By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel generally agrees that while ARM's move into chip manufacturing has potential long-term benefits, the current valuation appears to be pricing in perfection and ignores significant risks, including execution challenges, geopolitical volatility, and potential demand shifts. The panel is bearish on the current valuation of ARM Holdings.
Risk: Execution risk in ARM's pivot to in-house chip manufacturing, including potential delays in fab construction, ASML equipment delivery, and shifts in AI demand.
Opportunity: Potential long-term benefits of ARM's move into chip manufacturing, such as higher gross margins.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
In this episode of Motley Fool Hidden Gems Investing, Motley Fool contributors Tyler Crowe, Matt Frankel, and Jon Quast discuss:
- ARM Holdings’ and Advanced Micro Devices’ blowout earnings.
- ARM’s ambitious new goal to build its own chips.
- The bottlenecks to bringing on new chip capacity.
- DoorDash’s earnings missing guidance.
- Mailbag: Why do Starbucks and Domino’s have negative shareholder equity?
- Mailbag: How will the SaaSpocalypse affect Salesforce and Wix?
To catch full episodes of all The Motley Fool's free podcasts, check out our podcast center. When you're ready to invest, check out this top 10 list of stocks to buy.
Will AI create the world's first trillionaire? Our team just released a report on the one little-known company, called an "Indispensable Monopoly" providing the critical technology Nvidia and Intel both need. Continue »
A full transcript is below.
Ever feel like you missed the boat in buying the most successful stocks? Then you’ll want to hear this.
On rare occasions, our expert team of analysts issues a “Double Down” stock recommendation for companies that they think are about to pop. If you’re worried you’ve already missed your chance to invest, now is the best time to buy before it’s too late. And the numbers speak for themselves:
- Nvidia:if you invested $1,000 when we doubled down in 2009,you’d have $584,450! - Apple:if you invested $1,000 when we doubled down in 2008,you’d have $57,222! - Netflix:if you invested $1,000 when we doubled down in 2004,you’d have $469,293!*
Right now, we’re issuing “Double Down” alerts for three incredible companies, available when you join Stock Advisor, and there may not be another chance like this anytime soon.
*Stock Advisor returns as of May 11, 2026
This podcast was recorded on May 7, 2026.
Tyler Crowe: We've got earnings galore on Motley Fool Hidden Gems Investing. Welcome to the Motley Fool Hidden Gems Investing. I'm your host, Tyler Crowe, and today I'm joined by longtime contributors Jon Quast and Matt Frankel. We're going to do a whole bunch of earnings reactions today because it's been a busy week related to earnings, and, of course, we're going to hit our mailbag at the end of the show.
First, as we're going to start, we're going to talk about basically semiconductor earnings because it has been one of the big talking points of the week. Arm Holdings and Advanced Micro Devices, AMD, both reported within the past couple of days, and after both earnings, we saw shares explode as they blasted past earnings expectations, 15, 20% moves in the day. We're going to start Arm Holdings today because shares are quickly retreating after the company mentioned on its call after hours that mobile growth was, well, not really growth, and that rising costs were going to impact commodity mobile device sales. Jon, Matt, you two played rock-paper-scissors to cover the two. Jon, you happened to pick Arm Holdings as a result. What did you see in the earnings release and the conference call and was today's reaction to this, hey, maybe mobile growth isn't great? Was that an appropriate response, do you think, to what you saw?
Jon Quast: Tyler, I think the market reaction is appropriate, but not for the reason that you mentioned here, and so I just want to frame this. It is important that you mention the mobile aspect of the business because, if we zoom way out, I don't want to take for granted that all of our listeners know what Arm Holdings is. This is a company that really rose in prominence due to mobile devices. Its chips are more energy-efficient than other chips on the market, and that's a really big deal when you're looking at battery life in a mobile device, so it was able to rise. It does not make its own chips. Historically, it licenses these products to the manufacturers of the mobile devices, but if you look at what we have right now in AI, we have a bottleneck.
You've heard about many bottlenecks. The big one is electricity. Power is scarce, and this is driving AI companies to try to find more energy-efficient solutions, and so ARM makes CPUs, and it claims they're two times more efficient than conventional X86 architecture, and that's the kind that Intel makes, for example. Arm is claiming that they can save AI companies 10 billion per gigawatts in capital expenditures in a data center, so that's a really big deal. I think the big news here lately with Arm has been it's not going to just license the technology anymore. It's going to make its own chips. It's going to actually be a chipmaker, and it's a no-brainer. According to the company, it can make 10 times the gross profit per chip than just licensing it. That's a huge thing, and if you look, management says here in the most recent quarter, it already has $2 billion worth of demand over the next two years for its in-house chips. That's a really big adoption curve.
