Speaker · Updated
Kai Wu: latest market views
Sparkline Capital
Summary
Kai Wu is the founder and chief investment officer of Sparkline Capital, an investment firm that uses machine learning on unstructured data to measure intangible value and runs a suite of active ETFs, and he previously worked on Jeremy Grantham's asset allocation team at GMO.
Chorus has 20 views from Kai Wu across 2 episodes on Excess Returns since October 21, 2026. Their latest calls: bearish on Meta Platforms, Microsoft and Oracle.
Where they stand
Their latest bullish or bearish call on each company.
- Meta PlatformsBearish ·
- MicrosoftBearish ·
- OracleBearish ·
What they talk about most
The latest views
Newest first. Every quote is checked against the episode's transcript.
About half of the S&P 500 is tied to the AI infrastructure trade, so the index is highly concentrated in one bet.
Kai's holdings analysis classifies the S&P 500 into AI infrastructure (Mag Seven, chip stocks, and data-center firms such as Oracle), early adopters and laggards. He finds about 49-50% of the index sits directly in the AI infrastructure trade. Exposure has roughly doubled from about 25% in September 2019. Tech is about 35% of the index, and semiconductors are about half of that. Kai says the Nasdaq is about 67% in the same exposure, so QQQ and SPY holders who let winners run keep adding AI exposure unless they rebalance. He argues the index looks diversified by stock count but is highly correlated to one trade. Effective holdings are about 50, and the top 10 names are 36% of weight. He links this to market-cap weighting, which buys more of a stock as it rises. He notes that the S&P had its highest tech weighting right before the 2000 collapse. He attributes the rise both to fundamentals and to multiple expansion in AI infrastructure.
“when you do all the math, it turns out that... But 49%, let's call it 50% of the S and P is directly tied to this trade.”at 7:48
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
Enterprise AI adoption & AI software
AI will become table stakes, so today's laggard firms will shrink toward zero over time.
Kai uses the classic S-curve model of technology diffusion. Infrastructure is built first, then early adopters, then the early and late majority, with laggards last. He argues that AI will become table stakes, as the internet and calculators did. Companies that lean into AI now will hire, build products and streamline operations to gain share. Laggards will either go out of business or be forced to adapt. He cites e-commerce, where many retailers failed while Walmart adapted. He expects laggards to decline toward zero, though he says this may take a long time. He describes the current period as one of wide dispersion between adopters and laggards.
“in ten years, there will be no AI laggards. In ten years or twenty years, at some point, we'll all be using AI. In the same way we all use the Internet today, we all use the calculator.”at 9:41
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
Multinationals earned excess returns over domestic firms during the trade war; investors should stay global.
During the trade war, the popular move was to hide in domestic firms. Kai wrote a paper comparing multinationals with domestic firms and found that the excess returns went to multinationals. He attributes this to gains from trade, since big multinationals can choose their supply chains and office locations. He also cites selection bias, since multinationals are the best-positioned firms competing on a global stage. Third, multinationals are more intangible-intensive, and intangible assets scale over a larger market. He says investors shouldn't retreat into the less dynamic domestic sectors. He says investors shouldn't kill the golden goose of global multinational investing, whether the firms are US-based or international.
“Excess returns have been accrued to the multinationals. They're, like, the the world's best firms.”at 12:21
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
AI infrastructure & data centers
Infrastructure builders historically don't keep long-term AI profits; value flows to users of the technology.
Kai argues that in past technology revolutions the builders and providers of infrastructure did not end up as long-term winners. Railroads, electricity, automobiles, airplanes and the dot-com era all followed this pattern. In the dot-com case, telecoms laid excess fiber, the capital cycle collapsed, prices fell, and the telecom index is still well below its high. The value went to the users, such as Netflix, Meta and Google. He applies this to AI. If AI creates large value, the question is whether it accrues to consumer surplus, to businesses that use AI, or to AI producers. His historical view is that it accrues to users. So the current infrastructure winners may not be the ultimate long-term winners.
