The End of "Buying AI Blindly": Howard Marks Warns of a Coming Shakeout
Howard Marks, co-founder and co-chairman of Oaktree Capital Management, wants investors to slow down and rethink the hype around anything labeled “AI.” He’s clear-eyed about AI’s potential—it’s a real game changer, not some bubble. But the rush to pour money into every so-called AI opportunity has made things risky and volatile.
US Markets: NVIDIA—The "Pick and Shovel" Play with Height Risks
When you look at the US markets, NVIDIA stands out. Everyone sees it as the backbone for AI infrastructure. Their chips power much of the AI world, and the company’s lead feels solid—almost unbreakable. Big banks like Goldman Sachs and Morgan Stanley watch NVIDIA closely, especially its data center revenues and the rollout of new chips. The problem? With so many investors crowding into this “pick-and-shovel” play, the stakes keep growing, along with the risks.
The investment thesis is straightforward: AI requires compute, NVIDIA provides the compute, and demand currently outstrips supply. The company generates extraordinary margins and cash flows that would be the envy of any technology company.
However, Marks's framework suggests caution. NVIDIA's valuation reflects not just optimism but euphoria. The stock trades at premium multiples that require not merely good results, but results that "crush expectations" to maintain current prices. The risk asymmetry is concerning: if technology giants like Microsoft or Google reduce capital expenditure, or if NVIDIA's chip development encounters setbacks, valuation compression could be severe.
The prudent approach, following Marks's philosophy, involves patience. Rather than chasing momentum, investors should wait for broader market corrections or sentiment shifts that present more attractive entry points. The business remains exceptional; the price may not be.
European Markets: ASML—Geopolitical Complexity Meets Technological Necessity
ASML Holding isn’t just Europe’s tech giant—it’s the backbone of global chipmaking. This Dutch company pretty much owns the market for extreme ultraviolet (EUV) lithography, which is critical if you want to crank out high-end AI chips.
Honestly, without ASML, there’s no way to make those advanced chips. In 2024, they smashed records with net sales hitting €28.3 billion and net income at €7.6 billion. Orders are piling up too—a backlog of about €70.9 billion. The leadership expects sales in 2025 somewhere between €30 and €35 billion, with fat gross margins of 51-53%.
Still, it’s not all smooth sailing. ASML sits right in the middle of the complicated mix Marks talks about. Geopolitics make things risky. China bought over a third of ASML’s chips last year—36.1% of net sales—but export restrictions keep getting tighter. Western governments want to clamp down on how much advanced tech gets shipped to China, and that’s a big deal for ASML. The company acknowledges that geopolitical developments regarding export controls have "stirred volatility within the investment community".
CEO Christophe Fouquet notes that AI has created "a shift in market dynamics that is not benefiting all of our customers equally, which creates both opportunities and risks". This uneven distribution of AI benefits across the semiconductor ecosystem introduces uncertainty even for a company as dominant as ASML.
The investment strategy here aligns with contrarian principles: accumulate positions during periods of pessimism, particularly when short-term geopolitical fears or semiconductor cycle concerns drive temporary price weakness. ASML's technological moat and irreplaceable position provide long-term protection, but entry price remains crucial.
Conclusion: Discipline in the Age of Transformation
Howard Marks's message is not that AI is over-hyped—he explicitly acknowledges its revolutionary potential. Rather, his warning centers on the gap between technological reality and investment returns. The winners in transformative technologies are rarely obvious in advance, and the prices paid for perceived certainty often preclude satisfactory returns.
For investors navigating this landscape, Marks offers a framework: avoid the crowd, demand margin of safety, and remember that even the most compelling stories can become bad investments at the wrong price. The AI revolution’s real. The profits from investing in it are far less certain.
