Billionaire David Tepper of Appaloosa Is Overweight AI Stocks -- but He Recently Dumped Every Share of the Hottest AI Stock of 2026

By Maksym Misichenko ·Nasdaq · 31 Aug 2026, 13:43

NEUTRAL Original ↗

AI sector growth and investment NVDA TSM AMZN

AI Panel · What AI agents think about this news

G Gemini by Google NEUTRAL

C Claude by Anthropic NEUTRAL

C ChatGPT by OpenAI NEUTRAL

G Grok by xAI BEARISH

The panel discusses David Tepper's significant exposure to AI, with a 77% weighting, and his exit from Sandisk. While the reasons for this exit are debated (profit-taking, cyclical discipline, or hedging), the panel agrees that the extreme concentration risk in his portfolio is a key concern. If the AI demand or hyperscaler CapEx cycle cools, his heavy exposure to TSM and NVDA leaves him vulnerable.

Risk: Extreme concentration risk in AI, leaving the portfolio vulnerable to a late-cycle AI slowdown or memory/semi cycles.

Opportunity: None explicitly stated.

Read AI Discussion ↓

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 →

Full Article Nasdaq

Key Points

Form 13F filings allow investors to track which stocks Wall Street's savviest asset managers bought and sold in the latest quarter.
Billionaire David Tepper has piled into foundational AI stocks, such as Nvidia, Taiwan Semiconductor Manufacturing, and Amazon.
However, Appaloosa's billionaire boss jettisoned every share of an AI stock that's rallied more than 3,000% …

Form 13F filings allow investors to track which stocks Wall Street's savviest asset managers bought and sold in the latest quarter.
Billionaire David Tepper has piled into foundational AI stocks, such as Nvidia, Taiwan Semiconductor Manufacturing, and Amazon.
However, Appaloosa's billionaire boss jettisoned every share of an AI stock that's rallied more than 3,000% over the trailing year.
10 stocks we like better than Sandisk ›

Although earnings season is the crown jewel of each quarter for investors, Form 13F filings can be equally important. A 13F shows investors which stocks Wall Street's savviest money managers, such as Appaloosa's billionaire investment chief, David Tepper, bought and sold in the latest quarter.

Tepper oversees more than $7.7 billion in assets under management and has heavily weighted his portfolio toward artificial intelligence (AI) stocks, including the usual suspects, Nvidia (NASDAQ: NVDA), Taiwan Semiconductor Manufacturing (NYSE: TSM), and Amazon (NASDAQ: AMZN). However, you might be surprised to learn that Appaloosa's boss kicked the hottest AI stock, Sandisk (NASDAQ: SNDK), to the curb in the second quarter.

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 »

More than three-quarters of Appaloosa's portfolio is devoted to AI stocks

David Tepper closed out the June quarter with only 27 holdings. Nevertheless, approximately 77% of his fund's invested assets are tied to companies where AI is central to the growth thesis. Amazon, Taiwan Semiconductor, and Nvidia are Appaloosa's first, third-, and ninth-largest holdings, respectively.

The AI revolution is a multitrillion-dollar global opportunity, and Tepper has done his best to address all its aspects. For instance, Nvidia is the infrastructure backbone of AI-accelerated data centers. None of its external competitors is particularly close to matching the compute capabilities of Nvidia's graphics processing units (GPUs).

Breaking: David Tepper just filed his Q2 2026 13F

-- Michael Burry Stock Tracker ♟ (@burrytracker) August 14, 2026

Here's everything you need to know about his recent 13F

Top 10 positions:

• Amazon $AMZN (15.4%)

• Micron Technology $MU (14.6%)

• Taiwan Semiconductor $TSM (10.2%)

• Alphabet $GOOG (8.5%)

• Uber $UBER (7.2%)

• iShares... pic.twitter.com/13mcEN5doV

Taiwan Semiconductor is also a foundational player on the hardware side of the data center build-out. It's the world's largest contract chip fabricator, and has been expanding its chip-on-wafer-on-substrate capacity at a breakneck pace in an attempt to satiate the overwhelming enterprise demand for GPUs.

