Every quarter, investors look at 13F filings to see what the world’s best-known money managers bought and sold. The filings are delayed and incomplete, but they still matter. They reveal how long-term investors think when the market is noisy.

The latest quarter was not a calm period. Oil prices were volatile, interest rates remained uncomfortable, AI spending kept rising, and investors were still debating whether the market was expensive. In that environment, the most interesting point is not that famous investors bought stocks. They always do.

The more important point is where they chose to concentrate. Berkshire Hathaway leaned further into Alphabet and added Delta Air Lines. Stanley Druckenmiller’s Duquesne Family Office showed interest in diversified AI semiconductor demand and Argentina’s energy story. Bill Ackman moved into Microsoft after weakness in the stock. Li Lu’s Himalaya Capital bought businesses tied to capital-market infrastructure and Chinese digital consumption.

These are very different trades. But they share one message: the great investors were not simply buying the market. They were choosing specific trends they believed could survive volatility.

First, a warning about 13F filings

A 13F filing is useful, but it is not a live trading screen. It shows U.S.-listed long equity positions at the end of the quarter. It usually does not show cash, most foreign-listed holdings, short positions, or derivatives in a complete way.

That means investors should not copy these filings blindly. By the time the public sees the filing, the manager may already have changed the position. The filing also does not explain the exact reason for each trade.

Still, 13F filings can show direction. They can show which ideas are becoming larger, which positions are being reduced, and whether a manager is concentrating or spreading risk. For investors like Berkshire, Ackman, Druckenmiller, and Li Lu, that direction is often more important than the exact purchase price.

A 13F filing should not be treated as a shopping list. It should be read as a map of conviction.

Berkshire is still being read through the Buffett lens

Berkshire Hathaway’s latest filing attracted unusual attention because the company is now operating in the post-Buffett CEO era, with Greg Abel leading the company. Warren Buffett remains chairman and continues to be closely associated with Berkshire’s investment culture, but stock-level decisions are no longer automatically easy to attribute to one person.

That distinction matters. It is safer to say “Berkshire bought” rather than “Buffett bought.” But the portfolio is still interpreted through the Buffett framework because Berkshire’s culture was built around patience, balance-sheet strength, durable cash generation, and valuation discipline.

The filing showed three major messages. Berkshire increased its Alphabet position sharply. It initiated a new Delta Air Lines position. It reduced Chevron and exited several names, including Amazon, Visa, Mastercard, and UnitedHealth.

This is not the behavior of a firm trying to chase every rally. Berkshire remained selective. The overall portfolio became more focused, while new capital was directed toward businesses where the economics may look different from the market’s first impression.

Alphabet looks like an AI infrastructure bet, not only a search bet

Berkshire’s larger Alphabet position is the most symbolic move. For years, many investors thought of Alphabet mainly as a search advertising company. That is still true, but the AI era changes the way Alphabet’s balance sheet should be read.

Alphabet owns enormous infrastructure. Data centers, servers, chips, cloud systems, AI research, and distribution through Search, YouTube, Android, Gmail, and Workspace all sit inside the same corporate structure. In a market where AI models require huge computing power, that asset base becomes more valuable.

This may explain why Alphabet can fit a value-investor framework even though it is a technology company. The company is not only promising future innovation. It already has cash flow, scale, distribution, and infrastructure.

That makes it different from a speculative AI startup. Alphabet is spending aggressively on AI, but it is doing so from a position of enormous profitability. If its return on invested capital remains strong, heavy capital spending can be interpreted not as waste, but as reinvestment into a market where infrastructure may become the moat.

Alphabet is no longer just a search company in the market’s eyes. It is becoming an AI infrastructure owner with a cash machine attached.

Apple was reduced because the old certainty has become less obvious

Apple remains Berkshire’s largest disclosed holding, but the position is far smaller than it was at its peak. That shift should not be read as a simple rejection of Apple. It is more likely a recognition that Apple’s old investment case has become less one-sided.

Apple is still a world-class business. It has brand power, ecosystem lock-in, services revenue, huge cash generation, and a loyal customer base. But its growth profile is more mature, and the AI transition has created new questions.

If the next decade of technology value is increasingly tied to AI infrastructure, cloud compute, data centers, and model distribution, Alphabet may offer a more direct route to that trend. Apple still has the consumer interface. Alphabet has the AI infrastructure and advertising data engine.

So the move from Apple toward Alphabet can be understood as a shift from a device-and-services compounder toward an AI infrastructure compounder. It is not anti-Apple. It is a change in where the next marginal dollar looks more attractive.

Delta is the surprising Berkshire-style idea

Delta Air Lines was the more surprising Berkshire move. Airlines have historically been poor businesses. They require large capital investment, depend on fuel prices, face labor pressure, suffer from economic cycles, and can destroy capital during downturns.

So why would Berkshire buy Delta? The answer is that Delta is trying to become less like a traditional airline and more like a premium travel, loyalty, and financial-partnership platform.

