By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel's net takeaway is that while AI infrastructure spending remains robust, the current dip might not be the best entry point due to valuation reset risk and potential ROI scrutiny from hyperscalers. The 'intelligence superhighway' monetization is still years away, and any deceleration in capex guidance could trigger another leg down.
Risk: Deceleration or cratering of AI capex due to elusive ROI by 2025, leading to another 15-25% leg down in valuations.
Opportunity: Potential state-sponsored infrastructure demand decoupling from corporate profit cycles, creating a floor for capex.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
The rout in artificial intelligence (AI)-related stocks is ongoing.
Stocks of companies that provide major inputs into AI data centers have all fallen far from recent highs. That includes memory chip stocks such as Micron Technology and Intel, copper stocks such as Global X Copper Miners ETF, silver stocks such as iShares Silver Trust, and construction stocks such as Caterpillar.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
In recent months, investors have been dumping those stocks and many others like them as they began to question the unprecedented AI spending by companies like Meta Platforms and Alphabet and the potential returns.
Shares of companies involved in AI infrastructure fell by an average of nearly 7% in July.
Morgan Stanley (NYSE: MS) says to buy the dip.
In a July 28 note, analysts at the investment bank said they believe AI infrastructure will eventually become an "intelligence superhighway" that provides significant net benefits to economies worldwide. The analysts are bullish on that highway, despite the likelihood of speed bumps ahead, such as companies limiting their AI use due to cost concerns, and lower-priced competition from Chinese AI models.
Yet the analysts still like the rate of improvement in AI capabilities and the benefits of AI adoption, as well as the associated capital expenditures. As a result, given the pullback in associated AI stocks, their report said: "This point in time represents an unusually attractive buying opportunity."
In other words, buy the dip.
Before you buy stock in Morgan Stanley, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Morgan Stanley wasn't one of them. The 10 stocks that made the cut are built for long-term growth and could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you'd have $386,727! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $1,232,139!
That performance is why people listen. With a track record of beating the S&P 500 by 4x, Stock Advisor offers a distinct advantage. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built for the long haul.
Four leading AI models discuss this article
"The article glosses over concentrated capex risk and unproven AI ROI, making the 'unusually attractive buying opportunity' less obvious than presented."
Morgan Stanley's call to buy the AI infrastructure dip (MU, INTC, copper/silver miners, CAT) after a ~7% July selloff rests on long-term AI capex momentum outweighing near-term cost-pushback and Chinese competition. The article's bullish framing ignores that hyperscaler spending plans remain highly concentrated in a handful of names (META, GOOGL, MSFT, AMZN) whose ROI on AI is still unproven at scale. Forward P/E compression in memory and industrials already prices in mid-teens growth; any deceleration in 2025 capex guidance could trigger another 15-25% leg down. Valuation reset risk is real while actual 'intelligence superhighway' monetization stays years away.
The strongest case against is that Morgan Stanley may simply be talking its own book; if hyperscalers do not slow AI spend, the dip buyers win big and the article's cautionary speed-bumps prove immaterial.
"The market is shifting from pricing AI as a speculative growth narrative to demanding tangible ROIC, which will cause further volatility in capital-intensive infrastructure stocks."
Morgan Stanley’s 'buy the dip' thesis relies on the assumption that AI infrastructure spending will follow a linear path toward an 'intelligence superhighway.' However, this ignores the massive hurdle of return on invested capital (ROIC). We are seeing hyperscalers like Alphabet and Meta pivot from experimental spending to rigorous cost-discipline. If the revenue-generating applications of AI—beyond internal efficiency gains—don't materialize by Q4, we face a multi-year 'trough of disillusionment.' While I agree that physical infrastructure like copper (COPX) and power management is essential, the valuation compression in semiconductors like Micron (MU) reflects a cyclical supply-demand imbalance, not just a temporary 'speed bump.' Investors should prioritize power grid reliability over generic chip exposure.
If AI adoption triggers a genuine productivity supercycle, the current infrastructure pullback will be viewed as a generational entry point for the 'picks and shovels' providers.
"Morgan Stanley is right on long-term AI capex tailwinds but wrong to imply current valuations are panic-driven; they're merely 'less frothy,' which is not the same as cheap."