That's really good, but what is the hang-up here? The hang-up here is that, if you look out to Fiscal 2031, which mostly overlaps with calendar 2030, so just four years away from now, it's saying that, look, by then, we'll have 25 billion maybe in trailing 12-month revenue. Maybe we'll have $9 in adjusted earnings per share. You look at where the market cap was before earnings, and it's gone up a lot, mostly due to competitors' earnings results already. It was trading at over $250 billion market cap. Projecting maybe 25 billion in annual revenue in four years. That's over 10 times its four-year forward sales. You look at earnings. It's trading at somewhere in the ballpark of 23 times earnings on an adjusted basis four years out into the future. That's a really pricey valuation for a company that a lot of exciting things are happening, and I do believe that its products are going to be more and more needed for AI data centers, but it just got way out in front of its skates here.
Tyler Crowe: To say that high valuations, that seems to be part of the course for just about anything that's tangentially related to AI infrastructure or semiconductors, whatever. In that vein, we have another relatively highly valued company here with AMD, whose shares jumped as much as 20% yesterday after earnings release. Now, Matt, I didn't get a chance as much to look over the details, but I bet it had to do with AI spend. Prove me wrong.
Matt Frankel: Yeah, and it's not just the 20% gain yesterday. AMD has tripled over the past year, and yes, it has to do with AI spend. That's really the lazy explanation for it, though, so I'm going to go a little bit into depth with that. Revenue, of course, grew significantly faster than analysts thought, and the big driver was, as you say, the 57% growth in that data center segment, which is AI spend, but the guidance was a big part of the reaction to the stock. Second-quarter revenue guidance came in much higher than expected and implied a surprising acceleration in growth. Lisa Su, the AMD CEO, said AMD expects server growth to accelerate and that the company should deliver tens of billions of dollars in just data center AI revenue next year alone.
But really, the X factor here, this is what I meant by that AI spend just doesn't tell the full story, is the strong CPU business that AMD has. That's a big differentiator from Nvidia. AMD is a distant second to Nvidia on the GPU side of the business, which to this point has been generally synonymous with data center chips. AMD is a CPU leader, and this is becoming an increasingly important part of AI compute power, especially in the agentic age that we're approaching. Although the data center segment is the main story here, it's also important to note that the client segment, which includes the chips that AMD puts in PCs and laptops and things like that, that grew rapidly and indicated that the AMD rise in processors continue to take market share from Intel. That just underscores the strength of their CPU business and why the market might be so optimistic on them right now.
There's a lot to look forward to with AMD later this year. They're going to start shipping their Helios full rack system for AI data centers. That's a direct competitor with products Nvidia offers and charges about $3 million apiece for. OpenAI and Meta have already placed large orders. Meta, in particular, is an especially interesting deal because it's literally one of the single largest AI infrastructure deals that has ever been announced so far. There's a lot to like. It's tripled over the past year, but it's for a reason.
Tyler Crowe: I want to expand on what Jon was talking about with ARM getting into building their own chips now, because we're seeing more and more companies wanting to do this. Alphabet said they want to do it. I think Meta has even mentioned it. Tesla has floated the idea of the Terafab. It all sounds ambitious, and I understand why, but one of the things I think about with the semiconductor industry is that, yes, building fabs is nice and new. It definitely increased production, but also there are bottlenecks behind the bottlenecks. You have companies like ASML, Lam Research, as well as KLA Corporation. These companies that we think of, the bottleneck, it's Taiwan Semi or Intel. They're the only game in town in terms of chip manufacturing. ASML is the only game in town when it is the equipment to make the chip factories. I'm very curious when I hear these companies saying, we're going to do this, that they're all going to have to put in orders with these chip manufacturing equipment companies, and I do wonder to Arm's ambitious goals. How long are they going to have to wait in line for this equipment? How long is it going to take to build out?