Delta’s recent results show why this matters. The company has emphasized premium cabins, loyalty revenue, corporate travel, maintenance revenue, cargo, and its American Express partnership. In the March quarter of 2026, Delta said diversified, high-margin revenue streams represented 62% of total revenue, premium revenue grew 14%, loyalty and related revenue grew 13%, and American Express remuneration exceeded $2 billion.

This changes the business mix. A traditional airline tries to fill as many seats as possible. Delta is increasingly trying to sell higher-value seats, monetize loyalty, and capture spending from affluent and corporate customers.

The old airline model sold seats. Delta’s new model sells premium access, loyalty, and customer relationships.

This does not remove airline risk. Fuel, recession, labor costs, aircraft commitments, and travel demand still matter. But it makes Delta more interesting than the old airline stereotype. If the company can keep generating strong free cash flow and reduce balance-sheet risk, the investment case becomes less about cheap airfare and more about cash conversion from a premium travel ecosystem.

Chevron was reduced because valuation matters even in a strong oil market

Berkshire’s Chevron reduction is also instructive. Oil prices were supported by geopolitical risk, and energy companies benefited from the market’s renewed focus on supply security. But that does not automatically make every energy stock attractive.

Energy is cyclical. When oil prices rise sharply, earnings and cash flow can look very strong. But investors have to ask whether that strength is sustainable or already priced in.

In commodity businesses, valuation can become dangerous at the top of the cycle. If the stock price reflects peak cash flow, the downside can be large when oil normalizes. Reducing Chevron in that environment fits a disciplined approach: even a good business can become less attractive when the market prices it too generously.

This is one of the clearest Berkshire lessons. The firm is not simply buying “themes.” It is still asking whether the price gives enough future return.

Druckenmiller is looking beyond the first wave of AI chips

Stanley Druckenmiller’s Duquesne Family Office did not appear to make a giant all-in macro bet during the quarter. But some holdings point to a more nuanced AI view.

The key idea is that AI chip demand may broaden. In the first phase, Nvidia dominated the story because GPUs were the center of training and inference capacity. But as AI moves from data centers into networks, power systems, custom chips, vehicles, robots, and industrial equipment, the semiconductor opportunity becomes wider.

Broadcom fits the custom-chip and networking side of that trend. Large cloud customers increasingly want application-specific chips, networking solutions, and infrastructure optimized for AI workloads. Broadcom is one of the major companies positioned for that shift.

STMicroelectronics fits a different layer. It is exposed to power semiconductors, microcontrollers, sensors, automotive systems, and industrial electronics. If AI moves into the physical world through robots, machines, vehicles, and smart devices, the need for control chips and power management expands.

The first AI trade was about GPUs. The next AI trade may be about power, sensors, custom silicon, and physical-world control.

YPF is a bet on Argentina’s energy reset

Druckenmiller’s interest in YPF points to another theme: energy supply outside the traditional Middle East framework.

YPF is Argentina’s largest integrated energy company. The strategic asset behind the story is Vaca Muerta, one of the world’s most important shale oil and gas formations. If Argentina can develop that resource efficiently, it could become a more important energy exporter.

The investment case depends on two things. First, production must rise while costs fall. That is the classic shale learning-curve story: as operators drill more, improve techniques, and build infrastructure, productivity can improve.

Second, Argentina’s political and economic risk must decline. President Javier Milei’s market-oriented reforms are central to that hope. If Argentina becomes more investment-friendly, YPF could benefit from both resource development and country-risk re-rating.

But this remains a higher-risk idea. Argentina has a long history of inflation, default, capital controls, and policy instability. YPF is not only an energy bet. It is also a bet that Argentina’s reform path can hold.

Ackman’s Microsoft move follows his familiar playbook

Bill Ackman’s Pershing Square initiated a major Microsoft position after the stock weakened. This fits Ackman’s recent style: buy high-quality dominant platforms when the market becomes disappointed with near-term execution.

Microsoft’s problem was not that the business suddenly became weak. Revenue and earnings remained strong. The market’s concern was more specific: Azure growth did not fully satisfy expectations, AI capital spending was rising, and investors questioned whether Copilot and the OpenAI partnership would generate enough return.

Ackman’s view appears to be that the market punished Microsoft too harshly. The company still controls enterprise software distribution through Windows, Office, Azure, Teams, GitHub, LinkedIn, security tools, and cloud infrastructure. That gives Microsoft one of the strongest positions in enterprise AI.

The key question is whether Microsoft can turn AI spending into durable revenue. If Copilot, Azure AI, GitHub Copilot, and enterprise automation deepen customer lock-in, today’s capital spending may become tomorrow’s moat. If not, the market will continue asking whether Microsoft is spending too much to chase AI hype.

Ackman is not buying Microsoft because the market is excited. He is buying because the market became disappointed with a company that still owns the enterprise software layer.

