Good morning investors.
A more personal note to start this morning. I have been feeling something lately that I suspect a lot of you know well. Between the PhD, the day job, building my own book of business, and now writing two newsletters, there is rarely a moment where I feel like I have done enough. I finish one thing and the feeling is not satisfaction, it is a quiet sense that I should already be onto the next. I think a lot of ambitious people live in that gap, where no amount of output ever quite settles the feeling that more is expected, most of it self imposed. If you feel the same, you are not alone, and I have come to think the feeling is less a verdict on how much you are doing and more just the tax that comes with caring about the things you are building. Anyway, onto the markets, which have been giving all of us plenty to think about.
We are seeing something strange in this market, and TSMC put it on full display yesterday. The best chipmaker in the world posted its fifth record quarter in a row, beat across the board, raised its outlook, and the stock fell anyway. That reaction set off a broad chip rout that dragged the S&P down 0.5%, the Nasdaq 1.5%, and even pushed Nvidia's market cap back below $5 trillion. The Dow held up better, off just 0.2%. Oil kept climbing into the mid $80s as the Iran situation smolders and shipping traffic thins.
Futures are heavy again this morning, with the Nasdaq off nearly 2% and chips extending their slide worldwide. Japan's Nikkei closed down 4% overnight. The week's big earnings are behind us, so today the focus turns to Michigan consumer sentiment and how the market digests a soft reaction to Netflix.
The Bar Keeps Rising on the AI Trade
TSMC's quarter captures the mood of the market right now. Profit jumped 77% to $22 billion, management called demand extremely robust, and the company announced an additional $100 billion for Arizona on top of an already massive buildout. Those are extraordinary numbers. The stock fell anyway, and the reason tells you where investors' heads are. TSMC lifted its capital spending ceiling toward $64 billion, guided margins a touch below expectations, and the market focused entirely on the spending rather than the record results.
This is the pattern that has defined the group all summer. Every report out of the AI supply chain confirms that demand is accelerating and money is pouring in, and yet the stocks keep falling. The semiconductor ETF is down roughly 8% over the past month, and Micron and Nvidia are both off double digits from their highs, all while the companies keep smashing records. Banner quarters have become the baseline, and marginal good news no longer moves the needle the way it did.
There are two ways to read this. The charitable version is that investors are raising the bar on what counts as good news, which over time makes for a more rigorous and sustainable boom rather than a euphoric one. The cautionary version is that if rising capex keeps getting punished, the hyperscalers and chipmakers will eventually have to think harder about how freely they spend. The buildout itself may not slow, but the era of Wall Street cheering every dollar of it with no questions asked looks to be winding down. That is not a bad thing. A market that demands proof of returns is healthier than one that rewards spending blindly. It is just a different regime than the one that carried these stocks for two years, and the adjustment is uncomfortable.
The Damage, and the Twist
The technical picture has deteriorated. The semiconductor index broke below a key level yesterday and now sits less than 1% above the mark that would confirm a 20% decline from its June high, which is to say a bear market for the group. Global chip stocks have shed around $3.3 trillion since June 22. That is a genuine washout.
What caught my attention, though, was the twist in yesterday's session, and it matters for how I read all this. For most of the summer, when chips sold off, money simply rotated into the Magnificent Seven and the index held. Yesterday that changed. The megacaps fell too, but the equal-weighted S&P actually rose, lifted by transports, regional banks, homebuilders, and defensive names. This was not an everything down day. It was another hard step away from the old leadership and toward the parts of the market that have lagged. That rotation is exactly what I have wanted to see, and it is why I keep saying a chip correction is not the same as a market top. The money is not leaving equities. It is moving.
One more point worth making. Nvidia is nearly flat over the stretch that has erased trillions elsewhere and remains the best performer in the large cap chip group. The breakdown is happening despite Nvidia, not because of it, with the real pain concentrated in memory. That distinction matters. This is not a uniform collapse of the AI thesis. It is a violent repricing of the most crowded, most speculative corners while the highest quality name holds its ground.
