Closing Bell Overtime 10/13/26

Micron Technology

Across the shows: 7 speakers in other episodes, 3 bullish and 2 bearish calls, past 30 days

• Gil Luria · D.A. DavidsonBullish

  China's CXMT memory push is a limited threat to Micron, which doesn't compete in high-bandwidth memory.

  Luria says the Chinese memory makers, though newly public, have been catching up but only compete at the low end and do not participate in high-bandwidth memory. Micron's biggest customers, Microsoft, Amazon, Google, Apple and Nvidia, cannot buy Chinese chips, so any impact falls mainly on SK Hynix and Samsung rather than Micron. He adds that Chinese producers cannot meet their own domestic demand for a while, and that the competitive threat is already in Micron's supply and demand math.

  “They only participate in the low end of the market. They do not participate in the high bandwidth memory part of the market.”at 10:28

• Gil Luria · D.A. DavidsonBullish

  Micron no longer sells a commodity, so its ~6x forward multiple should converge toward AMD and Intel.

  Micron trades at about 6x forward earnings while AMD and Intel trade above 40x, even though Luria sees the three as similarly competitive and similarly exposed to the AI cycle. He says the gap persists because investors still treat memory as a commodity with its own boom-bust cycle. He argues that has changed: large customers now sign long-term deals, so Micron sells a non-commoditized product, and the cycle hitting Micron is the same one hitting the rest of the chip sector. He says management expects tightness to worsen next year and the year after, and that they know the competitive roadmap, including China's. Micron is about 3% of Nasdaq 100 market cap but about 10% of its next-twelve-month earnings. He attributes the stock's underperformance to psychology: investors who sold after the run think they are done, and those who missed it won't buy in. If Micron reaches a market multiple in a year, that is about seven points of performance from one position.

  “So Micron no longer sells a commodity... the cycle that impacts Micron is the cycle that impacts everybody else.”at 13:06

Figure AI

• Andrew Kang · RoboStrategyBullish

  Figure AI's humanoid is as advanced as, or slightly more advanced than, Tesla's Optimus.

  Kang names Figure AI as one of the most prominent companies and says his firm thinks Figure's humanoid is just as advanced, if not slightly more advanced, than Tesla's Optimus. He offers no benchmark, but the claim frames Figure as a leading humanoid player despite being private.

  “we think their humanoid is just as advanced, if not slightly more advanced, than Tesla's [Optimus].”at 34:47

• Andrew Kang · RoboStrategyBullish

  A general-purpose humanoid form factor matters because it can scale to hundreds of millions or billions of units.

  The guest argues the humanoid is the most general-purpose form factor, citing general-purpose devices such as smartphones and GPUs that run many tasks. He says that generality is what allows scale to hundreds of millions or billions of units, which specialized robots cannot match as easily. He adds that specialized robotics will also be a large market, so he sees room for both. The speaker's link to the company is inferred from context, not stated in this part of the transcript.

  “We believe that's important because it allows you to scale. Scale to hundreds of millions, billions of units, something you can't do as much with a specialized device.”at 37:21

Apple

Across the shows: 14 speakers in other episodes, 4 bullish and 6 bearish calls, past 30 days

• Alex Kantrowitz · Big TechnologyUncertain

  Apple's home hub is not a step forward today, but it is strategically needed for the ambient AI shift.

  Asked whether Apple's new HomeView device is a real step forward, he says no today. Its current use is mainly information retrieval, such as recipes and FaceTime. He argues that ambient computing will become important as voice assistants gain agentic abilities such as booking restaurants and calling rides. Over the long term, he thinks ambient computing could make the phone and computer less important, since the AI does the work in the background. He says the launch matters because Apple needs a device like this, so that future AI updates run on Apple hardware rather than Google, Amazon or OpenAI devices.

  “that's why this is an important launch for Apple because it needs a device like this to take off and also just to be there for when the AI gets good enough”at 40:48

Meta Platforms

Across the shows: 17 speakers in other episodes, 19 bullish and 6 bearish calls, past 30 days

• Alex Kantrowitz · Big TechnologyUncertain

  Meta's Muse is more capable than expected and close to OpenAI's Dot, but trust in Meta remains a barrier.

  He admits he underestimated Meta's Muse, which he now says is much more capable than he expected. He credits standard models from Meta and xAI, not just frontier models from OpenAI and Anthropic, with doing a good job for agentic tasks. He concedes the gap between OpenAI's Dot and Muse is not large, even though they target different users. He still would not give Meta access to health records, email or bank accounts, so trust remains an uphill battle. He notes that many users will accept risk if a product gets the job done, and says Muse is doing that for many people.

  “I didn't give it a big chance to succeed and I'll say right here, I was wrong that it's much more capable than I anticipated”at 41:44

Robotics & autonomous vehicles

• Andrew Kang · RoboStrategy

  Physical AI robots are deployed now in narrow tasks, then generalize over time, like Waymo.

  Kang says robots like Taku are being deployed today, though not in the millions, starting with narrow tasks such as napkin folding, laundry folding and data center operations. He uses the Waymo approach as the model: narrow deployments come first and models generalize over time. As evidence he cites Path Robotics, which has been developing AI welding for almost a decade and needed hundreds of thousands of welding hours of data to make its model commercially viable. Path is now deploying with Huntington Ingalls, which recently signed a $600 million contract. His RoboStrategy fund invests in private robotics and physical AI companies.

  “So robots like Taku are actually being deployed today. They're not being deployed in the millions yet. They're starting out in specialized tasks like napkin folding, laundry folding, data center operations.”at 32:13

• Andrew Kang · RoboStrategy

  Humanoid robotics won't be winner-take-all; it will split into segments like the auto industry.

