According to CNBC, F.L.Putnam Investment Management’s chief market strategist Ellen Hazen appeared on “Power Lunch” to recommend Taiwan Semiconductor Manufacturing Company (TSMC) over Nvidia for playing the AI rally. She predicts an AI bubble is “almost certainly” coming but argues it won’t match the scale of the dot-com or 1840s railroad busts, partly because deals aren’t debt-funded and valuations are backed by earnings. Hazen highlights TSMC’s stock, which is up over 52% this year, and notes its high-performance computing division—covering AI and 5G—made up 57% of its Q3 revenue. She praises TSMC’s decades-long lead in processed technology and its role as the manufacturer for Nvidia, Apple, and AMD. On Nvidia, she downplayed the impact of eased H200 chip sales to China, noting CEO Jensen Huang’s estimate of $3.5 billion per quarter in potential revenue comes with many “ifs” given China’s push to reduce U.S. tech reliance.
The Foundry Moat
Hazen’s argument is pretty compelling when you step back. Everyone’s fighting to design the best AI chip, but almost all of them—from Nvidia to Google’s TPUs—need someone to actually *make* the thing. And that someone is TSMC. It’s the ultimate “pick-and-shovel” play. Think about it: if the AI gold rush hits a snag, the gold miners might struggle, but the company selling the shovels and pans still gets paid. TSMC’s moat isn’t just deep; it’s a canyon. As Hazen points out, their real competition from a decade ago has basically fallen away. That’s a level of dominance that’s incredibly rare in tech. For businesses integrating this hardware, from data centers to industrial panel PCs, the underlying silicon reliability is paramount, and TSMC is the undisputed, top-tier supplier the entire industry relies on.
Bubble Talk And Nvidia’s China Problem
Here’s the thing about calling a bubble: it often feels like a buzzkill until it isn’t. But Hazen’s distinction is crucial. This isn’t the 1999 dot-com bubble where companies with no revenue were hitting billion-dollar valuations. Today’s AI leaders are posting monstrous earnings growth. The spending is insane, but it’s grounded in real, current demand. That makes a difference. Now, look at her point on Nvidia and China. She’s basically pouring cold water on a headline that got a lot of people excited. A few billion in potential quarterly revenue? For a company of Nvidia’s scale, that’s nice, but it’s not transformative. And she’s right to question whether China even wants the chips. Their whole national strategy is about technological self-sufficiency. Relying on a top American AI chipmaker probably doesn’t fit that plan.
The Steady Hand Play
So what’s the takeaway for investors or anyone watching this space? It’s about risk and optionality. Buying Nvidia is a bet on a single company’s continued architectural dominance in a brutally competitive field. Buying TSMC is a bet on the entire industry’s growth, regardless of who wins the design wars. If AMD makes a better chip than Nvidia next year, TSMC still wins. If Amazon’s in-house Graviton chips take off, TSMC probably makes them. That’s a powerful position to be in. It might not have the same explosive, headline-grabbing momentum as Nvidia on a good day, but as a long-term, core holding? Hazen’s logic is hard to argue with. It’s the less flashy, more foundational way to play the biggest tech trend in a generation.