How I'm Thinking About It
I have said for weeks that I would let the semiconductors come to me rather than chase them, and this is the environment that view was built for. Barclays put it well this morning, that while tech volatility may persist near term, the reset in positioning should ultimately prove healthy and create more attractive entry points for long term investors targeting the structural AI theme. That is close to my own read. I am not trying to catch the exact bottom in a falling group. I am watching for the point where the selling exhausts itself and the quality names are on sale, and we are getting closer to that than we were a month ago.
The counterweight worth respecting comes from the Bank for International Settlements, which cautioned in its annual report that boom-bust cycles are a regular feature of past investment surges driven by transformative technologies. That is the honest long term risk, and it is the same cyclical caution I have carried on memory all year. I believe this buildout is structural, backed by contracted demand and sold out order books, but intellectual honesty means acknowledging that every transformative technology in history drew in more capital than it could immediately return, and the reckoning came later. Watching the capex discipline the market is now imposing is one way that risk gets managed in real time.
Netflix Was Fine, and That Was the Problem
Netflix reported a solid, largely in-line quarter and the stock fell more than 8% after hours. I wrote about this yesterday, and the read holds. There was not much wrong with the report itself. Revenue grew 13%, earnings edged past estimates, advertising is still on track to roughly double to around $3 billion this year, and live programming keeps driving subscriber growth. The problem was expectations. Management narrowed its full year revenue outlook rather than raising it, kept margin guidance unchanged, and said it will only publish engagement data annually going forward. None of those is a real negative on its own, but together they left investors with nothing new to get excited about.
A good quarter is no longer enough when the valuation already prices perfection. I have not been particularly bullish on Netflix for a while, mostly because the price reflected so much optimism. If this weakness continues, though, it becomes a name I will look at more closely. Sometimes a great business simply needs time for the valuation to catch up with the fundamentals, and a selloff is how that gap closes.
China's AI Push and a Deepening Slowdown
Two things out of China worth noting. First, Moonshot AI unveiled a new model, Kimi K3, that it says surpasses some of the leading US systems on certain benchmarks. It still trails the very best from Anthropic and OpenAI overall, but it beat their second tier models on coding and agents, and it did so despite China's hardware constraints. Chinese models keep closing the gap while staying cheaper to run, which is a real competitive dynamic for the US labs and part of the pricing pressure I have flagged in the token economics. The release rattled Chinese rivals, with several competitor stocks falling sharply.
Second, China's economy grew just 4.3% last quarter, its slowest since 2022 and below the government's own target, with property investment down 18%, the worst on record. Industrial production and AI-linked exports are propping up the headline while the domestic side sags. It is a reminder that the AI boom is holding up parts of the global economy that would otherwise look considerably weaker.
Final Thought
The week ends with the chips in a genuine correction and the mood around the AI trade souring, even as the underlying numbers keep confirming that demand is strong. That tension is the whole story. Records are no longer enough, capex is being scrutinized rather than celebrated, and the most crowded corners are repricing hard. I do not read this as the boom ending, but as the market maturing into a phase where it demands returns rather than rewarding spending on faith.
The rotation out of chips and into the neglected parts of the market is healthy, not a warning, and yesterday's session where the equal-weight index rose while chips fell is the clearest sign yet that money is moving rather than leaving. I continue to favor the sectors picking up the baton, I am watching the quality semiconductor names for a better entry as the reset plays out, and I keep energy on as a hedge with Iran still live and oil in the mid $80s. The bar has risen, and for long term investors a market that asks harder questions is one worth being in.
Watch Michigan sentiment this morning, watch whether the chip index reclaims its broken level or slides toward the next one, and enjoy your weekend. If any of this has you wondering whether your portfolio fits your goals rather than the day's headlines, that is a conversation I am always glad to have.
Best regards,
Dan Sheehan [email protected]
Subscribe to Market Pulse: https://substack.com/@dansheehan3
Market Pulse with Dan Sheehan is a personal market commentary newsletter and is independent from my employer. The content is provided for informational and educational purposes only and reflects my views as of the publication date, which may change without notice. Nothing contained herein should be construed as personalized investment, legal, tax, or financial advice, or as a recommendation to buy or sell any security. Any positions discussed represent my own views and may not be suitable for every reader's objectives, financial situation, or risk tolerance. Information is derived from publicly available sources believed to be reliable, but accuracy and completeness cannot be guaranteed. Readers should conduct their own research and consult their own professional advisers before making financial decisions.