Fed Decision and Tech Earnings

Meta Platforms

Across the shows: 53 speakers in other episodes, 46 bullish and 28 bearish calls, past 30 days

• David Schassler · VanEckBullish

  Meta's September rally shows AI paying off via ad growth and news adoption.

  He points to Meta in September, which was up 27% on increased advertising and news adoption, as proof of concept that AI is already producing results. He expects more of that as the technology moves through its early stages.

  “look at Meta in September. Right? It was up 27% off of increased advertising as well as news adoption. Right? You're gonna see more and more of that.”at 27:17

Microsoft

Across the shows: 17 speakers in other episodes, 13 bullish and 1 bearish calls, past 30 days

• Ed Ludlow · BloombergNeutral

  Microsoft's 60,000 job cuts are belt-tightening to protect free cash flow amid heavy AI spending.

  Wall Street now cares about free cash flow. Microsoft is still essentially a software sales and distribution business, so it does not need as many people. If it stays committed to spending at its current AI levels, something has to give, and workforce reduction is the mechanism for belt-tightening. Ludlow frames it as a straightforward story rather than a directional call on the stock.

  “Wall Street cares about free cash flow now. It really does. And if you are committed to spending at the levels that Microsoft is something has to to be a mechanism for belt tightening.”at 32:24

Alphabet

Across the shows: 18 speakers in other episodes, 17 bullish and 1 bearish calls, past 30 days

• Ed Ludlow · BloombergUncertain

  Cloud is the upside surprise; market doubts AI lifts core search profits.

  Google Cloud is likely to grow about 91% year on year, and the market sees the cloud juggernaut as the source of any upside surprise. Alphabet has had problems, including key talent leaving to start their own companies. Investors still tie Google to core search and want evidence that search is made better and more profitable by its AI work. Cloud growth above 90% is evidence AI is going well only if the market links it to the core business, and it has not yet.

  “it is the juggernaut of Google's cloud business. ... People still think it's the core search business and they want to see evidence that that core search business is somehow made better, is somehow made more profitable by all of the work that's gone into AI.”at 34:27

AI infrastructure & data centers

• Dominic Konstam · Mizuho Americas

  AI chips depreciate slower than feared, so capital deepening is greater than past cycles.

  Chips from a few years ago are still heavily used, so depreciation runs slower than many expect, and that is one of the issues behind AI valuations. Because less capital has to be replaced quickly, the net capital stock grows more, making this a more profound capital deepening than in past cycles when depreciation was much faster. That undercuts the bear argument that AI capex is quickly wasted.

  “I understand you're still using a lot of chips that came out a few years ago that still have a tremendous amount of sort of use in them still, which is basically allowing the capital deepening to be even greater basically because there's less depreciation.”at 17:57

• David Schassler · VanEck

  Beaten-down AI infrastructure builders are a buy, since power demand holds whoever wins the model race.

  Winners in the AI model race keep changing, since a new model gets only six to eight weeks before being leapfrogged. Whoever wins, everybody pays their electric bill, so the infrastructure builders benefit. Many of these stocks are beaten down and trade near where they started the year, so he advises buying the dip very broadly. VanEck is leaning into this through infrastructure exposure.

  “No matter who wins on the technology side, everybody's gonna pay their electric bill. Infrastructure development stocks, a lot of these are beaten down significantly... So buy the dip on those, do it very, very broadly.”at 21:40

• Ed Ludlow · Bloomberg

  Industry is moving AI workloads to on-prem and edge data centers, not everything to the cloud.

  Ludlow pushes back on the common premise that AI data centers are being built around the country and everything is moving to the cloud. He argues the direction of travel has been the opposite: companies big and small are building small on-premises data centers, which he calls edge computing, because they want their AI to run locally. Tom Keene's follow-up paraphrases this as AI running on personal computers, closed off from the world. The view rests on the stated motive (local control of AI workloads) and a described industry trend, with no quantified evidence.

  “Actually the direction of travel in industry has been the opposite. Companies big and small are building little data centers on prem ... because they want to have all of that A. I. Run locally to them.”at 36:59

US rates & the Fed

• Constance Hunter · Economist Intelligence Unit

  Whether the Fed hikes today or in December barely matters for the underlying economy.

  Her team is split. One side sees underlying non-energy inflation moderating, so it expects one hike and then a downward trajectory. The other side wants more hikes. The key question is whether energy price spikes spill into services. Whether the Fed hikes today or in December, or insures a cut while signaling toughness on inflation, changes little for the underlying economy.

  “If you look at the underlying economy, does it matter if they cut... If they hike today or hike in December, it really doesn't matter that much for the underlying economy.”at 4:45

• Dominic Konstam · Mizuho Americas

  Fed's inflation credibility anchors long-end yields, keeping 10- and 30-year rates from reaching 6%.

  Inflation expectations, especially at the front end, matter far more for inflation than the Fed itself, though a Fed hike may help bring them down. The Fed's main value is credibility: by being seen to care about inflation, it stabilizes the long end of the market. That credibility is what stops tens and thirties from rising to around 6%, with tens already sitting near 5%.

  “But the bigger issue is by being seen to be caring about inflation, they get credibility for the long end of the market, which stops interest rates going up to say 6% in tens or thirties.”at 15:01

US inflation

• Constance Hunter · Economist Intelligence Unit

  Rising global frictions push up prices and demand, unlike the friction-reducing 1990s.