We've been talking about cost inflation and things like that, and I bring this up specifically because I've been thinking about this a lot lately as much as this is an explosive growth, and we have AI infrastructure, basically, finding any chip that they can find, whether it's reused crypto mining or whatever, it seems like whatever spare parts or compute power they can get their hands on, they're going to use it, but it is still a cyclical industry. As ambitious as all this growth is, how much capacity expansion can we have in chip manufacturing before something really starts to shift because, even if we have this five-year growth period, and we bring all this new capacity online, we could be looking at it 6-7 years from now, and all of a sudden, we're way over capacity. I feel like that's a major risk for, especially, somebody like Arm Holdings who doesn't have this yet and wants to get into it. Are you guys seeing something similar, or is it I think you're just shaking at the wrong problem here?
Matt Frankel: I do see that as a problem. I don't see it as a problem yet, I'll put it that way. Like you said, Arm Holdings is a different animal because they're building this chip business from scratch, essentially. AMD already has enough capacity for what it's doing now. It has somewhat of a backlog, but it's very managed. The big question is, how long can we see this exponential growth go for, and how much are they going to invest in infrastructure and production capacity and things like that before things turn? Right now, supply and demand are clearly not in equilibrium. There's far more demand than there is supply in the chipmaking industry that's why we're seeing companies, like Micron, the memory companies, they literally can't build their products fast enough. Same with Nvidia and AMD. Nvidia used the word "sold out" in its latest earnings report several times to talk about products. For now, yes, there's a backlog on ASML, the equipment that the chipmakers are using to make their products, but right now it's working out in favor of AMD and Nvidia. With Arm, and I'm curious to get Jon's thoughts here, it's a little bit of a different animal. Can they scale quickly enough while the demand is still on the rise, while they still have the ability to turn this into a significant revenue stream? I don't know the answer to that.
Jon Quast: I think that's the key, Matt. If these companies could snap their fingers today and increase the production to meet the current demand, then I think that it would be a higher risk of overcapacity, but because these things do take multiple years and because there are bottlenecks even to them increasing their production capabilities, and you mentioned ASML, I think that's a good point right there, that is going to mitigate some of that risk because they can't increase the capacity as much as they would like to right now. It's multiple years out into the future to bring the supply up, and I guess it really depends on where you fall personally on the growth curve of the ongoing AI revolution. Does the demand continue to increase from here for these products and services? If so, then the supply is still going to tend to lag behind for multiple years, but if demand is plateauing already while supply is ramping, yes, that is the higher risk right there. I'm personally in the camp that I think that the demand for the products are going to continue to rise, at least, with the supply, so I don't see the big risk, as much cyclicality risk, as I've seen in the past.
Tyler Crowe: Just wrapping it up here, I think I'm more or less in line with you, guys, but I reserve the right on some curveball of algorithmic efficiency, where power and compute use goes way down relative to what we're seeing out of Anthropic, OpenAI, and the big power users today. Maybe they'll start seeing some DeepSeek-esque drop in compute power per token or however we want to measure it. I think it's there, but I think we should all be ready for those curves that could happen. We've seen it in numerous other industries before. After the break, Matt and Jon are going to walk me through what I don't understand in DoorDash's earnings.
ADVERTISEMENT: Dell PCs with Intel inside are built for the moments that matter, for the moments you plan and the ones you don't, built for the busy days that turn into all-night study sessions, the moment you're working from a cafe and realize every outlet is taken, the times you're deep in your flow and the absolute last thing you need is an auto update throwing off your momentum. That's why Dell builds tech that adapts to the way you actually work, built with a long-lasting battery so you're not scrambling for the closest outlet and built-in intelligence that makes updates around your schedule, not in the middle of it. They don't build tech for tech's sake. They build it for you. Fine technology built for the way you work at dell.com/dellpcs, built for you.
Tyler Crowe: Like I said, before the break here, I had a really hard time understanding what's going on with DoorDash's earnings a
Four leading AI models discuss this article
"ARM's pivot into hardware manufacturing introduces excessive execution risk and capital intensity that the current 10x forward sales valuation fails to appropriately discount."
The market's reaction to ARM Holdings and AMD highlights a dangerous decoupling of valuation from operational reality. While the AI infrastructure thesis is robust, ARM’s pivot to in-house chip manufacturing is a massive strategic gamble, not a 'no-brainer.' It introduces significant execution risk, capital expenditure bloat, and competitive friction with its own licensing partners. AMD’s growth is impressive, but the 'agentic age' narrative assumes a linear adoption curve that ignores the massive, looming bottleneck of power availability and cooling infrastructure. Investors are currently pricing in perfection across the entire semiconductor stack, ignoring that even minor delays in ASML equipment delivery or a shift in model efficiency could trigger a violent multiple contraction.