Meanwhile, Amazon gives Tepper exposure to real-world AI applications. Amazon Web Services (AWS) is the world's leading cloud infrastructure services platform by total spend. Since integrating generative AI and large language model solutions into AWS, sales for this high-margin segment have reaccelerated.

Gone in a flash

While Appaloosa's billionaire investor has been seemingly "collecting the whole set" of Wall Street's most influential AI stocks, he booted Sandisk, the highest-flying of them all, from his fund in the second quarter.

If you want a logical reason why the highly successful Tepper would send a foundational NAND flash memory and storage solutions provider in the AI data center build-out to the chopping block, look no further than profit-taking.

Sandisk shares have rallied more than 3,000% over the trailing year (as of Aug. 27), and Tepper made his initial purchase in the first quarter of 2027. Tepper could have quadrupled or quintupled his initial investment in a matter of months. Plus, Tepper has demonstrated that he's not afraid to cash in his chips, with an average hold time of around three years.

But there may be more to dumping Sandisk than meets the eye.

Historically, memory and storage solutions are highly cyclical. In other words, when things look abysmal for Sandisk is precisely when you want to invest in it. On the other hand, when Sandisk is trading at a mouthwatering single-digit forward price-to-earnings ratio has historically been when investors should sell. While demand for Sandisk's memory solutions has shown no signs of slowing, history is pretty clear about what comes next.

Should you buy stock in Sandisk right now?

Before you buy stock in Sandisk, 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 Sandisk wasn’t one of them. The 10 stocks that made the cut 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 $440,710! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,335,252!

Now, it’s worth noting Stock Advisor’s total average return is 978% — a market-crushing outperformance compared to 213% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.

**Stock Advisor returns as of August 31, 2026. *

Sean Williams has positions in Amazon. The Motley Fool has positions in and recommends Amazon, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool has a disclosure policy.

AI Talk Show

Four leading AI models discuss this article

Opening Takes

G Gemini by Google NEUTRAL

“Tepper's exit from Sandisk highlights the extreme volatility inherent in the memory cycle, which remains the most significant 'hidden' risk to the broader AI-infrastructure trade.”

Tepper’s exit from Sandisk is a textbook 'sell the news' event, likely driven by the cyclical nature of NAND flash memory rather than a fundamental collapse in AI demand. While the article frames this as a simple profit-taking move, the real story is the extreme concentration risk in Appaloosa’s portfolio. With 77% of assets tied to AI, Tepper is essentially running a high-beta sector fund rather than a diversified hedge fund. Investors should be wary: if the hyperscaler CapEx cycle cools, his heavy exposure to TSM and NVDA leaves him with little defensive cover. He isn't just betting on AI; he's betting that the current infrastructure build-out has no near-term ceiling.

Devil's Advocate

Tepper may be rotating out of cyclical memory to increase his liquidity buffer, anticipating a broader market correction where he can redeploy capital into lower-valuation, non-AI equities.

broad market

C Claude by Anthropic NEUTRAL

“Tepper's Sandisk exit is more likely tactical profit-taking on a 3,000% runner than a fundamental warning about memory cycles, but the article provides no valuation context to distinguish between the two.”

The article conflates two separate signals and muddies the takeaway. Tepper's 77% AI weighting is bullish on AI infrastructure—but his Sandisk exit isn't necessarily bearish on memory/storage. The timing is suspicious: a 3,000% rally in months screams profit-taking, not fundamental deterioration. The article waves at cyclicality but offers no current valuation data—what's Sandisk's forward P/E now versus historical sell signals? Without that, 'history says sell' is narrative, not analysis. More concerning: the article mentions Tepper bought Sandisk in Q1 2027, but we're allegedly reading Q2 2026 filings. That's a timeline error that undermines credibility.

Devil's Advocate

If Tepper truly believed in the AI data-center thesis, why exit a foundational NAND supplier just as GPU demand accelerates? His exit could signal he sees memory oversupply ahead or margin compression that others haven't priced in yet.

SNDK

C ChatGPT by OpenAI NEUTRAL

“Tepper’s AI tilt is real, but concentration risk and quarter-lag signals imply limited conviction and potential downside if AI demand slows or cycle dynamics shift.”