  The 1990s backdrop had falling frictions: free trade was advancing, more countries were becoming democratic and capitalist, institutions supported coordination, and the peace dividend followed the fall of the Berlin Wall. Today frictions are rising and global institutions are not easing them, so security and supply constraints are pushing up prices and in some cases demand. That makes today's inflation backdrop different from the 1990s tech boom.

  “we are increasing frictions, where global institutions are not easing global frictions... security, supply... are really pushing up prices and in some cases pushing up demand.”at 6:00

• Constance Hunter · Economist Intelligence Unit

  AI is a small share of CPI for now, but its inflation effect could grow.

  AI-related spending is not a large percentage of the consumer price basket, so it is not a big inflation driver today. The open question is whether that share grows as AI expands, which would add demand-pull inflation. She says it is unknown and that the team is closely monitoring it.

  “It's not a large percent of the basket, but the question is, can that grow... As this continues to grow? And that is, I think, an unknown, but something that we are closely monitoring.”at 6:25

US consumer

• Constance Hunter · Economist Intelligence Unit

  Card debt tracks consumption, but delinquencies are rising among a significant stretched cohort.

  Measured against overall consumption, credit card debt and spending are growing in line, so the aggregate ratio is not alarming. The concern is that delinquencies are starting to tick up. The stress sits among a significant cohort of households and consumers who are really stretched.

  “I think you need to look at credit card debt and credit card expenditures as a ratio of overall consumption. ... it's really in line. But then if we look at where our delinquency is starting to tick up... And there's a significant cohort of them.”at 8:01

AI & jobs

• Dominic Konstam · Mizuho Americas

  AI job cuts are reabsorbable as productivity rises; bigger risk is income and capital distribution.

  The workforce is shrinking, yet the economy still adds enough jobs to hold unemployment flat, and Microsoft's cuts may reflect AI-driven change. Past productivity events were followed by labor demand recovering and workers being reabsorbed as productivity expands, so he is not afraid of permanent mass unemployment. The real issue is income and capital distribution: labor may be reabsorbed on less favorable real-wage terms, which hurts people's ability to own pensions and houses. That K-shaped outcome is where policy has to act.

  “I wouldn't be afraid of this sort of Luddite sort of idea that, we're forever gonna create a mass unemployment that somehow can't be reabsorbed. I think the bigger issue is actually income and capital distribution.”at 13:06

US Treasuries & yields

• Dominic Konstam · Mizuho Americas

  10-year Treasury yields at 5% are too high; fair level nearer 4%.

  With inflation at 2% and real growth around 3%, a government security should yield close to 4%, not 5%. High rates are a big problem for debt service and limit the government's discretionary spending and its ability to stabilize debt. Yields could end up somewhat lower, but the inflation problem is still there and will take a long time to fix.

  “I think the idea is that inflation is running at 2% and a real growth, even if it was 3%, you still probably ought to have a government security that's close to four, not five.”at 15:51

Enterprise AI adoption & AI software

• Dominic Konstam · Mizuho Americas

  Business take-up of language models is just beginning; companies aren't spending much yet.

  Based on the clients he interacts with and his own bank, take-up of language models is still only just beginning. Early use was mostly a substitute for search. He uses Claude a lot now, but the business side is only starting and customers are not yet spending much on it.

  “the take up of these language models is really still only just beginning. ... So I think from the business side, this is thing is really just beginning. And we're not spending that much money on it yet in terms of a customer.”at 18:49

AI bubble & tech valuations

• David Schassler · VanEck

  AI is the most-hated trade, but it is very early and the trend will run a long time.

  People are already writing AI off, yet the technology is still being refined and improving faster and faster, infrastructure is still being built, and companies are still working out how to use it. Bubbles are not popped when people know exactly how it will happen, and bull markets don't end this way. He thinks the trend runs a very long time, with five to seven years likely to be needed before payoff is visible. VanEck is leaning into the theme.

  “bubbles aren't popped when people exactly know how it's gonna happen. Right? Bull markets don't end this way. We're very, very early in this. We think it runs for a very, very long period of time.”at 22:50

Gold & precious metals

• David Schassler · VanEck

  Gold heads well above $5,500; catalysts now stronger than when it peaked.

  Gold peaked above $5,500 and has pulled back to around $4,000. The reasons people hold gold, including debasement, inflation, geopolitical shocks, weaponization of the dollar and broad uncertainty, are significantly more elevated now than at the peak. The earlier correction came from overbought prices, higher rates and a stronger dollar, which are not the same forces now. He thinks gold goes a lot higher, beyond $5,500.

  “The catalysts for gold are significantly more elevated now than they were at $5,500. ... We think gold goes a lot higher.”at 25:09

Commodities (broad)

• David Schassler · VanEck

  Scarce physical assets are the ultimate bottleneck, so commodities should benefit as AI drives growth.

  His bottleneck framework says scarcity preventing growth is now in physical assets, which are the building blocks of growth. AI needs physical assets for its build-out, and the productivity gains it promises would raise growth and living standards, adding to demand. He wants exposure across the whole commodity complex, from copper, metals and energy to rails, MLPs and the data centers built from them.

  “And scarcity are now physical assets. Physical assets are the building blocks of growth. They they are the ultimate bottleneck.”at 26:35