Morgan Stanley's 'buy the dip' call on AI infrastructure is defensible on capex fundamentals—Meta and Alphabet aren't slowing spend materially—but the article conflates two separate theses: (1) AI capex remains robust, and (2) current valuations reflect a dip worth catching. On (1), credible. On (2), we need specifics. MU trades ~11x forward P/E; CAT ~13x. Neither screams 'desperation pricing' versus history. The real risk: if AI ROI remains elusive through 2025, capex doesn't decelerate gradually—it craters. The article assumes a 'speed bump' narrative; reality could be a cliff.
If enterprise AI adoption stalls because LLMs plateau in utility while costs remain high, capex cycles can reverse violently in 12-18 months, making 'dip buying' a value trap rather than a contrarian opportunity.
"The AI hardware capex cycle remains highly cyclical and may deteriorate before the MS thesis of a durable rebound materializes, making the dip potentially riskier than it appears."
Morgan Stanley’s note hinges on AI infrastructure becoming a lasting macro beneficiary and a mid‑term dip creating a compelling entry point. The strongest counter is that the AI capex cycle is highly cyclical and likely to be a bell curve rather than a straight ascent: cloud demand could slow, price competition—from U.S. and Chinese players—worsen margins, and hardware rebuilds may be delayed or pared back. The ‘AI infrastructure’ label also covers large, commodity‑sensitive names (copper, mining, construction) that are not pure margin stories. If funding cycles compress or policy bites bite, the bounce may be shallower or delayed.
Bull case: AI spend persists and accelerates as cloud deployments scale, tight hardware supply cushions margins, and a rapid rebound in AI infra equities follows a dip. This could trap late buyers who expect a slow return to prior highs.
"Commodity infrastructure names have higher embedded beta to any 2025 capex deceleration than semis."
Claude's cliff vs speed-bump framing misses that hyperscaler concentration already embeds the crater risk. MU's 11x forward P/E assumes 19% EPS growth into 2026; if Meta/Google ROI scrutiny delays even one generation of GPU refreshes, memory pricing collapses 30-40%. Nobody has flagged how silver/copper miners' capex leverage amplifies this binary outcome far more than semis.
"Sovereign AI infrastructure spending provides a non-commercial floor for capex that mitigates the risk of a total market collapse."
Grok, your focus on memory pricing is correct, but you're ignoring the sovereign risk factor. Hyperscalers are no longer the only buyers; national governments are now subsidizing sovereign AI clouds to avoid dependence on Silicon Valley. This creates a floor for capex that isn't tied to immediate ROI, but rather to geopolitical security. Even if Meta or Google hits a 'cliff,' state-sponsored infrastructure demand could decouple from corporate profit cycles, making the 'value trap' thesis premature.
"Sovereign AI capex creates a floor, not a ceiling—it delays but doesn't prevent a memory pricing reset if corporate ROI falters."
Gemini's sovereign AI capex floor is real but overstated. State-backed infrastructure spending is fragmented, slow-moving, and typically lower-margin than hyperscaler capex. It doesn't prevent a memory pricing collapse if corporate ROI scrutiny hits hard—it just softens the landing. The geopolitical demand thesis also assumes governments will sustain spending even as domestic AI utility remains unproven. That's a 2-3 year bet, not a near-term valuation anchor.
"Memory pricing is less likely to crash 30–40% due to price floors from contracts and sovereign demand, which cushions downside for MU/CAT."
Re Grok: The memory-price crash thesis hinges on a near-perfect demand unwind; in reality, memory pricing is shaped by inventory cycles, supplier discipline, and long-term contracts that create price floors. Even if ROI slows, a 30–40% collapse seems unlikely this year; sovereign capex could cushion demand and keep price support. That said, MU/CAT remain vulnerable to multiple compression if AI ROI stays murky and hyperscaler capex stalls.
The panel's net takeaway is that while AI infrastructure spending remains robust, the current dip might not be the best entry point due to valuation reset risk and potential ROI scrutiny from hyperscalers. The 'intelligence superhighway' monetization is still years away, and any deceleration in capex guidance could trigger another leg down.
Potential state-sponsored infrastructure demand decoupling from corporate profit cycles, creating a floor for capex.
Deceleration or cratering of AI capex due to elusive ROI by 2025, leading to another 15-25% leg down in valuations.