rest of tech. Why is Intel is replacing Bob Swan? Investors are cheering the move. Intel will likely turn to Taiwan semiconductor for chips. Here's who benefits most. So let's take a look at some of the opinions that are inside these articles. So first one I'm going to pull out. Intel has indicated a willingness to try new things and investors expect the company to announce a hybrid manufacturing approach in January. Now, if you take a look at that and you quote CEO Swan, he says, what has changed is that we have much more flexibility in our designs. And with that type of design, we have the ability to move things in and move things out. And that gives us a little more flexibility about what we will make and what we might take from the outside. So let's unpack that a little bit. I mean, Intel, we know is a highly vertically integrated workflow from design to manufacturing production. But to me, the designers are the artists and the flexibility you would think would come from outsourcing manufacturing to give designers more flexibility to take advantage of say seven nanometer or five nanometer process technologies versus having to wait for Intel to catch up. It used to be that Intel's process was the industry's best and it could supercharge a designer to master certain design challenges so that Intel could maintain its edge, but that's no longer the case. Here's a sentiment from an analyst, Daniel Denali. Denali is at city. It says, he's confident, Denali is confident that Intel's decision to outsource more of its production won't result in the company divesting its entire manufacturing segment. And he cited three reasons. One, it would take roughly three years to bring a chip to market. Two, Intel would have to share IP. And three, it would hurt Intel's profit margins. He said it would negatively impact gross margins by 10 points and would cause a 25% decline in EPS. Now, I don't know about this. To that, I would say one, Intel needs to reduce its current cycle time to go from design to production from let's say three to four years where it is today. It's got to get it under at least the two years, maybe even less. Second, I would say was what good is intellectual property if it's not helping you win in the market? And three, I think profitability is nuanced. So here's another take from a UBS analyst. His name is Timothy Arcuri. And he says, quote, we see but no option, but for Intel to aggressively pursue an outsourcing strategy. He wrote that Intel could be 80% outsourced by 2026. And just by going to 50% outsourcing, he said would save the company $4 billion annually in CapEx and 25% it would drop to free cash flow. So look, maybe Gelsinger has to sacrifice some gross margin in EPS for the time being, reduce the cost of goods sold by outsourcing manufacturing, lower his CapEx and fund innovation in design with free cash flow. Here's our take. Pat Gelsinger needs to look in the mirror and ask, what would Andy Grove do? You know, Grove's quote at only the paranoid survive it's famous, less well-known are the words that preceded that quote. Success breeds complacency and complacency breeds failure. Intel in our view is headed on a path to a long drawn out failure if it doesn't act aggressively. It simply can't compete on cost as an integrated manufacturer because it doesn't have the volume. So what will Pat Gelsinger do? You know, we've probably done 30 cube interviews with Pat and I just don't think he's taking the job to make some incremental changes to Intel to get the stock price back up. Why would that excite Pat Gelsinger? Trends, markets, people, society, he's a dot connector and he loves Intel deeply and he's a legend at the company. Here's what we strongly believe. We think Intel has to do a deal with TSM or maybe Samsung, perhaps some kind of joint venture or other innovative structure that both protects its IP and secures its future. You know, both of these manufacturers would love to have a stronger US presence in markets where Intel has many manufacturing facilities. They may even be willing to take a loss to get this started and deeply partner with Intel for some period of time. This would allow Intel to better compete on a cost basis with AMD. AMD would protect its core data center revenue and allow it to fight the fight in PCs with better cost structures. Maybe even gain some share that could account for another $10 billion to the top line. Intel should focus on reducing its cycle times and unleashing its designers to create new solutions. Let a manufacturing partner who has the learning curve advantages enable Intel designers to innovate and extend ecosystems into new markets. Autonomous vehicles, factory floor use cases, military security, distributed cloud, the coming telco explosion with 5G, AI inferencing at the edge, bite the bullet, give up on yesterday's playbook and reinvent Intel for the next 50 years. That's what we'd like to see. And that's what we think Gelsinger will conclude when he channels his mentor. What do you think? Please comment on my LinkedIn post. You can DM me at dvalante or email me at david.valante at siliconangle.com. I publish weekly on wikibon.com and siliconangle.com. These episodes, remember, are also available as podcasts for your listening pleasure. Just search, Breaking Analysis Podcast. Many thanks to my friend and colleague David Floyer who contributed to this episode and has done great work in the last, better part of the last decade and has really thought through some of the cost factors that we talked about today. Also, don't forget to check out ETR.plus for all the survey action. Thanks for watching this episode of Cube Insights powered by ETR. Be well and we'll see you next time.