Bloomberg Surveillance · · 51 min

AI and the Equity Selloff

With David Schassler, Fiona Greig, Gene Munster, Greg Peters, Heath Terry, Ted Mortonson

ChorusBring the market's conversation to your AI. Learn more

Summary

Bloomberg Surveillance examined an equity selloff tied to AI, with guests Gene Munster, Ted Mortonson, Heath Terry, Greg Peters, David Schassler and Fiona Greig. Munster argued Nvidia's fundamentals beat expectations and that falling token prices, through a Jevons-paradox logic, will drive usage rather than hurt Anthropic-type model makers, while he sees AI adoption under 1% and the build still early. Terry stressed hyperscaler returns of about 28%, compute costs up 40%, and nearly $800B of debt and equity still needed for capex, arguing that supply pressure is pushing up debt rates and pulling down equity multiples. Peters described an economy where AI build-out drives growth, with chip financing emerging as a large asset class at yields above 8%, and Schassler said rising rates and mega-cap funding are pressuring nearly all assets, including gold, which he is positioned around by holding intermediate Treasuries and tilting toward technology. The panel broadly saw the selloff as a cyclical correction within a secular AI trend, while disagreeing on whether capex and rates will hold up.

Key takeaways

  • Gene Munster (bullish on NVIDIA): Nvidia's fundamentals came in far better than expected, yet the stock hasn't reflected it.
  • Ted Mortonson on Open-source AI models: Cheaper open-source models with about 80% of the functionality are a headwind to frontier models.
  • Heath Terry on AI infrastructure & data centers: Hyperscaler AI build-out earns high returns, but power, labor and permit limits cap how fast it grows.
  • Greg Peters on US growth & recession risk: US growth now depends on the AI build-out, so any slowdown in it has clear economic impact.
  • David Schassler on AI bubble & tech valuations: The market is disciplining AI: heavy borrowing and no credible ROI path are being punished.

What was said, by company and theme

24 views from 6 speakers. Every quote is checked against the episode's transcript (how).

NVIDIA

Across the shows: 179 speakers in other episodes, 187 bullish and 177 bearish calls, past 30 days

Where everyone stands on NVIDIA

  • Gene Munster · Deepwater Asset ManagementBullish

    Nvidia's fundamentals came in far better than expected, yet the stock hasn't reflected it.

    Nvidia reported much better fundamentals than people expected the day before the Anthropic news. Munster says the stock did not reflect that. He sees the setup for the AI trade as easy for stocks to move higher, which supports a bullish view on the stock.

    “We had much better fundamentals from NVIDIA than people were expecting. The stock didn't reflect that.”at 4:19

AI infrastructure & data centers

Where everyone stands on AI infrastructure & data centers

  • Ted Mortonson · Baird

    The infrastructure build needed over the next two to three years is massive; peak-earnings worries are overdone.

    Mortonson says the infrastructure build, which is years in the making, has been pushed off. The Street worries about peak earnings, and he says he understands that argument. He argues the build needed in the next two to three years is massive, so he does not see peak earnings.

    “Well, from my standpoint, this infrastructure build is gonna be years in the [building], and now it's getting pushed off... I understand the argument, but I think if you look at the infrastructure build that needs to be done in the next two to three years, it's massive.”at 10:47
  • Gene Munster · Deepwater Asset Management

    Hyperscalers will likely signal faster calendar 2027 build-outs at January earnings, a bullish update.

    Munster expects end-of-January hyperscaler earnings to say that, despite political changes, companies still plan to build faster in calendar 2027 than people think. He expects no slowdown in major headwinds, which would ease worries about 2028. Labor changes may raise data center costs by about 5%, but he says continued commitment is highly likely. He calls this the golden update and says confidence that the AI build is still early would follow.

    “I think the message we're gonna hear at the end of January is that despite any of the change in politics that they continue... These companies continue to believe that they're gonna build at a pace in calendar twenty seven faster than what people think.”at 12:49
  • Heath Terry · Citi Research

    About $800B of debt and equity still needed for AI capex will push up debt rates and compress equity multiples.

    Terry says about $200B of debt and equity came in the last week of October, and roughly another $100B of equity came this week. Nearly $800B of equity or debt still needs to be raised to fund capex for next year and the year after. He applies supply and demand: when supply is this large, debt rates rise and equity multiples fall, which he says is what is happening now.

    “there's nearly $800,000,000,000 of equity or debt left to be raised to fund the CapEx investments for next year and the year after. ... when there's that much supply, rates go up in debt and multiples go down in equity.”at 21:08
  • Heath Terry · Citi Research

    Hyperscaler AI build-out earns high returns, but power, labor and permit limits cap how fast it grows.

