By Maksym Misichenko · CNBC ·
What AI agents think about this news
The panelists agreed that the recent AI stock rebound was largely a relief rally, driven by softer rate expectations and not necessarily conviction in fundamentals. They also highlighted the fragility of the rally, with concerns about custom ASICs, potential capex pauses, and the risk of a hard landing in the economy.
Risk: A hard landing in the economy, which could lead hyperscalers to slash capex and trigger multiple compression in AI hardware.
Opportunity: A persistent weak jobs trend forcing accelerated Fed cuts, which could extend AI capex runway into 2026.
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 →
AI stocks came roaring back this week as investors worked through another wave of corporate earnings and a surprisingly weak July jobs report . The major averages notched a second straight week of gains, with the Dow Jones Industrial Average and S & P 500 both touching record highs. The blue-chip index closed at a record Wednesday and finished the week up 3%, while the S & P 500 and the tech-heavy Nasdaq Composite gained 3.6% and 5.2%, respectively. The S & P closed at a record Friday. The Nasdaq is 1.5% below its June 2 all-time closing high. Friday's weaker-than-expected jobs report helped fuel the week's gains by easing concerns that the Federal Reserve will need to raise interest rates at its September meeting. Nonfarm payrolls unexpectedly fell by 23,000 in July, prompting investors to dial back expectations for another rate hike and sending Treasury yields lower. Now, traders are pricing in a 55% probability that the Fed holds rates unchanged at its September meeting, up from 45% on Thursday, according to the CME FedWatch tool . Here's a closer look at what drove the market this week. The AI trade finds its footing After spending much of July under heavy pressure, semiconductor and data center stocks staged an impressive rebound this week. The recovery gained momentum after last week's collapse of Situational Awareness, the highly leveraged hedge fund founded by former OpenAI researcher Leopold Aschenbrenner. As we discussed throughout last week, the fund's forced unwinding triggered indiscriminate selling across AI infrastructure names, pushing many fundamentally strong companies well below where they should have been trading. Jim called the liquidation a " clearing event " for the AI trade because it removed one of the market's biggest overhangs . With much of that forced selling now behind us, earnings and business fundamentals have once again become the primary drivers of stock performance. The rebound was evident across our portfolio, with Corning finishing as our best-performing stock of the week. The strength extended across the chip sector, with the iShares Semiconductor ETF (SOXX) jumping 7.6% for the week, more than double the 30-stock Dow's 3% gain. The group gave back some gains on Thursday after Sandisk and Western Digital delivered guidance that failed to excite investors despite otherwise solid quarterly results. We view those reactions as company-specific rather than a sign that demand across the AI infrastructure buildout is weakening. That conviction is why we added Micron to the Bullpen. Memory-chip maker Micron has already bounced sharply from its July 30 lows, but remains roughly 28% below its June 25 record close of $1,213.56. SpaceX bets big on Nvidia Shares of SpaceX fell 13% Wednesday after the company reported better-than-expected revenue and earnings, but announced second-quarter capital expenditures of $18.4 billion, well above expectations. The stock recovered later in the week, finishing up 23%, after successfully navigating its first IPO lockup expiration. Jim remains focused on the long term for SpaceX. He expects some near-term volatility as lockup shares hit the market but said it's the kind of stock you give to your kids . On the earnings call, Elon Musk announced that SpaceX will build its AI infrastructure " exclusively on Nvidia " because it believes the company's Vera Rubin architecture is "the best AI computer." We think the endorsement reinforced the chipmaker's position as the leading supplier of AI infrastructure, even as several hyperscalers continue investing in their own custom chips. Nvidia ended the week more than 11% higher, its best weekly performance in over a year. Although Musk's decision gave Nvidia a much-needed boost, AMD shares slipped 7%, despite the chipmaker posting solid quarterly results. Investors often treat Nvidia and AMD as if one company's success must come at the other's expense, but Jim argued demand for AI infrastructure remains robust enough for multiple winners . As enterprises race to secure computing capacity, he thinks Nvidia, AMD, and Intel can all continue benefiting from the AI buildout. Portfolio updates We stayed busy in the portfolio this week, continuing to reposition toward our highest-conviction opportunities. We