If the transition to custom silicon allows hyperscalers to achieve the 10x gross profit margins ARM projects, the current premium is actually a bargain compared to the long-term cash flow potential of a dominant AI architecture provider.
"ARM's chip ambitions are strategically rational but the stock has priced in 4 years of flawless execution while ignoring fab construction delays and the risk that algorithmic efficiency improvements outpace capacity additions."
ARM's move into chip manufacturing is structurally sound (10x gross margin lift) but the valuation is already pricing in perfection: 23x forward P/E on $9 EPS in 2030 assumes flawless execution across fab construction, ASML equipment queuing, and sustained AI demand. The real risk isn't the strategy—it's the timeline. Fabs take 3-5 years to operationalize. If algorithmic efficiency (like DeepSeek's token-efficiency gains) materializes before ARM ships volume, or if ASML capacity constraints delay them past the AI capex peak, they've locked in $250B market cap for a licensing business with speculative chip revenue. AMD's 57% data center growth is real today; ARM's $2B pipeline is real but 2-year-out. The article conflates near-term momentum with long-term value.
ARM's energy-efficiency claim (2x better than x86, $10B capex savings per gigawatt) is exactly what hyperscalers desperately need right now, and $2B in pre-orders suggests genuine demand, not vaporware—the valuation may compress but the business case is sound.
"N/A"
[Unavailable]
"ARM's plan to become a chipmaker is a high-risk, multi-year undertaking that may not deliver the projected margin uplift or growth fast enough to justify current valuation if AI demand slows or supply-chain constraints persist."
ARM's pivot from licensing to in-house chipmaking could unlock higher margins if execution lands, but the path is riddled with risk. The industry’s supply chain is already stretched (ASML, Lam Research, KLA), and chip-capex cycles take years to ramp. Even with $2B of near-term demand for in-house ARM chips, achieving meaningful scale requires massive upfront spending, factory uptime, and favorable wafer/ equipment availability. Valuation appears to bake in an optimistic AI-driven growth story four to seven years out; any delays, higher costs, or softer AI demand could compress the earnings trajectory and margins far more than the market expects.
Counterpoint: if ARM sustains licensing as a moat while the manufacturing venture scales gradually, the downside may be limited; and if AI demand proves stickier than feared, the long-run margin uplift could still materialize, albeit slowly.
"ARM's transition to manufacturing shifts their risk profile from neutral IP licensing to high-stakes geopolitical target, threatening their Chinese revenue stream."
Claude and Gemini are missing the geopolitical volatility inherent in ARM’s pivot. Moving from licensing to manufacturing forces ARM into the crosshairs of export controls and trade protectionism in a way that pure IP licensing never did. If they build their own silicon, they become a direct target for US-China chip wars, potentially losing their largest market overnight. This isn't just an execution risk; it is a fundamental shift in regulatory exposure that the current valuation completely ignores.
"ARM's manufacturing move doesn't meaningfully increase regulatory exposure versus licensing; the real geopolitical risk is customer concentration among US hyperscalers."
Gemini's geopolitical angle is sharp, but overstated. ARM's licensing business already operates in this minefield—Huawei, SMIC, and others navigate export controls today. Manufacturing doesn't materially worsen their regulatory exposure; it actually diversifies it. The real geopolitical risk is hyperscalers themselves becoming targets if they rely on ARM-designed chips. That's a demand-side risk, not ARM's problem to solve. The valuation still ignores this downstream concentration.
[Unavailable]
"The manufacturing pivot's capex/depreciation and supply-chain risks likely erode required IRR, making licensing upside the safer play even if margins improve in theory."
Claude's line about 23x forward P/E assuming flawless fab ramp ignores the debt/return hurdles of multibillion-dollar capex. Even with a 10x gross margin lift, the required scale, uptime, depreciation, and wafer-tool shortages push IRR well below a 23x multiple if AI demand softens or yields lag. The real risk isn't just timeline—it's whether ARM can monetize manufacturing at scale before capital costs erode margin; licensing-only upside remains the safer path.
The panel generally agrees that while ARM's move into chip manufacturing has potential long-term benefits, the current valuation appears to be pricing in perfection and ignores significant risks, including execution challenges, geopolitical volatility, and potential demand shifts. The panel is bearish on the current valuation of ARM Holdings.
Potential long-term benefits of ARM's move into chip manufacturing, such as higher gross margins.
Execution risk in ARM's pivot to in-house chip manufacturing, including potential delays in fab construction, ASML equipment delivery, and shifts in AI demand.