    Terry says hyperscalers earn a 28% cash return on invested capital, and spot and contract prices are very high. Compute costs are up 40% year over year, which he says should not happen given Moore's Law. The market is tight, but companies cannot build fast enough because of data center moratoriums, capacity, power, labor and capital limits. He says hyperscalers have less than one turn of debt and that the spending is voluntary, so they could slow the build if demand signals change.

    “The hyperscalers generate a 28% return, cash return on cash invested from the money they put into building this.”at 22:00

US Treasuries & yields

Where everyone stands on US Treasuries & yields

  • Greg Peters · PGIM

    High starting yields and bonds' protection if AI build-out stalls make fixed income attractive for 2027.

    Peters says the starting point matters because yields are clearly higher. Income and carry levels on recent deals, such as the SpaceX GPU deal at over 8%, are very attractive for any asset. Bonds also have a protective quality if the economy slows, for example if the build-out is slowed by power or political constraints. Together these point to a very different fixed income landscape in 2027.

    “So it's somewhat perverse, but I think the starting point matters. Yields are clearly higher. ... These are very attractive income and carry levels for any asset.”at 29:26
  • David Schassler · VanEck

    Negative on bonds because of rate pressure, yet long bonds for their stability.

    Schassler says he is negative on bonds, because higher rates are pressuring everything. He also says bonds are providing stability in the current market, so his firm is long bonds.

    “You're negative on bonds, but bonds are providing stability. So we are long bonds.”at 36:17
  • David Schassler · VanEck

    Adding 7-10 year Treasuries as portfolio ballast after holding fewer for some time; fixed income now near neutral.

    His firm is significantly more optimistic on fixed income than before and is running close to neutral. It owns a lot more Treasuries than it has for some time, specifically the 7 to 10 year maturities, which provide the portfolio's ballast. The aim is to hold equity exposure while diversifying with fixed income and real assets, so the portfolio can capture upside when markets turn while cushioning the downside.

    “We own a lot more treasuries than we have for some time, specifically the seven to ten year. That's providing our ballast, and then we're diversifying it all with real assets.”at 40:12

AI bubble & tech valuations

Where everyone stands on AI bubble & tech valuations

  • Heath Terry · Citi Research

    This AI selloff is a normal cycle correction; fundamentals and AI demand remain intact.

    Terry says this is the seventh correction of over 10% or 20% since the AI trade started, and he treats it as part of a cycle. He says investors should anchor on fundamentals. He points to enterprise demand for AI services and the returns customers earn as evidence that the underlying demand is still there.

    “This is part of a cycle. What we go back to are the fundamentals. We always have to ground ourselves in the fundamentals.”at 26:14
  • David Schassler · VanEck

    The market is disciplining AI: heavy borrowing and no credible ROI path are being punished.

    Schassler says the market is the adult in the room on AI. Companies that borrow too much get punished, and companies without a credible path to ROI get punished too. His firm is a trend follower, and the AI theme got so big that it pulled in everything, so the unwind now hits everything and makes markets more volatile.

    “The market's being the adult in the room as it comes to AI. You borrow too much money, you get punished for it. You don't show a credible path towards an ROI, you get punished for it.”at 35:21

US stock market outlook

Where everyone stands on US stock market outlook

  • David Schassler · VanEck

    Asset prices want to fall as capital drain and higher rates pressure nearly everything.

    Schassler says most assets his firm invests in are falling. Interest rates are rising because funding mega-cap IPOs and the CapEx build drains liquidity, and higher inflation adds to rates. Higher rates pressure everything. He notes the VIX near 30 and that the S&P and Nasdaq are close to giving up all their gains.

    “I look out at a market right now that wants to fall. Why do I say that? Because most of the things that we're investing in are falling.”at 34:55
  • David Schassler · VanEck

    Neutral on equities but overweight technology; the current selloff should fade and markets will turn.

    VanEck is neutral on equities overall but long technology. Boring stocks that are not growing are not the place to go, so investors should keep equity exposure and lean into growth, which means technology. He expects the current weakness to fade and markets to turn, so the goal is to capture as much upside as possible while cushioning the downside. Investors cannot take excess risk, but they also cannot run and hide.

    “We're neutral equities, but we're long technology. Running into boring stocks that aren't growing isn't the place to go.”at 39:59

Gold & precious metals

Where everyone stands on Gold & precious metals

  • David Schassler · VanEck

    Gold is falling under higher rates and a stronger dollar after running up too fast last year.