“I would just, you know, just venture to say that never in history... has the builder and the provider of the infrastructure been the long term winner? ... It's always, you know, ultimately gone to the consumers.”at 20:08
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
Enterprise AI adoption & AI software
AI adopters trade at no premium to laggards, and leadership should shift to adopters in the next phase.
Kai's data shows infrastructure firms have done well while early adopters and laggards have gone roughly flat. Early adopters trade at parity with laggards, so investors pay no premium for AI adoption. He draws on the history of technology cycles: after the infrastructure build-out, leadership usually transitions to the adoption phase. He argues that if AI generates ROI for users, early adopters should benefit through multiple expansion and business improvements. That makes identifying them the key question. He thinks the market is currently in the adoption phase, though he qualifies that the transition has not yet happened. The argument rests on the historical pattern and the flat relative performance of adopters, and he doesn't put a timeframe on it.
“the infrastructure firms have done really well. And the laggards and early adopters are basically just kind of sitting there... the valuations of early adopters are at parity with laggards... There's no premium for those guys.”at 20:35
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
Core S&P or Nasdaq holders already have enough AI exposure, so dedicated AI funds add redundant risk.
Kai ran a holdings analysis of a 30-fund AI group. Its top holdings are Nvidia, Micron, AMD, Broadcom, Alphabet, TSMC and SK Hynix, and its composition is close to the Nasdaq's, with about two-thirds in AI infrastructure. Kai says these AI funds are mostly infrastructure, with slightly more software and hardware and less Mag Seven than the Nasdaq. He concludes that an investor who owns a core SPY or QQQ position doesn't need a second dose of AI. The investor is already overweight the theme. If anything, they should seek diversification away from the capex cycle.
“I don't think you need a second helping of AI, dedicated AI funds, because you're already full.”at 26:09
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
AI infrastructure & data centers
Hyperscalers shifting to heavy capex is diluting their quality; capital intensity predicts weaker stock returns.
Kai argues that the megacap platforms earned their returns as asset-light businesses, such as Google's search model. The AI arms race is now pushing them into capital-intensive, asset-heavy operations, and outsiders like SpaceX and Oracle are joining. He cites a factor chart showing that the Mag Seven's capex-to-sales has risen over time. He says capital-intensive businesses are generally worse businesses, since utilities take a lot of capital just to stand still. He also cites academic factor evidence: firms that grow capex aggressively tend to underperform, and capex intensity predicts poor returns across sectors. He stops short of saying these companies will go bankrupt. His point is that the quality of their business is being diluted.
“capital intensive businesses are just worse businesses... the overall quality of their business is being diluted as they transition [from asset-]light to asset heavy.”at 29:20
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
OracleBearish
Oracle's entry into AI capex, with Larry Ellison's money, erodes the moats of incumbents.
Kai lists Oracle, under Larry Ellison, among the outsiders now racing into AI infrastructure, along with SpaceX and the neoclouds. He argues that the AI arms race is turning former asset-light businesses into capital-intensive ones. He asks where the moats are when a well-funded entrant can simply deploy money and create a new competitor. This implies Oracle's entry adds to overinvestment and weakens the competitive position of the existing players. The view is an implication drawn from the broader capex-arms-race argument rather than a detailed analysis of Oracle's own financials.
“Where are the moats when someone like Larry Ellison can just, you know, he brings a bunch of money and then, you know, suddenly a new competitor, all the Neo Clouds as well.”at 29:28
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
MicrosoftBearish
Microsoft's capital intensity now rivals utilities and dot-com-peak telecoms, so it is less attractive for investors.
Kai argues that capital-intensive businesses tend to have weaker moats and lower returns on invested capital. Microsoft's capital-to-sales ratio is higher than AT&T's at the height of the dot-com boom and higher than many utilities today, so he says it is turning itself into a utility. Based on the data, he thinks that is not where investors should be putting their money.
“You look at Microsoft and Meta. I mean, their capital to sales is higher than [AT&T] and at the height of the .com boom. It's higher than, you know, many utility firms today.”at 34:15
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
Meta PlatformsBearish
Meta's capital intensity now rivals utilities and dot-com-peak telecoms, so it is less attractive for investors.
