I’ve been watching the big money moves for years, and every so often a filing lands that makes you sit up a little straighter. This one did. When the latest 13F numbers for Appaloosa Management hit the tape, it became clear that David Tepper had done something decisive. He didn’t just tinker at the edges. He took real chips off the table in the hottest corner of the AI hardware trade and quietly built larger positions in the names almost everyone already knows by heart.
What Tepper Actually Did With The Portfolio
The second quarter of 2026 will be remembered as the period when Appaloosa leaned harder into the so-called Magnificent Seven while dialing back exposure to memory names that had delivered eye-watering gains. Amazon remained the largest holding and grew even larger. Meta Platforms saw a substantial increase. Alphabet received a meaningful top-up. And for the first time in a while, a sizable new position in Apple appeared. At the same time, the fund cut its Micron stake by more than 40 percent and completely exited a large Sandisk position that had been worth over $400 million at the previous filing.
That combination tells a story. Tepper has never been shy about riding powerful themes. He was early and aggressive on the AI hardware wave. Yet the latest moves suggest he decided the risk-reward in pure memory plays had shifted enough to take some money off the table and redeploy it into the broader platform companies that sit at the center of the same technological shift.
Amazon Stays At The Top Of The Heap
Amazon was already Appaloosa’s biggest position heading into the quarter. The fund increased that stake by more than 15 percent, pushing the market value close to $1.2 billion. In a portfolio of this size, that is not a casual adjustment. It is a statement that the combination of cloud growth, advertising strength, and the longer-term AI infrastructure story still looks attractive even after years of gains.
I’ve always found it interesting how certain investors treat Amazon as both a growth engine and a kind of ballast. The business is so diversified that it can absorb periods of weaker consumer spending while the cloud and AI-related segments keep compounding. Raising the position at this stage of the cycle suggests Tepper still sees meaningful runway left.
Meta Platforms Gets A Heavy Add
The increase in Meta was even more aggressive in percentage terms. Appaloosa expanded the position by nearly 55 percent, bringing the value to roughly $380 million. That is a sizable bet on a company that has already been one of the strongest performers in the group.
Meta’s story has shifted over the past couple of years from pure social media to a more balanced mix of advertising recovery and AI-driven product improvements. The efficiency gains from earlier cost discipline gave the company more flexibility to invest in the technology that now powers better ad targeting and new consumer experiences. Adding at this size after such a strong run is not the move of an investor who thinks the easy money is already gone.
Alphabet And A Fresh Apple Position
Alphabet received a more measured but still meaningful increase of nearly 7 percent. The stake now sits above $661 million. Google’s cloud business and the broader AI research efforts continue to make the company a core holding for many large funds. The slight top-up fits the pattern of reinforcing positions in the companies that own both distribution and infrastructure.
The new Apple position stands out more. Appaloosa established a stake valued at more than $241 million during the quarter. Apple has sometimes been treated as the more defensive member of the group because of its hardware cycle and services mix. Building a position of that size suggests the fund sees the combination of installed base, services growth, and gradual AI feature integration as still underappreciated by the market at current levels.
The Memory Stock Exit That Caught Attention
The other side of the ledger is where the real drama sits. Micron Technology had been a standout performer. Shares more than doubled during the second quarter alone, delivering the strongest three-month gain in the company’s history as investors poured money into anything tied to AI server demand. Appaloosa cut the position by more than 41 percent. Even after the reduction, Micron remained the second-largest holding at around $1.125 billion. That tells you how large the original stake had become.
The timing is worth noting. Micron shares peaked and then pulled back nearly 16 percent from the start of July. The broader memory group experienced a similar cooling after the parabolic move. Tepper appears to have reduced exposure before the sharpest part of that pullback, which is the kind of move that separates experienced operators from momentum chasers.
Sandisk tells a cleaner story. The entire position, previously valued at more than $400 million, was sold. The stock had soared about 258 percent in the March-to-June period only to give back nearly 28 percent in the third quarter so far. Completely zeroing out a winner of that magnitude is rarely easy, but it removes the risk of giving back a large portion of the gains if the memory cycle softens further.
Taking profits in names that have already delivered multi-bagger returns while redeploying into higher-quality platforms is the kind of discipline that compounds over decades rather than quarters.
Why The Memory Trade Cooled
Memory stocks are famously cyclical. Even when the long-term demand story from AI training and inference looks intact, the short-term supply response and pricing dynamics can create violent swings. After such an extreme upward move in a single quarter, some digestion was almost inevitable. Analysts still largely remain constructive on Micron, with the average target implying substantial upside from current levels. Yet the near-term volatility has been real, and funds that rode the wave hard had logical reasons to lock in gains.
I’ve watched similar cycles play out before. The companies that supply the critical components often experience the sharpest moves both up and down. The platform companies that use those components tend to show more durable earnings power. Shifting weight from the former toward the latter is a classic way to stay involved in a powerful theme while reducing the amplitude of portfolio swings.
