Howard Marks’s view is that an AI bubble has not been established—not that AI stocks are safe. Marks focuses on whether investors have become oblivious to risk; widespread debate about a bubble may suggest they have not.
Published byEdited with GPT-6 LunaImages generated with GPT Image 2
Research answer

Create a landscape editorial hero image for this Studio Global article: Why does Oaktree Capital Management co-chair Howard Marks remain unconvinced that AI stocks are in a bubble despite warnings from Ray Dalio. Article summary: Marks’s position is “a bubble has not been established,” rather than “AI stocks are safe”: reporting describes him as unconvinced even as other prominent investors warn of speculative excess. [1] His central distinction . Topic tags: general, news, general web, user generated. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts wi
Howard Marks’s position is narrower than a bullish verdict: he is unconvinced that the AI trade has reached bubble conditions, but that does not mean he considers every AI stock fairly valued or every infrastructure investment sound. His distinction is about investor behavior—especially whether market participants have become so swept up in optimism that they ignore risk. 1
10
The Nasdaq-100 fell 5.5% over the reported week, while the Nasdaq Composite lost 4.5%. Those declines show that investors were repricing risk; they do not, by themselves, establish why prices had risen or whether the market had reached a bubble. 7
Marks’s reported reasoning turns on the presence of caution. When investors are openly debating whether a bubble exists, that awareness may be different from the risk blindness associated with speculative excess. It can restrain enthusiasm—but it cannot prove valuations are justified or prevent a bubble from forming later. 10
That distinction helps explain how Marks can remain unconvinced even amid steep losses and prominent warnings. Ray Dalio’s reported “classic bubble” assessment emphasizes debt accumulation and rising interest rates. Michael Novogratz has called AI the “biggest bubble of our lifetime,” while also expecting the speculative frenzy to keep growing. Their comments reflect more alarm about bubble conditions; Marks’s reported stance is that the defining psychology has not yet been established. 2
9
10
Marks’s December 2025 memo, Is It a Bubble?, discusses concerns about circular deals among AI companies. One example described in the memo is a technology company financing a customer that then spends money back with technology suppliers. Such arrangements merit scrutiny because money circulating among companies may not be the same as independently funded customer demand. 36
Debt-funded expansion is another concern discussed in reporting on Marks’s memo. If companies borrow to build AI infrastructure, investors need to assess whether the resulting capacity can generate enough earnings to support the borrowing—especially if demand or expected returns fall short. 11
The pressure is not hypothetical for every company in the sector: reporting has identified CoreWeave’s debt as a source of investor concern. But that concern about an infrastructure provider does not, on its own, establish that the entire AI investment cycle is a bubble. 8
The useful question is not only whether AI is a bubble. It is whether specific companies can turn demand into durable revenue and earnings, and whether their spending and financing make sense relative to those returns. Marks’s memo highlights reasons to examine how AI deals are financed, not simply how much money is flowing into the sector. 11
36
So the 5.5% Nasdaq-100 decline is a warning about market sensitivity—not a conclusive answer. Marks’s skepticism about the bubble label rests on investor awareness of risk, while his concerns about circular financing and debt show why caution remains warranted. Neither point guarantees that AI valuations are safe. 7
10
36
Studio Global AI
This page includes a source-backed answer you can continue inside Studio Global.
Howard Marks’s view is that an AI bubble has not been established—not that AI stocks are safe.
Howard Marks’s view is that an AI bubble has not been established—not that AI stocks are safe. Marks focuses on whether investors have become oblivious to risk; widespread debate about a bubble may suggest they have not.
He still flags risks including circular financing and debt backed AI expansion, which make the quality of demand and the returns on infrastructure spending important to watch.
Howard Marks’s view is that an AI bubble has not been established—not that AI stocks are safe. Marks focuses on whether investors have become oblivious to risk; widespread debate about a bubble may suggest they have not.
Published byEdited with GPT-6 LunaImages generated with GPT Image 2
Research answer

Create a landscape editorial hero image for this Studio Global article: Why does Oaktree Capital Management co-chair Howard Marks remain unconvinced that AI stocks are in a bubble despite warnings from Ray Dalio. Article summary: Marks’s position is “a bubble has not been established,” rather than “AI stocks are safe”: reporting describes him as unconvinced even as other prominent investors warn of speculative excess. [1] His central distinction . Topic tags: general, news, general web, user generated. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts wi
Howard Marks’s position is narrower than a bullish verdict: he is unconvinced that the AI trade has reached bubble conditions, but that does not mean he considers every AI stock fairly valued or every infrastructure investment sound. His distinction is about investor behavior—especially whether market participants have become so swept up in optimism that they ignore risk. 1
10
The Nasdaq-100 fell 5.5% over the reported week, while the Nasdaq Composite lost 4.5%. Those declines show that investors were repricing risk; they do not, by themselves, establish why prices had risen or whether the market had reached a bubble. 7
Marks’s reported reasoning turns on the presence of caution. When investors are openly debating whether a bubble exists, that awareness may be different from the risk blindness associated with speculative excess. It can restrain enthusiasm—but it cannot prove valuations are justified or prevent a bubble from forming later. 10
That distinction helps explain how Marks can remain unconvinced even amid steep losses and prominent warnings. Ray Dalio’s reported “classic bubble” assessment emphasizes debt accumulation and rising interest rates. Michael Novogratz has called AI the “biggest bubble of our lifetime,” while also expecting the speculative frenzy to keep growing. Their comments reflect more alarm about bubble conditions; Marks’s reported stance is that the defining psychology has not yet been established. 2
9
10
Marks’s December 2025 memo, Is It a Bubble?, discusses concerns about circular deals among AI companies. One example described in the memo is a technology company financing a customer that then spends money back with technology suppliers. Such arrangements merit scrutiny because money circulating among companies may not be the same as independently funded customer demand. 36
Debt-funded expansion is another concern discussed in reporting on Marks’s memo. If companies borrow to build AI infrastructure, investors need to assess whether the resulting capacity can generate enough earnings to support the borrowing—especially if demand or expected returns fall short. 11
The pressure is not hypothetical for every company in the sector: reporting has identified CoreWeave’s debt as a source of investor concern. But that concern about an infrastructure provider does not, on its own, establish that the entire AI investment cycle is a bubble. 8
The useful question is not only whether AI is a bubble. It is whether specific companies can turn demand into durable revenue and earnings, and whether their spending and financing make sense relative to those returns. Marks’s memo highlights reasons to examine how AI deals are financed, not simply how much money is flowing into the sector. 11
36
So the 5.5% Nasdaq-100 decline is a warning about market sensitivity—not a conclusive answer. Marks’s skepticism about the bubble label rests on investor awareness of risk, while his concerns about circular financing and debt show why caution remains warranted. Neither point guarantees that AI valuations are safe. 7
10
36
Studio Global AI
This page includes a source-backed answer you can continue inside Studio Global.
Howard Marks’s view is that an AI bubble has not been established—not that AI stocks are safe.
Howard Marks’s view is that an AI bubble has not been established—not that AI stocks are safe. Marks focuses on whether investors have become oblivious to risk; widespread debate about a bubble may suggest they have not.
He still flags risks including circular financing and debt backed AI expansion, which make the quality of demand and the returns on infrastructure spending important to watch.