Kai argues that capital-intensive businesses tend to have weaker moats and lower returns on invested capital. Meta's capital-to-sales ratio is higher than AT&T's at the height of the dot-com boom and higher than many utilities today, so he says it is turning itself into a utility. Based on the data, he thinks that is not where investors should be putting their money.
“You look at Microsoft and Meta. I mean, their capital to sales is higher than [AT&T] and at the height of the .com boom. It's higher than, you know, many utility firms today.”at 34:15
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
Emerging markets have shifted from an AI diversifier to a second helping of the AI trade.
Kai's rolling one-year correlation between the EM index (EEM) and his AI stock index averaged around zero, peaked near 0.5 in 2021, then dipped slightly negative. Over the past year or two it has run around 0.7. He attributes this to top EM holdings TSMC, Samsung and SK Hynix, which have become a large part of the index. On his AI exposure metric, EM is about as high as the S&P 500, so it no longer diversifies away from the AI trade.
“EM is as high as S and P on this dimension. So EM has gone from being a diversifier to being kind of a second helping on the AI trade.”at 38:18
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
Swapping AI infrastructure for diversifiers just swaps it for laggards, roughly doubling to quadrupling laggard exposure.
Kai tested diversifiers against an AI-infrastructure-heavy portfolio: traditional value (IWD), developed ex-US (EFA), equal-weight S&P (RSP) and US small caps (IWM). Infrastructure exposure falls to a low single-digit share, but laggard exposure rises from about 24% to somewhere between 40% and 80%. He concludes the mission is not accomplished because investors have traded one problem for another. He argues that AI should be seen as a spectrum rather than an all-in or all-out binary.
“The trade off is that you've just swapped out infrastructure for laggards... You've doubled to quadrupled your exposure to laggards.”at 39:17
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
Enterprise AI adoption & AI software
Early AI adopters across sectors and markets offer long-term AI upside without infrastructure's capital intensity.
Kai argues that if AI is truly transformative, the real risk is owning companies that early adopters put out of business. Holding early adopters lets investors capture the long-term upside of AI while avoiding the high valuations and capital intensity of infrastructure stocks. He finds these adopters outside US tech, in German and Japanese companies, small caps, industrials and pharmaceuticals, so the thesis diversifies across markets and industries too.
“Being put out of business by the early adopters. So I do think of early adopters as being kind of... threading the needle... where you can get exposure to the long term upside of AI while... not having the same exposure to high valuations or the capital intensity of these... infrastructure stocks.”at 40:16
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
Healthcare, as an information-intensive sector that has lagged, looks like an unpriced AI beneficiary.
Healthcare is one of the most information-intensive industries in the world. As a category it has underperformed over the past ten years despite long-term outperformance. Kai argues that if anything has AI upside that is not priced in, healthcare is an interesting place to look. He notes this is the thesis behind Cathie Wood's roughly 30% healthcare weight in ARKK, which is mostly biotech, life sciences and medtech.
“Health care is one of the most information intensive industries in the world. As a category, has actually underperformed in the past ten years despite having long term outperformance. If there's anything that has AI upside yet is not priced, health care as a category seems like an interesting place to look.”at 47:21
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
ARK Innovation ETFNeutral
ARKK's underperformance reflects factor tilts toward junk stocks, not poor stock picking by Cathie Wood's team.
In a CAPM regression ARKK's beta of 1.6 shows it has underperformed by about 2 percentage points a year since inception. A seven-factor model gives a beta of 1.26 and shows a small-cap tilt, a strong tilt toward expensive stocks away from value, and an even stronger tilt toward unprofitable stocks. Attribution shows the junk-stock tilt cost about 25 percentage points, and small-cap exposure hurt in a mega-cap market. Alpha is positive, so the stock picking is good but was given back through factor exposures that have not worked in this period.
“If you look at the alpha, it's actually positive. So what this says is actually, Cathy and her team, they are good stock pickers. They just unfortunately given a lot of that back by, you know, having factor exposures that have been less good over this time period.”at 50:41
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
Asset managers, brokers & exchanges
Fund performance dispersion is mostly explained by factor exposures, not stock-picking skill.