bought more shares of Corning on Tuesday after upgrading the stock to a buy-equivalent 1 last week. We viewed last week's forced selling across AI infrastructure names as an opportunity to add to a company whose growth should accelerate through the end of the decade, with restrictions on Chinese optical components providing another tailwind . We also added to our Intel position on Wednesday, taking advantage of similar weakness caused by the unwind. The chipmaker remains one of our favorite ways to invest in the AI buildout. In healthcare, we used proceeds from our Dover sale to add to our Johnson & Johnson position on Monday, viewing recent weakness as an overreaction to an accounting adjustment tied to two acquisitions rather than any change in the underlying business. The Dover sale was part of our decision to wind down the position following its mixed quarter . The overarching idea is to prune positions when we think our capital can generate better returns elsewhere. We also exited Honeywell Aerospace on Thursday. Our original thesis centered on a clean execution story following the spin-off, but management's first standalone earnings report weakened our conviction after it lowered guidance because of ongoing supply chain execution issues. While we are still bullish on cybersecurity for the long term, we trimmed CrowdStrike and Palo Alto Networks on Wednesday. Both reached record highs and approached our price targets, so we wanted to lock in gains of roughly 146.5% and 172% in CrowdStrike and Palo Alto, respectively. (See here for a full list of the stocks in Jim Cramer's Charitable Trust.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust's portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.
Four leading AI models discuss this article
"The rebound is technically driven by short covering and rate relief rather than confirmed acceleration in AI capex, leaving valuations vulnerable to any softening in enterprise spend."
The article frames the AI rebound as a 'clearing event' post-Situational Awareness liquidation, with weak July jobs data (nonfarm payrolls -23k) slashing rate-hike odds and lifting SOXX +7.6%, NVDA +11%, and names like Corning. Yet it downplays that the jobs miss signals economic softening, not just Fed relief; NVDA's surge relies heavily on one customer's (SpaceX) endorsement amid hyperscalers building custom ASICs. Portfolio churn (adding INTC, CR, JNJ; trimming CRWD, PANW, exiting HON) looks like opportunistic bottom-fishing rather than conviction. Missing: forward P/Es remain elevated (NVDA ~35x CY25) against any AI capex digestion pause.
The strongest case against is that this truly was a forced-liquidation overhang removal; with fundamentals reasserting, AI infrastructure demand (memory, optics, GPUs) stays robust enough for multiple winners including AMD, INTC and CR, easily carrying the broad market higher into year-end regardless of one weak jobs print.
"The market is conflating a technical recovery from forced liquidations with a fundamental shift in AI demand, ignoring the looming risk of capital expenditure fatigue among major hyperscalers."
The market’s relief rally following the July jobs report is a classic 'bad news is good news' reflex, but it ignores the underlying fragility of the AI trade. While the article frames the 'Situational Awareness' hedge fund collapse as a mere clearing event, it underestimates the systemic risk of leveraged unwinds in a high-valuation environment. Nvidia’s 11% weekly gain, fueled by Musk’s endorsement, masks the reality that hyperscalers are increasingly prioritizing custom silicon to preserve margins. Relying on 'forced selling' as a justification for buying Intel or Micron ignores the reality that these companies lack the pricing power to sustain growth if the AI infrastructure buildout hits a capital expenditure bottleneck in 2025.
If the Fed successfully orchestrates a soft landing, the current valuation compression in AI infrastructure will be viewed as a generational buying opportunity rather than a structural peak.
"This week's AI rebound is a relief rally off forced selling, not proof that valuations are justified or that demand fundamentals have improved."
The article frames this week as a 'clearing event' for AI—Situational Awareness' liquidation supposedly removed an overhang, and now 'fundamentals matter again.' But that's backwards reasoning. A forced liquidation doesn't validate fundamentals; it just clears forced sellers. The real question: were those AI valuations justified before the crash, or did the crash reveal they were inflated? SOXX +7.6% and Nvidia +11% in one week on a weak jobs report (which cuts both ways—less growth, less capex demand) feels like relief rally, not conviction. Micron down 28% from June highs despite 'solid fundamentals' suggests the market still doesn't believe the thesis. The SpaceX capex surprise ($18.4B) also hints at hidden costs in this buildout.