    Schassler says gold is suffering from the same pressures hitting other assets: higher interest rates and a stronger dollar. He says gold falls when rates rise, as in past cycles. Gold also went up too much too fast last year, so his firm changed positioning toward energy.

    “Gold is suffering from the same thing we were just describing about higher interest rates, a stronger dollar.”at 37:08
  • David Schassler · VanEck

    Gold is temporarily under pressure from rates but should catch a bid once rates stabilize; a long-term hedge.

    Gold compounds around 8% over the long run, versus about 11% for global equities and 5% for bonds. Gold tends to do best when equities and fixed income are under the most pressure. Its recent fall is a mean reversion after rising too far, too fast, and has turned into a downtrend because interest rates are pushing it lower. Once rates stabilise, gold should catch a bid. He expects it may still face more downward pressure, but investors who buy now should look smart 18 to 36 months out, and gold is a great hedge over ten years.

    “Gold's temporarily under pressure because it went up too much too fast, so was a mean reversion that's now transitioned into a falling trend simply because of interest rates are pushing it. When interest rates stabilize, you expect gold to catch a bid.”at 37:55

AI pricing & unit economics

Where everyone stands on AI pricing & unit economics

  • Gene Munster · Deepwater Asset Management

    Falling token prices will raise usage rather than produce profitless prosperity for model makers.

    Tom Keene argued the log-convex decline in token prices is a negative signal. Munster disagrees, calling the view that falling prices mean profitless prosperity for companies like Anthropic a well-traveled narrative. He invokes Jevons paradox: as pricing falls, usage rises, and he says the industry has already seen this. He also notes the threat that companies could bring models in house.

    “this whole Jevons paradox that's talked a lot about as pricing goes down, usage goes up. And we've seen that.”at 7:09

Open-source AI models

Where everyone stands on Open-source AI models

  • Ted Mortonson · Baird

    Cheaper open-source models with about 80% of the functionality are a headwind to frontier models.

    Mortonson cites Palo Alto's Nikesh Arora, who sees roughly 60% of his client base adopting open source. Open-source models are community-built, cheaper, and deliver about 80% of the functionality, which he calls a headwind for frontier models. He notes that Jensen Huang and Alex Karp both described open source as a big alternative to frontier models. He says institutional investors are debating how it plays out and that both could do well.

    “An open source model would be something that is more community and it's lower price and it's about 80% of the functionality. So that's a headwind as it relates to frontier models.”at 8:08

AI: overall outlook

Where everyone stands on AI: overall outlook

  • Gene Munster · Deepwater Asset Management

    AI adoption is under 1%, so the market is roughly a third of the way through value creation.

    Munster's firm was buying stocks during the selloff because it believes fundamentals are strong. His confidence comes from the view that AI is still early, with adoption under 1%. He estimates the industry is about a third of the way through creating value for investors. He cannot say when the market will recover, but sees the selloff as an opportunity.

    “The part that gives us confidence to still buy is a belief that we're still early and, you know, one less than 1% of adoption. I think we're probably a third of the way through the value creation for investors in this.”at 9:53

AI chips & semiconductors

Where everyone stands on AI chips & semiconductors

  • Ted Mortonson · Baird

    Semiconductor names are now secular rather than cyclical and should be bought from the fearful.

    The semiconductor industry has been hit hard on multiple concerns. Mortonson's firm is focused on 15 sectors and about 50 stocks, and he calls the relevant chip names secular, not cyclical. He says these names should be bought from the fearful. He also warns that investors should not catch a falling knife until it stops spinning.

    “And they're secular now. They're not cyclical. So there are secular names that should be bought from the fearful.”at 11:27

Cybersecurity

Where everyone stands on Cybersecurity

  • Ted Mortonson · Baird

    Security-driven upgrades of networking, storage and servers are a new catalyst under the AI trade.

    Mortonson argues that the infrastructure trade misses a security issue. Companies must upgrade their digital core, and many are scrambling for networking, storage and servers. If they do not move to new architecture and firmware, they risk being hacked through patches. He also says aging global infrastructure is being replaced faster than expected, with about a trillion dollars of security spend shifting, so he sees this as another catalyst.

    “All these companies are scrambling for networking, storage, servers. If they don't go to new architecture and firmware, they risk being hacked because of these patches. So we have another catalyst underneath the AI umbrella.”at 14:40

Enterprise AI adoption & AI software

Where everyone stands on Enterprise AI adoption & AI software

  • Heath Terry · Citi Research

    Enterprises are getting returns on AI spend; adoption will see air pockets but no straight-line reversal.