Kai ran his attribution tool on three active funds. Goldman's GVIP, a concentrated hedge-fund-style basket, did okay, but its results can be explained by owning momentum and junk stocks. The VanEck social sentiment ETF (BUZZ) did well, but its returns can be explained by owning junk and high-beta stocks. The Moat ETF underperformed, but its anti-momentum bets explain the shortfall. Over the constrained three-year period, junk stocks did well and quality stocks did poorly, which explains all three funds' results. He concludes investors should adjust for factor exposures first and then judge any residual idiosyncratic alpha.
“Over this constrained past three year period, actually, junk stocks have done well, and quality stocks have done poorly. I think that can be used to basically explain all the performance of these three examples.”at 54:14
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
Asset managers, brokers & exchanges
Dividend, value and low-vol ETFs have near-identical factor profiles, so one could cover all three.
Kai screened the largest dividend (VIG, SCHD), value (VTV, IWD) and low-volatility (USMV, SPLV) ETFs. All six have market beta below one, near-zero SMB, positive value and conservative-minus-aggressive exposure, and negative momentum and intangible value. Their excess-return correlations range from about 0.75 to 0.93. He concludes they are essentially the same fund under different names and sponsors, so one could pick any of them.
“The correlation between these six funds ranges from point seven five to point nine three. So they're basically the same fund. Here's six funds that are just doing all the same thing.”at 58:07
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
Traditional value's underweight to intangible-heavy, asset-light companies will keep costing it returns.
Traditional value funds hold laggards, have light exposure to new-economy companies, and spend less on R&D. Their holdings also have lower ROE because they are capital-intensive. Price-to-book captures tangible capital but misses intangible capital, which Kai links to innovation, R&D, global brands and high ROE. Traditional value therefore underweights asset-light businesses such as Google, and he expects companies like that to keep doing well.
“So the asset light businesses, like the Googles of the world that have done so well and likely will... Maybe not Google, but like companies like that, which will do well moving forward.”at 1:01:01
Excess Returns: The S&P 500 Is Just 50 Stocks. The Nasdaq Is a 67% AI Bet. Kai Wu on What Else Is Hiding in Your Funds ·
If enterprises see no AI ROI and stop buying, chipmakers and memory stocks will suffer.
Kai argues that AI infrastructure makers cannot earn a profit on what they build unless customers buy it, and customers will only buy if they see ROI. Early on, companies experiment. Eventually enterprises will ask whether an AI tool is worth about a million dollars a year, and if not they will shut it down, hitting memory stocks and chipmakers. He says the market cannot have it both ways, either AI transforms the economy and everyone wins or it does not.
“the fix and shovels can't actually drive a profit on ... all the infrastructure they're building unless people actually buy it. ... And no one will buy it if they don't see ROI.”at 42:08
Excess Returns: Rob Arnott on the Hidden Costs of Passive Investing and the Real Winners From AI ·
AI infrastructure & data centers
AI chips go obsolete fast, making data center capex a treadmill with a narrow ROI window.
Kai argues that buildings, land and fiber last a long time, as in past infrastructure booms, but GPUs become obsolete quickly. Companies must keep reinvesting just to stand still, and the window in which they can earn returns on this spending is narrow and must come quickly. He calls it a treadmill, a tough business, and suggests AI companies could come to resemble utilities or telecoms, with AI tokens as a commodity.
“But the GPUs themselves, because of how fast technology progresses, they go obsolete pretty quick. ... You have to just keep investing in these ships just to stand still ... It's a treadmill. It's a very, like, tough business to be in.”at 45:50
Excess Returns: Rob Arnott on the Hidden Costs of Passive Investing and the Real Winners From AI ·
Shows
- Excess Returns · 2 episodes
Ask your AI what Kai Wu thinks
Chorus gives Claude, ChatGPT and other assistants every view Kai Wu has shared on the shows it covers, next to everyone else's. Ask where they stand, what changed, and who disagrees.
Chorus reports what speakers said. It isn't investment advice.