Bloomberg Surveillance · · 55 min
AI Scares and Inflation
With Bruce Wolfe, Celine Woo, Jason Wu, Masa Takeda, Tani Fukui
Summary
Bloomberg Surveillance covered AI-driven market jitters and inflation. Celine Woo of Lazard argued that headline uncertainty over oil, rates, the Middle East and elections has detached sentiment from strong AI fundamentals, pointing to recent earnings from ASML, TSMC, SpaceX and Meta, and she defended Alphabet's negative free cash flow as justified by cloud growth near 80%, while naming Nvidia her top holding in automation and robotics. Jason Wu of the IMF warned that AI capex, forecast to peak near 3% of GDP, is increasingly debt-financed through opaque channels such as private credit and circular deals, creating financial stability risks if rates rise. Masa Takeda argued Japan's governance reforms are lifting capital efficiency, with valuations up from about 13x to 17x earnings, and praised Hitachi's shift to grid services. Tani Fukui said tariffs are a small share of the economy, the Fed cannot target household inequality, and the 10-year yield near 5% is more consistent with fundamentals.
Key takeaways
- Tani Fukui on Tariffs & trade policy: Tariffs are a small share of the economy; uncertainty and modest CPI effects matter more.
- Celine Woo (bullish on Alphabet): Alphabet's negative free cash flow from AI capex is justified by cloud growth beating peers.
- Masa Takeda (bullish on Hitachi): Hitachi's shift to grid and infrastructure services brings recurring income; bullish.
- Jason Wu on AI infrastructure & data centers: AI buildout shifting from free cash flow to debt; financial stability risk if rates rise or AI slows.
- Bruce Wolfe on US entitlements: Social Security will need adjustments within five to ten years, though their shape is unclear.
What was said, by company and theme
24 views from 5 speakers. Every quote is checked against the episode's transcript (how).
Alphabet
Across the shows: 8 speakers in other episodes, 7 bullish and 1 bearish calls, past 30 days
Celine Woo · Lazard Asset ManagementBullish
Alphabet's negative free cash flow from AI capex is justified by cloud growth beating peers.
Alphabet turned free cash flow negative in the last quarter, which spooked the market. Woo argues the cash flow change is justified because cloud services revenue growth exceeded peers by a big margin, close to 80% last quarter by her recollection. The capex is building capabilities that support that growth, so she is not worried about the dip.
“Alphabet, for example, they already turned free cash flow negative just the last quarter. That's what spook the market, but they were justifying that cash flow direction change because the cloud services revenue growth was actually exceeding their peers by a big margin.”at 17:19
NVIDIA
Across the shows: 30 speakers in other episodes, 25 bullish and 11 bearish calls, past 30 days
Celine Woo · Lazard Asset ManagementBullish
Nvidia is central to automation and robotics; valuation compelling after recent correction.
Following the recent correction, many AI-exposed companies look compelling on valuation. Nvidia, Lazard's top holding, sits front and centre in powering automation and robotics. Woo frames automation as the core of Lazard's investment philosophy, which drives customer productivity gains.
“given the correction we just talked about, a lot of companies are providing pretty compelling evaluation. I mean, Nvidia, which is a household name, which is our top holding name is at the center, front and center powering this automation and robotics as well.”at 19:53
Hitachi
Masa Takeda · SPARX Asset ManagementBullish
Hitachi's shift to grid and infrastructure services brings recurring income; bullish.
Hitachi is no longer the ailing, low-margin manufacturer it once was. It has become a power grid solutions provider that looks after mission-critical social infrastructure, including power grids and railway systems. That makes it more service-oriented, with recurring income, which Takeda sees as a more attractive business model.
“So Hitachi is no longer the ailing low margin manufacturer that it once was. Has turned into a power grid solution provider. ... So, it's much more service oriented and recurring income”at 27:12
AI infrastructure & data centers
Celine Woo · Lazard Asset Management
Hyperscaler capex builds full-stack AI capabilities; Lazard has visibility on free cash flow recovery.
Lazard believes it has good visibility on when free cash flow recovers, because it has conviction in the strategy behind the spending. The capex builds full-stack capabilities: selling compute, custom chips, data, distribution and software, which she says is critical for the AI agent-driven economy. That is why hyperscalers stand out.