New Positions Outside Pure Tech
The filing also revealed two fresh positions that sit outside the usual technology concentration. Appaloosa established stakes in Boeing valued at about $173 million and American Airlines at roughly $136 million. These are not trivial sizes. After years of focusing heavily on the AI and megacap tech complex, adding exposure to commercial aerospace and airlines marks a noticeable diversification step.
Boeing has been working through a long recovery process on both the commercial and defense sides. American Airlines, like most carriers, remains sensitive to fuel costs, labor negotiations, and travel demand. Building these positions suggests the fund sees a potential inflection that is not yet fully reflected in valuations. Whether that thesis plays out will depend on execution and the broader economic backdrop, but the willingness to step outside the familiar tech names is itself interesting.
What Was Sold Beyond Memory
On the sell side, the fund also exited or reduced positions in Corning, L3Harris, and RTX. These moves further tilt the portfolio away from certain industrial and defense-related names and toward the megacap technology core plus the new airline and aerospace exposure. Portfolio construction is never static. Each quarter’s changes reflect an evolving view of relative opportunity and risk.
Reading The Bigger Picture
Stepping back, the pattern feels deliberate. Appaloosa stayed heavily involved in the AI theme but chose to express that exposure more through the companies that own the customer relationships, the cloud capacity, and the software layers rather than solely through the pure-play hardware suppliers. At the same time, the fund introduced two new ideas in sectors that have lagged the tech rally for years.
This is not the behavior of a fund that has abandoned growth. It looks more like a fund that is trying to own higher-quality growth while pruning the names most exposed to cyclical pricing pressure. The memory stocks delivered extraordinary returns in a short window. Taking some of those chips off the table and putting them to work in Amazon, Meta, Alphabet, and Apple is a way of staying long the structural trend without accepting the same level of near-term volatility.
Lessons For Individual Investors
Most of us will never manage a multi-billion-dollar portfolio, yet the principles still apply. When a single theme becomes crowded and valuations stretch, it often makes sense to shift some exposure toward the higher-quality companies that benefit from the same trend but with stronger balance sheets and more durable competitive positions. That does not mean abandoning the original idea. It means expressing it in a way that can survive the inevitable pullbacks.
Position sizing matters too. Even after the Micron cut, the remaining stake was still large enough that the fund stayed meaningfully exposed. Complete exits are rare when the long-term case remains intact. Partial profit-taking combined with selective new ideas is usually a more sustainable approach than all-or-nothing timing.
- Identify the durable beneficiaries of a powerful theme rather than only the purest plays
- Take partial profits after extreme short-term moves instead of waiting for the perfect exit
- Use periods of concentration to selectively diversify into under-owned sectors
- Keep position sizes large enough to matter but not so large that a single name dominates risk
The Road Ahead For These Names
Micron and the broader memory group still face a constructive long-term demand backdrop from AI infrastructure buildouts. Near-term pricing and inventory dynamics will determine how quickly the recent pullback reverses. The platform companies Appaloosa added to face their own challenges around competition, regulation, and capital intensity, yet their scale and installed bases give them more ways to monetize the same technological wave.
Boeing and the airlines introduce a different set of variables. Execution risk remains elevated in aerospace, and consumer travel demand can shift with economic conditions. These positions will need to prove themselves over coming quarters. Their presence in the portfolio, however, shows a willingness to look beyond the consensus growth trade of the moment.
Final Thoughts On The Filing
Filings like this one offer a rare window into how sophisticated capital is actually positioned rather than what the headlines claim. Tepper has built a reputation for making concentrated bets when conviction is high and for adjusting course when the risk-reward changes. The second-quarter moves fit that pattern. Heavy reinforcement of the Magnificent Seven core, meaningful reduction in the most extended memory names, and selective new ideas outside technology paint a picture of a fund that remains growth-oriented but is no longer willing to accept the same degree of cyclical risk in the purest AI hardware expressions.
Whether these changes prove correct in hindsight will depend on how the next phase of the AI investment cycle unfolds and whether the newer positions deliver the expected recovery. What is already clear is that Appaloosa refused to sit still. In markets that reward adaptability as much as foresight, that willingness to rebalance remains one of the more valuable traits an investor can maintain.
The next set of filings will show whether these adjustments were the start of a broader rotation or simply a mid-cycle pruning. For now, the message from one of the more closely watched hedge funds is straightforward. The AI story is far from over, but the preferred way of owning it has shifted toward the platforms with the deepest competitive moats and away from the purest cyclical expressions of the same demand.
That kind of nuance is easy to miss when the market is racing higher. It becomes visible only when you slow down and read the actual position changes rather than the narratives built around them. In that sense, this filing does more than update a portfolio snapshot. It offers a reminder that even the strongest themes eventually require thoughtful repositioning if the goal is to keep compounding rather than simply riding every wave until it breaks.