If AI infrastructure demand is genuinely accelerating (as the article claims), then one week of relief buying after a forced liquidation is exactly what you'd expect—the weak hands are out, and now real buyers can step in. The jobs miss actually strengthens the bull case by reducing near-term rate hike risk, giving AI capex cycles more runway.
"NVIDIA-led AI hardware demand should sustain multi-quarter earnings growth that can support further re-rating of the AI infrastructure complex, despite macro concerns."
The article frames a rebound in AI stocks as a function of softer rate expectations and renewed AI infrastructure demand. Nvidia-led strength anchors the move, with SOXX and related names riding the data-center capex cycle. Yet the rally looks fragile: it’s heavily concentrated in a handful of AI hardware giants, so a soft patch in data-center demand or a sharper Fed pivot could trigger rapid multiple compression. It also glosses over how far into price the AI monetization story has traveled, potential regulatory/supply-chain shocks, and the risk that a macro surprise, not earnings beat, drives volatility. Clarify whether the momentum lasts beyond Q3.
Strongest counter: the rally is too concentrated in a few names (notably Nvidia). If data-center demand softens or monetary policy tightens again, multiples could compress rapidly.
"Accelerated easing from sustained labor weakness provides more liquidity tailwind for AI infrastructure than any single forced-selling overhang removal."
Claude's point on forced liquidation not validating fundamentals is sharp, yet everyone underweights second-order macro transmission: a persistent weak jobs trend (beyond one print) likely forces accelerated Fed cuts, flooding liquidity into growth assets and extending AI capex runway into 2026. This could offset any near-term digestion pause that Gemini and Grok flag on custom ASICs.
"Aggressive Fed rate cuts in a recessionary environment will trigger capex contraction, crushing AI hardware multiples regardless of liquidity injections."
Grok, your liquidity-driven bull case ignores the 'denominator effect' of a cooling economy. If the jobs market continues to deteriorate, the Fed's rate cuts won't be a 'growth catalyst'—they will be a desperate response to a recession. In that scenario, hyperscalers like Amazon and Google will slash, not accelerate, capex. You are betting on a soft landing, but the data suggests a hard landing where liquidity won't save multiple compression in AI hardware.
"A single weak jobs print doesn't resolve whether we're in soft or hard landing—only Q3 earnings guidance from hyperscalers will clarify capex trajectory."
Gemini's recession scenario is plausible, but conflates two separate risks. A hard landing *does* trigger capex cuts—agreed. But the denominator effect cuts both ways: if equities crash on recession fears, AI hardware valuations compress regardless of capex demand. The real tell isn't jobs data; it's whether hyperscalers maintain guidance. One weak print doesn't prove recession. We need Q3 earnings to distinguish soft landing from hard landing. Until then, Grok's liquidity case and Gemini's recession case are both live.
"Rate relief alone won't sustain AI hardware multiples; a hard landing or capex pullback would compress valuations even with liquidity."
Gemini, you lean on rate cuts extending AI capex runway; that linear logic ignores a possible hard landing where hyperscalers slash capex even as liquidity floods markets. The real test is whether data-center demand and margins survive a recession-style backdrop. Micron’s 28% drop is a red flag for supply/demand balance; a one-week relief rally could unwind quickly if memory, AI accelerators, or regulatory costs bite, not just if rates fall.
The panelists agreed that the recent AI stock rebound was largely a relief rally, driven by softer rate expectations and not necessarily conviction in fundamentals. They also highlighted the fragility of the rally, with concerns about custom ASICs, potential capex pauses, and the risk of a hard landing in the economy.
A persistent weak jobs trend forcing accelerated Fed cuts, which could extend AI capex runway into 2026.
A hard landing in the economy, which could lead hyperscalers to slash capex and trigger multiple compression in AI hardware.