“We do. I mean, we think we have pretty good visibility because we have a good conviction in their strategy why they're doing it. Listen, ultimately this CapEx is to build the full stack capabilities and I think that's where its hyperscalers stand out.”at 18:23
Jason Wu · International Monetary Fund
AI buildout shifting from free cash flow to debt; financial stability risk if rates rise or AI slows.
The IMF observes that free cash flows across the AI ecosystem have been exhausted, so investment is increasingly financed by debt. Some debt is in public investment-grade bonds, which is transparent, while other financing is opaque: circular deals, private credit, and debt offered in return for future revenue commitments. The IMF worries about financial stability if interest rates keep rising or AI investment slows, especially for institutions holding that debt.
“what we have observed is that the free cash flows have been exhausted. So investments are turning increasingly to other source of financing, namely debt. And we do wonder in a environment of rising interest rates as well as potential slowdown in AI investments, what that might entail for financial stability”at 33:26
Jason Wu · International Monetary Fund
Peak AI investment is forecast at 3% of GDP, about double the internet-era 1.5% peak.
The IMF's peak-year forecast for AI investment reaches 3% of GDP. The early-2000s internet-era peak was about 1.5%, so this phase commands far more investment. Whether that level becomes unsustainable depends on the productivity gains AI delivers and its impact on corporates, which the IMF is still studying.
“This is a very important question. If we look at some numbers, we will see that the peak year forecast in terms of AI investments will reach 3% of GDP. By comparison, during the early two thousands in the Internet era, the peak investment was 1.5%”at 35:32
International equities
Masa Takeda · SPARX Asset Management
Japan's governance reforms are lifting ROE, a trend that can last for years.
Japan's governance reforms are about improving capital efficiency, not just checks and balances. They have run for about ten years, and only in the last three did the world catch up. ROE improved 5 to 10 points over the decade but is still very low versus the international average. Because Japan is a slow-moving culture, Takeda thinks the trend can be long-lasting.
“Essentially, Japan's corporate governance reforms is really about not just improving checks and balances, but it's really about improving the capital efficiency of businesses. ... return on equity has improved 5% to 10 in the last ten years, but compared to international average is still very low.”at 24:39
Masa Takeda · SPARX Asset Management
Japan's market rerated from 13x to 17x earnings, so stock picking must be more selective.
Japanese stock valuations have risen from about 13 times earnings several years ago to about 17 times now. Takeda says this means investors need to be more selective going forward. The market should be driven higher mainly by improving capital efficiency, not by multiple expansion.
“Just several years ago, was 13 times, for earnings now it's 17 times. So, we have to be a little more selective going forward, but it's really the improvement of capital efficiency that should drive the market forward.”at 25:36
Masa Takeda · SPARX Asset Management
Japan has market breadth with dozens of large companies, unlike Taiwan and Korea's few mega-caps.
In Taiwan and Korea, one or two dominant stocks skew the entire market, and each now has a market cap above $1 trillion. Japan lacks a company of that scale, but has dozens of companies with $300 billion to $400 billion market caps. That gives Japan breadth, beyond the AI and semiconductor trade that drove markets in the first half of the year.
“Taiwan and Korea, you always have one or two stocks, dominant stocks skewing the entire market... Now we don't have company of that scale in Japan, but then we have $300,000,000,000 market cap, $400,000,000,000 market cap companies by the dozens.”at 28:06
US rates & the Fed
Tani Fukui · MetLife Investment Management
Fed can't set different rates by income; it must focus on the macro economy.
Fukui argues the Fed's hands are tied because it lacks tools to redistribute wealth or set different interest rates for households at different income levels. It has to look at the macro economy, so it cannot focus on struggling households or AI-driven segments alone. Fed staff do analyse data on the distribution of earnings and spending.
“the Fed has its hands tied because it doesn't have the tools to redistribute wealth or have different interest rates for different levels of household income.”at 3:50
US Treasuries & yields
Tani Fukui · MetLife Investment Management
Current roughly 5% 10-year yield level makes sense given economic fundamentals.
The 10-year yield was near 5% (4.98% at the time) after a sell-off in which investors were afraid to step in front of rising yields. Fukui says the current level is in a better ballpark than a couple of weeks earlier and makes much more sense given the economy's fundamentals.
“Now I think this level approximately makes a lot more sense given the fundamentals of the economic economy.”at 5:33
Tariffs & trade policy
Tani Fukui · MetLife Investment Management
Tariffs are a small share of the economy; uncertainty and modest CPI effects matter more.