Li Lu is buying the pipes of capitalism

Li Lu’s Himalaya Capital made several notable moves. The most interesting are S&P Global, Moody’s, and MSCI. These are not flashy AI companies. They are infrastructure companies for capital markets.

S&P Global and MSCI benefit from the rise of index investing, ETFs, benchmarks, data products, analytics, and institutional portfolio construction. When passive investing grows, index providers become more important. Their products become part of the operating system of global finance.

Moody’s benefits from credit markets. As companies issue bonds, refinance debt, structure financing, or raise capital for AI infrastructure, ratings and credit analytics remain essential. Higher debt complexity increases the need for trusted credit evaluation.

These businesses share several attractive traits. They are asset-light, highly profitable, data-rich, and deeply embedded in financial workflows. Their pricing power can be strong because customers depend on their standards, ratings, data, and benchmarks.

In simple terms, Li Lu is not only buying companies. He is buying toll roads inside global capitalism.

Tencent Music is a bet on Chinese digital consumption

Li Lu also added Tencent Music, which points to a different theme: Chinese consumer spending moving into digital entertainment and online culture.

Tencent Music is not only a music-streaming platform. It also has social entertainment, online audio, artist services, fan communities, and digital content monetization. If China gradually shifts toward a more consumption-led economy, platforms that control digital leisure time could benefit.

This is not a risk-free idea. China’s consumer recovery has been uneven, regulation remains a factor, and youth unemployment and household confidence still matter. But the investment logic is clear: as middle-class consumers spend more on digital services, entertainment platforms can capture recurring revenue.

Li Lu has often been associated with patient, concentrated investing. Buying Tencent Music suggests he may be willing to look beyond short-term pessimism about China if he sees long-term consumer behavior shifting in favor of digital platforms.

Crocs shows Li Lu’s classic value angle

Crocs is a more traditional value setup. The stock has been under pressure since the post-pandemic boom faded and the HeyDude acquisition disappointed investors. The acquisition has been criticized because HeyDude’s performance failed to justify the price paid.

But Crocs still has attractive underlying economics. The core Crocs brand has high margins, strong cash generation, global recognition, and a simple product model. The company has also used cash flow for share repurchases.

The key question is whether HeyDude can stabilize. If the acquisition remains a drag, Crocs may stay cheap for a reason. If management can repair the brand and the core Crocs franchise keeps producing cash, the market may eventually re-rate the stock.

This is the kind of situation value investors often like: a damaged story, but not necessarily a damaged core business.

Crocs is not a perfect story. That is exactly why a value investor may find it interesting.

The common message: they did not buy the market; they bought structure

These portfolios are very different, but they have a common theme. The great investors did not simply say “stocks fell, so buy everything.” They selected structures.

Berkshire bought into AI infrastructure through Alphabet and a transformed premium travel model through Delta. Druckenmiller looked at semiconductor demand beyond GPUs and energy development outside the Middle East. Ackman bought Microsoft as a high-quality platform temporarily punished for AI spending anxiety. Li Lu bought capital-market infrastructure and Chinese digital consumption.

This is the key lesson. In a volatile market, broad timing becomes hard. But structural trends remain investable if the price is right.

AI infrastructure, premium consumer segmentation, capital-market plumbing, energy diversification, and digital consumption are all long-term themes. The difference among great investors is which theme they trust most and what price they are willing to pay.

What ordinary investors should take from this

The wrong lesson is to copy the stocks. The right lesson is to copy the process.

First, they concentrate only when they understand the business. Berkshire did not buy every AI stock. It bought a cash-generating infrastructure platform.

Second, they look for business-model change. Delta is not attractive because airlines are suddenly wonderful. It is attractive only if its high-margin revenue mix has changed the economics.

Third, they buy temporary disappointment when the long-term franchise remains strong. That is Ackman’s Microsoft logic.

Fourth, they look for toll roads. Li Lu’s S&P Global, Moody’s, and MSCI purchases show interest in businesses that earn money from the growth and complexity of markets themselves.

Fifth, they separate trend from price. Chevron may benefit from higher oil, but if valuation already reflects too much optimism, disciplined investors reduce exposure.

Great investors do not ask only, “What is the theme?” They ask, “What business captures the theme, and what price am I paying?”

Conclusion: conviction became narrower

The latest 13F filings do not show panic. They also do not show blind optimism. They show selective conviction.

Berkshire is willing to own more Alphabet and start a meaningful Delta position, but it is also cutting Chevron and exiting several holdings. Druckenmiller is not making one giant market call, but he is positioning around AI chip diversification and Argentina’s energy upside. Ackman is buying Microsoft when the market questions AI spending. Li Lu is buying financial infrastructure, Chinese digital consumption, and a beaten-down consumer brand with cash-flow potential.

That is the important market signal. The best investors are not treating this as a simple bull market or bear market. They are treating it as a market where only certain structures deserve capital.

The simplest way to read the latest 13F season is this: the masters are not buying because stocks went up or down. They are buying where business quality, structural change, and valuation finally meet.