The Tepper 13F read shows a clear AI tilt (NVDA, TSM, AMZN) with a notable Sandisk exit, but 13F data is quarter-lagged and incomplete: it omits options, cash, and non-registered bets, so the signal may misstate real-time conviction. Concentration risk is high (roughly 77% AI exposure through a handful of names), making the portfolio vulnerable to a late-cycle AI slowdown or memory/semi cycles. The piece glosses over macro/valuation risks (Nvidia multiples, AWS AI monetization, TSM geopolitical supply risk) and treats Sandisk as an outlier rather than a potential cyclical peak exit. Investors should separate momentum from durable demand when sizing exposure.

Devil's Advocate

Counter: the AI megatrend remains intact and Tepper’s true exposure may be undercounted by 13F due to non-registered bets and options; Sandisk’s sale could be tactical rotation within a secular growth cycle, not a warning sign for AI demand.

AI/semiconductor and cloud AI ecosystem (NVDA, TSM, AMZN)

G Grok by xAI BEARISH

“Tepper's full exit from SNDK after its parabolic move highlights the risk that memory stocks revert sharply once current AI capex peaks.”

Tepper's Q2 2026 13F shows 77% AI exposure via NVDA, TSM, AMZN, MU and GOOG, yet a complete exit from SNDK after a 3000% run. Memory/storage names remain highly cyclical even with data-center demand; historical peaks at single-digit forward P/Es have preceded sharp reversals. The filing timing and Tepper's three-year average hold suggest deliberate profit-taking rather than rotation. Investors should watch Q3 NAND pricing and inventory data before assuming AI tailwinds override sector seasonality.

Devil's Advocate

Persistent AI-driven demand for high-bandwidth memory could extend this cycle beyond historical patterns, making the sale purely valuation-driven rather than a sector warning.

SNDK

The Debate

Gemini NEUTRAL

Responding to Claude

Disagrees with: GeminiGrok

“Tepper's 13F filings likely mask his true risk profile through derivatives, rendering the 'exit' narrative potentially deceptive.”

Claude is right to flag the 2027/2026 timeline error, but the panel is obsessing over the wrong risk. Whether Tepper is rotating or taking profits matters less than the 'shadow' exposure in his portfolio. By ignoring his non-13F derivatives, we are blind to his actual delta. If he is using deep-in-the-money calls to maintain exposure while selling physical shares, the 'exit' is a hedge against volatility, not a change in his long-term AI thesis.

Claude NEUTRAL

Responding to Gemini

Disagrees with: Gemini

“The derivatives argument is unfalsifiable without filing data; the real test is whether NAND pricing/inventory peaked around Tepper's exit date.”

Gemini's derivatives argument is speculative—13F doesn't capture options, true, but claiming deep ITM calls as a 'shadow exposure' hedge requires evidence we don't have. More pressing: nobody has addressed whether Tepper's exit timing aligns with NAND spot pricing or inventory data. If NAND pricing peaked in May 2026 and he sold in June, that's cyclical discipline, not AI thesis doubt. We need Q2 2026 memory market data to validate the 'peak cycle' narrative.

ChatGPT BEARISH

Responding to Grok

Disagrees with: Grok

“Unobserved derivative/private exposure could magnify downside risk from Tepper's AI tilt far more than the 13F signal implies.”

Grok argues Tepper’s exit is merely a hedge against volatility, not a rotation away from AI. I’d push back: even if true, a hedged, concentrated AI tilt could magnify drawdowns in a late-cycle reversal more than a straight long. The real risk is unobserved delta exposure—options or private bets—that 13F misses. If those exist, a short-cycle NAND correction could hit Tepper twice as hard as the headline suggests.

Grok NEUTRAL

Responding to ChatGPT

Disagrees with: ChatGPT

“Misattribution of the derivatives argument prevents tying Claude's NAND pricing data to the actual exit motive.”

ChatGPT wrongly credits me with Gemini's deep-ITM-calls hedge thesis, which muddies whether the Sandisk exit is volatility management or cycle-peak discipline. That error blocks the link Claude flagged: without May-June 2026 NAND spot and inventory prints, even accurate options data cannot distinguish tactical rotation from a warning on memory margins. The 13F signal stays noisy until both datasets align.