From prior federal work analysing tariffs, Fukui says tariffs are always a relatively small share of the total economy, around 2%. They are very big in certain sectors and regions, but overall not as much of a big deal. The uncertainty they create weighs more on businesses, and they show up in CPI only modestly, which is why the effect was not as large as some expected.
“it's always a relatively small share of the total economy... Very big in certain sectors and in certain parts of the country, but overall, not as much of a big deal. The uncertainty is what it was get... Gets a lot of the, businesses.”at 7:18
Oil & fuels
Tani Fukui · MetLife Investment Management
Higher oil prices are a net growth enhancer for the US, though painful for consumers.
Higher oil prices hurt consumers, but the US has become a net exporter of oil, so higher prices boost domestic GDP. Fukui concedes this is a GDP effect, not one felt by households, and notes skepticism about the claim.
“I know this is terrible for consumers, but oil, higher oil prices are a net growth enhancer for The US right now.”at 8:11
US consumer
Tani Fukui · MetLife Investment Management
A stock or retirement-account decline would halt consumer spending via the wealth effect.
Net wealth across the board has improved and growth has improved since last year, even for the 20th percentile and up, which is part of why consumers are doing well. A legitimate correction or bear market would be a problem. If 401(k)s and other investments fall, people will stop spending, which is a big problem for many households.
“If we finally get a legitimate correction or dare I say a bear market, what does that do to the American economy? It's a problem. ... If your retirement is going down, big problem for a lot of people, they'll stop spending.”at 9:53
AI chips & semiconductors
Celine Woo · Lazard Asset Management
Market underestimates memory capacity buildout; memory is the critical AI performance driver.
Memory companies are building large new capacity. Woo says it will take time to build the fabs and improve yields, and that the market focuses on the big numbers and may create oversupply. She argues the market underestimates why they are doing it: memory is becoming the critical driver of AI performance, and AI customers say memory bandwidth is the bottleneck preventing better performance.
“I think market is probably underestimating why they are doing it. ... memory companies are doing it because memory becomes the critical driver for the performance.”at 14:49
AI bubble & tech valuations
Celine Woo · Lazard Asset Management
Market sentiment has disconnected from strong AI and chip earnings; she expects that to persist.
The market is dominated by headline uncertainties, including oil, rates, the Middle East and elections, and does not want to deal with uncertainty during earnings season. Woo sees a growing disconnect between sentiment and fundamentals. Recent results from ASML, TSMC, SpaceX and Meta showed strong earnings and very strong guidance, and she expects the high-level trend to remain the same.
“We think it's really puzzling... What we see instead is growing disconnect between this market sentiment and the fundamentals. Look at the earnings so far, ASML, TSMC, SpaceX, Meta, strong earnings and very strong guidance.”at 15:45
Robotics & autonomous vehicles
Celine Woo · Lazard Asset Management
Automation drives productivity gains across technology waves; robotics is a subset of automation.
Woo says automation is what drives productivity gains for customers. Looking at history, different technologies drove productivity in different eras: steam engines, electricity, then transistors, the internet and now AI. Robotics is a subset of automation technologies, so Lazard looks for the best companies worldwide that deliver productivity gains.
“at the heart of our investment philosophy is automation, is what is really critical to drive the productivity gains for the customers... right now that is AI. I mean, at the end of the day, robotics is actually a subset of the automation technologies.”at 19:08
Manufacturing & industrial policy
Celine Woo · Lazard Asset Management
TSMC's Arizona fab shows a US workforce can be trained faster than expected.
Woo was once more sceptical that moving skilled fab labor to the US would take much longer. TSMC's Arizona fab changed her mind. Training, bringing people from home countries and working together can get it done much better than expected. She sees the US fab push as strategically sensible because the biggest customers are based in the US.
“I used to be more skeptical that that transition could be much more time consuming, but in the case of TSMC with their Arizona fab, I have changed my mind. You can actually get that done much better than people's expectation”at 21:32
Japan's economy
Masa Takeda · SPARX Asset Management
A weak yen helps Japanese corporate fundamentals; the fund's portfolio is naturally hedged.
The fund is broadly invested in global Japanese companies. A weak yen helps their business fundamentals rather than hurting them. Takeda says the portfolio is positioned so that whichever way the yen moves, it is naturally hedged for dollar-based investors.
“No. I think it's actually the opposite because we are broadly invested in global Japanese companies. So the weak [yen] actually helps the business fundamentals.”at 23:39
AI & jobs
Masa Takeda · SPARX Asset Management
Japanese workers are less worried about AI job losses because of labor shortages.