    Terry says CIOs and CTOs universally report that they are getting a return on AI spend. Technology moves faster than banks, pharma companies and hospitals can keep up with, so there will be air pockets and pauses. Adoption will still continue, as hyperscaler backlog grew 142% from enterprise customers such as Citi, JPMorgan, Unilever, Walmart and Pfizer. He says technology is never adopted in a straight line.

    “And the answer to that universally, if you talk to CIOs and CTOs is yes. We will see air pockets in this because the fact is this technology is moving far faster than any normal company can keep up with it.”at 26:35

US growth & recession risk

Where everyone stands on US growth & recession risk

  • Greg Peters · PGIM

    US growth now depends on the AI build-out, so any slowdown in it has clear economic impact.

    Peters describes a single-factor economy in which US growth is driven by the AI build-out. Any sign that the build-out is slowing or stalling has clear economic impact. This explains why stocks and AI now show an anti-correlation with the bond market.

    “We're in a single factor economic world where US growth in particular is driven by this build out. And so any whiff or any sense or any inkling of that build out slowing down or stalling just has clear economic impact.”at 28:28

Credit markets

Where everyone stands on Credit markets

  • Greg Peters · PGIM

    AI chip financing is a new asset class; its supply will crowd out capital and keep spreads wide.

    Peters calls chip financing an emerging asset class because of its sheer size. A $32B deal at over 8% yield came the day before, while only about $20B had come before. Next year's issuance is slated at $350B, so he expects wide spreads not to tighten soon. He is very much in the crowding-out story, which he sees as anticipatory now, with sovereign bond financing picking up in the back half of next year. He expects higher yields, wider spreads and more volatility.

    “the chip financing side is an emerging asset class... before this $32,000,000,000 deal yesterday, there was only 20,000,000,000 to date... next year is slated to be $350,000,000,000.”at 30:39

Commodities (broad)

Where everyone stands on Commodities (broad)

  • David Schassler · VanEck

    Commodities have been the rock star; VanEck's commodity strategy is still up about 40% this year.

    Schassler says commodities have been the rock star. His firm's commodity strategy is still up about 40% this year after being up about 60%. Last year's gold run drove performance, and the firm rotated toward energy, which was the 2026 theme.

    “Commodities have been the rock star. You know, we... We've run a commodity strategy still up 40% this year. Was up 60. Still up 40, though.”at 36:17

US tax & spending policy

Where everyone stands on US tax & spending policy

  • Fiona Greig · Vanguard

    Automatic enrollment in retirement plans lifts participation by about 30 points; Vanguard recommends it.

    Vanguard's research proposes several retirement reforms. The signature recommendation is automatic enrollment: about two in five plans and workers are currently auto-enrolled. Workers can still opt out, but the data show the feature boosts retirement participation by about 30 percentage points. Vanguard also recommends allowing default investments in IRAs, as employer 401(k) plans already do, because money rolled into IRAs often sits in cash for long periods.

    “Automatic enrollment boosts retirement dramatically to the tune of about 30 percentage points.”at 46:31

Across the shows

What other episodes Chorus covered said about the same companies in the 30 days up to this one.

  • NVIDIA179 speakers · 187 bullish, 177 bearish calls

Ask your AI who they are and what they said

Topics

  • AI adoption stage
  • AI build-out as growth driver
  • AI chip financing supply
  • AI correction as cycle
  • AI theme unwind
  • bond positioning
  • capex financing supply
  • commodities performance
  • data center returns and supply constraints
  • earnings fundamentals
  • enterprise AI adoption
  • fixed income positive tail
  • gold as hedge
  • gold under rate pressure
  • hyperscaler capex pace
  • infrastructure build-out
  • intermediate Treasuries as ballast
  • market downside
  • open-source vs frontier model competition
  • retirement savings policy
  • sector rotation into tech
  • secular chip demand
  • security spending on infrastructure upgrades
  • token prices and usage

More from Bloomberg Surveillance

All Bloomberg Surveillance episodes

Ask your AI about this episode

Chorus gives Claude, ChatGPT and other assistants every view from this episode and every other show Chorus follows. Ask who agrees, who changed their mind, and what else was said about the same companies.

A weekly email with the calls that mattered from the most important voices in investing. Unsubscribe any time. Privacy Policy

Chorus reports what speakers said. It isn't investment advice. How Chorus works · About Chorus