  Kang argues the robotics market will have many winners, not a single winner-take-all company. He compares it to the auto industry, where BMW, Lamborghini and Ferrari serve different segments, so each market segment may need a specific robot. Many of the most interesting companies are private, which is why his fund focuses there.

  “It's not gonna be winner take all. We think it looks something like the automotive industry where you have, right, the BMWs of the world, the Lamborghinis, the Ferraris.”at 34:47

• Andrew Kang · RoboStrategy

  US industrial policy will likely support robotics the way it backs rare earths and chipmakers.

  Kang says the US government wants to reindustrialize and build factories, and it has backed rare earths companies with large long-term loans and invested directly in semiconductor companies. He argues that if the government acts in similar ways, it is very likely to support robotics. He says that gives an opportunity in America, especially after the FCC ban on Chinese companies. His fund invests mainly in commercial and home robots and only a little in defense.

  “You've seen them support industries like the rare earths companies providing these really large long term loans... very likely they're gonna support the robotics industry in a similar way.”at 36:18

• Andrew Kang · RoboStrategy

  General-purpose humanoid form factors matter, as Apple and Nvidia's general-purpose devices show.

  Kang argues the human form is the most general-purpose form factor, which is why it makes sense for robots designed to do many things. He points out that the biggest companies, Apple and Nvidia, build general-purpose devices, citing the smartphone as an example of doing a huge range of tasks. He concludes general purpose matters.

  “So the human form factor is the most general purpose form factor. We think general purpose is important because if you actually look at the biggest companies in the world today, Apple, Nvidia, they have general purpose devices.”at 36:48

AI chips & semiconductors

• Gil Luria · D.A. Davidson

  AI chip supply-chain results and outlooks keep improving, contrary to the stocks' weak trading.

  Luria says results reported over the last eight to ten weeks have been positive across the value chain, including chip equipment makers, chipmakers, hyperscalers and model companies, and that companies say outlooks and visibility have improved. He sees this consistency as evidence the sell-off is not about fundamentals. He points to Broadcom's commentary as consistently positive, and attributes the recent drop to a misunderstanding about OpenAI's gross versus net revenue reporting, not a deterioration in demand.

  “All the results that have been reported over the last eight to ten weeks have been not only positive, but companies across the value chain have said that the results have improved, that the outlook has improved, and the visibility has improved.”at 8:22

AI bubble & tech valuations

• Marta Norton · Empower

  Chip demand is strong, but the question is how much valuation headroom remains after the rally.

  Norton says the fundamentals for chips are very strong and hard to argue with on the demand side. But much of that information was already understood and helped fuel the first-half rally, so the concern is how much valuation is left in the space, not whether the business is good. For index investors, she notes semis are roughly 19% of the S&P 500, so they may already have AI exposure and need not look for new ways to access the trade.

  “So I don't know if we necessarily question how good those business fundamentals are as much as we question how much valuation is left in the space.”at 15:47

US Treasuries & yields

• Marta Norton · Empower

  Bonds are the one attractive valuation area; with the 10-year near 5.3%, investors are paid for risk.

  Norton says the valuation discount investors look for is not obvious in equities, with the exception of the bond market. The 10-year yield has moved up toward 5.3%, which she says prices real risk that hasn't been priced for some time, and she says investors are now being paid for taking that risk. In her view, investors can dollar-cost average into bonds, and she would fund that from areas that have run, such as the AI trade or health care.

  “But now you're seeing the ten year move higher, close to 5.3%. You're getting paid for some of that risk. And so, in my mind, you can dollar cost average in here into the bond market.”at 21:43

US stock market outlook

• Marta Norton · Empower

  Health care stocks look aggressively expensive after their run, so they may be a trim candidate.

  Norton says health care is now screening aggressively expensive in her view, after doing pretty well this year. Given that valuation, she suggests trimming the sector and redeploying the proceeds, for example into bonds, rather than adding to it.

  “Health care is now screening pretty aggressively expensive in our view, so that may be an area you could trim as well.”at 22:09

AI & jobs

• Andrew Kang · RoboStrategy

  Robots take tedious restaurant tasks like napkin folding, letting workers move to higher-skill jobs.

  Kang describes a robot deployed at Din Tai Fung in New York, where hundreds of employees, sometimes several at once, fold napkins for long shifts. He argues that robots take over this tedious hospitality work, such as towel and clothes folding, so employees can do higher-skill jobs like making dumplings. He presents this as a productivity gain for workers rather than a job loss.

  “a lot of that tedious work that happens in these restaurants, in hospitality, ... is taken off the hands of those employees so that they can do, you know, higher skilled jobs elsewhere.”at 33:59

Venture capital & startups

• Andrew Kang · RoboStrategy

  Leading AI and tech companies will stay private longer, until they reach about a trillion dollars.

  Kang says companies like Anthropic, SpaceX and OpenAI are staying private for longer and only go public around a trillion-dollar valuation. He expects the trend to continue because corporations, sovereign wealth funds and other large pools of capital are funding these companies in private markets. His fund's focus on private robotics companies follows from this.

  “these companies stay private for longer and longer. Now it's... Until they're a trillion dollars that they start to go public, and we think that trend is gonna continue.”at 35:17

AI: overall outlook

• Alex Kantrowitz · Big Technology

  Agentic consumer AI use cases will arrive sooner than most people expect, using basic models.

  He says AI is progressing faster than most people expect, himself included. He points to basic models from Meta and xAI, not only the best frontier models, as already performing well on agentic tasks. From this he concludes that agentic use cases, such as the voice-driven tasks on home devices, are not far off. He argues that users will not have to wait much longer for them to show up in products like the iPhone and Meta's properties.

  “Yeah, I mean, [I] think we're getting there very quickly. Quicker than a lot of people imagine, including myself.”at 41:24