Unlike in the US, people in Japan are not too worried about AI taking away jobs. The reason is the labor shortage seen everywhere in the country. Japanese companies can therefore leverage AI rather than resist it.
“I think unlike in The US, people are not worried about AI taking away your jobs too much because of the labor shortage issues that we observe everywhere.”at 29:39
Stablecoins & crypto policy
Jason Wu · International Monetary Fund
Tokenized assets are gaining real use but remain small; scaling raises financial stability risks.
Tokenization is a growing area, particularly for tokenized equities and bonds, which now have real use cases even though they remain fairly small. The IMF's message is that removing barriers lets them scale, but as they scale, financial stability implications must be watched. One example is whether more tokenized real-world assets could lead to higher leverage in the financial system.
“we observed that tokenization is a growing area, particularly assets, including tokenized equities, tokenized bonds. Those things are now having real use cases. They remain fairly small. ... as they scale, we also need to be attentive to financial stability implications.”at 36:52
US–China AI & chip race
Jason Wu · International Monetary Fund
Cheaper Chinese AI models may question whether current US AI investment is needed, though they could also curb excess spending.
China plays a major role in the AI ecosystem through chips and some frontier models. Wu sees a possible positive outcome: if the AI frontier is pushed forward collectively, this might reduce some excessive investment without compromising the power of the end product. He also sees a negative scenario, in which cheaper models from China raise substantial questions about whether all of these investments are needed. He presents both as possibilities rather than a firm forecast.
“this might take away some of the excessive investments if you will... some of the cheaper models coming out from China is gonna raise substantial questions about whether all of these investments are needed.”at 39:30
Cybersecurity
Jason Wu · International Monetary Fund
AI is speeding cyberattacks on finance; same-day exploitation of vulnerabilities rose from about 20% in 2018 to nearly 80% in 2026.
Finance is more digitalized, so the attack surface is larger and cyberattacks can hit banks and financial infrastructure. Wu says AI has also increased the speed of attacks. The IMF's Global Financial Stability Report chart shows the share of vulnerabilities discovered and exploited on the same day rising to close to 80% in 2026, from about 20% in 2018. He links this to the pace AI models can bring to the cyber threat landscape.
“in an AI driven world, the speed of attack has also increased... has risen to close to eighty percent in 2026 compared to just twenty percent in 2018.”at 40:43
US entitlements
Bruce Wolfe · Alight
Social Security will need adjustments within five to ten years, though their shape is unclear.
Wolfe says adjustments will have to be made over the next five or ten years, but he says it is too early to tell what they will look like. He notes that younger workers are worried Social Security won't be there and are saving more as a result. That behavioral response is his only supporting evidence, and his view on the form of the reform is explicitly uncertain.
“I think the reality is there'll be adjustments made over the next five or ten years, so there's gonna have to be. What those adjustments look like is, kind of too early to tell.”at 48:45
Across the shows
What other episodes Chorus covered said about the same companies in the 30 days up to this one.
Topics
- 10-year yield level
- AI capex financing
- AI capex share of GDP
- AI-driven cyber attack speed
- China's AI models and investment needs
- Fed mandate limits
- Japan corporate governance reform
- Japan equity valuations
- Japan market breadth
- Social Security solvency adjustments
- automation and productivity
- automation and robotics leadership
- chip fab reshoring labor
- grid infrastructure services
- hyperscaler capex payoff
- hyperscaler free cash flow
- labor shortage and AI
- memory capacity buildout
- oil price impact on GDP
- sentiment vs fundamentals
- tariff economic impact
- tokenized real-world assets
- wealth effect on spending
- yen weakness and corporate earnings
More from Bloomberg Surveillance
- Bloomberg Surveillance TV: October 30th, 2026 · October 30, 2026
- Markets Price in Tech Spending and Midterm Elections · October 30, 2026
- Bloomberg Surveillance TV: October 29th, 2026 · October 29, 2026
- Outlook for Inflation, Oil, and Tech · October 29, 2026
- Instant Reaction: Fed Stands Pat; Lisa Cook & Beth Hammack Dissent · October 28, 2026
Ask your AI about this episode
Chorus gives Claude, ChatGPT and other assistants every view from this episode and the rest of the market's podcasts. Ask who agrees, who changed their mind, and what else was said about the same companies.
Chorus reports what speakers said. It isn't investment advice. How Chorus works · About Chorus