Alphabet Is Now Berkshire’s Third-Largest Stock Holding, Marking a Sharp Break From Its Tech Caution

Alphabet became Berkshire Hathaway’s third-largest disclosed stock holding after the conglomerate added about 48.1 million shares during the second quarter of 2026. The position’s rapid ascent matters because Berkshire spent decades treating most technology companies as too unpredictable for a concentrated investment.

The purchase was not a minor portfolio adjustment. Berkshire increased its Class A position by 45.2% and expanded its smaller Class C position more than sevenfold between March 31 and June 30. Combined ownership reached almost 106 million shares at quarter-end.

That scale places Google’s parent behind only Apple and American Express in Berkshire’s reported U.S. stock portfolio. It also pushes the company ahead of established Berkshire holdings such as Coca-Cola, Bank of America, Chevron, and Occidental Petroleum.

The main tension is not simply technology versus traditional value investing. Berkshire is betting that Google’s advertising cash flows, cloud operation, and AI infrastructure can remain economically durable despite extraordinary capital requirements. That judgment now sits near the center of Greg Abel’s first year as Berkshire’s chief executive.

Berkshire’s Q2 Alphabet Purchase Changed the Portfolio Hierarchy

The defining Q2 move was Berkshire’s decision to keep buying after its existing stake had already grown dramatically.

Berkshire’s Q2 holdings filing, submitted on August 14, covers positions held on June 30. It reports 78,791,167 Class A shares and 27,188,433 Class C shares.

Three months earlier, Berkshire reported 54,249,798 Class A shares and 3,585,215 Class C shares. The conglomerate therefore added 24,541,369 Class A shares and 23,603,218 Class C shares during Q2.

That was an increase of approximately 48.1 million shares across both classes. It lifted Berkshire’s combined position to 105,979,600 shares, compared with about 57.8 million at the end of March.

The quarter-end market value of the two share classes was approximately $37.8 billion. That made the combined stake Berkshire’s third-largest disclosed equity investment, behind Apple and American Express.

The ranking requires one important qualification. A Form 13F reports certain U.S.-listed securities held by an institutional investment manager. It does not represent Berkshire’s entire balance sheet, operating-business portfolio, cash position, foreign securities, or every possible investment exposure.

The filing is also a quarter-end snapshot. It shows what Berkshire reported owning on June 30, not what it owns today. Trades made after that date will remain undisclosed until a later filing or another applicable disclosure.

Even with those limits, the direction is unmistakable. The conglomerate did not merely retain shares obtained through a large June transaction. It also accumulated additional shares, according to the change between its March and June holdings.

That distinction matters because Berkshire had already committed substantial capital to Google’s parent before the quarter closed. The later filing showed that the commitment was broader than one negotiated transaction.

Berkshire’s Q1 filing provides the comparison point. At March 31, its reported Class A position was valued at roughly $15.6 billion, while its Class C holding was valued near $1 billion.

The March total was already more than triple Berkshire’s position at the end of 2025. Q2 then added another major layer, turning a new investment into one of the conglomerate’s defining public-stock commitments.

Other changes reinforce the concentration of that decision. Berkshire added to Delta Air Lines, Lennar, Macy’s, and The New York Times during Q2. None approached the scale of the additional Google-parent purchases.

At the same time, Berkshire reduced Bank of America, Capital One, Kroger, Nucor, DaVita, and Ally Financial. It exited Constellation Brands entirely. The portfolio was not simply expanding across the board.

Berkshire was reallocating capital. It sold or reduced several financial, consumer, health-care, and industrial positions while placing far more weight behind one large technology platform.

That creates the article’s central reversal. A company famous for avoiding difficult-to-forecast technology businesses has made one of the world’s largest technology companies a top-three reported holding.

Why Google’s Parent Now Fits Berkshire’s Value Framework

Berkshire’s move suggests it sees a durable economic system, not merely a fast-growing technology stock.

Warren Buffett frequently explained that Berkshire avoided businesses whose competitive positions could not be forecast with reasonable confidence. Rapid technical change often made software, hardware, and internet companies difficult to place inside that framework.

Apple eventually became the major exception. Berkshire treated the iPhone maker less like a speculative hardware company and more like a consumer franchise with high customer loyalty, pricing influence, and recurring economic relationships.

The latest investment extends that logic into a more complicated business. Google’s parent combines search advertising, YouTube, cloud computing, mobile software, productivity tools, AI models, custom processors, and global infrastructure.

Search remains especially important to the investment case. It connects billions of user queries with advertisers seeking measurable commercial intent. That feedback loop has historically supported high margins and immense cash generation.

YouTube adds another distribution system spanning advertising, subscriptions, creators, television viewing, and music. Android, Chrome, Maps, Gmail, and Workspace give the company additional routes into consumer and enterprise activity.

Google Cloud provides the clearest link between that established franchise and the current AI investment cycle. Customers can rent computing capacity, use managed data services, deploy models, and build applications without operating every infrastructure layer themselves.

This breadth makes the company look less like a single-product technology vendor. It resembles a collection of digital infrastructure and service franchises whose products reinforce one another.

Berkshire does not need to predict every model benchmark to find that structure attractive. Its thesis can rest on more durable questions about distribution, customer relationships, infrastructure economics, and long-term cash generation.

The timing also points to a valuation judgment. Berkshire enlarged the position while investors were debating whether extraordinary AI spending would produce adequate returns. That debate created a conflict between current business strength and future capital intensity.

The purchase therefore looks different from chasing a new product announcement. Berkshire was committing capital while the market questioned how much infrastructure spending the company could absorb.

The scale of the position matters here. Berkshire can hold a small investment without making it central to the portfolio. Moving one company into third place signals a much stronger assessment of its economic durability.

It also solves a practical problem for Berkshire. The conglomerate needs opportunities large enough to absorb billions of dollars without forcing it into thinly traded securities or complete acquisitions.

Few companies meet that requirement. A candidate must offer enormous market liquidity, a defensible business, capable management, and enough potential return to matter against Berkshire’s overall size.

Google’s parent qualifies on scale alone, but scale is not sufficient. Berkshire’s willingness to keep buying indicates that the investment team viewed the underlying economics as understandable enough for a concentrated position.

The purchase still does not reveal a precise internal valuation. Berkshire does not normally explain quarterly stock trades, and a 13F does not identify which investment manager initiated each decision.

That uncertainty should prevent observers from calling the purchase a blanket endorsement of every AI project. It is better understood as a judgment about the entire company, including the advertising engine financing its infrastructure expansion.

Greg Abel Is Breaking With Berkshire’s Old Technology Caution

The investment is becoming an early test of how Berkshire allocates capital after Warren Buffett’s transition from chief executive.

Greg Abel became Berkshire’s CEO at the beginning of 2026, while Buffett remained chairman. Abel inherited a company with an enormous cash reserve, a concentrated stock portfolio, and unusually high expectations surrounding every major capital decision.

The timing makes the position impossible to separate from Berkshire’s leadership transition. The company entered the investment before Abel took the top job, then expanded it sharply during his first two quarters as CEO.

An earlier portfolio disclosure showed that Berkshire had more than tripled its stake by March 31. That same filing revealed new investments in Delta Air Lines and Macy’s, alongside exits from several long-held or technology-related names.

Q2 made the shift harder to dismiss as portfolio cleanup following a management change. Berkshire continued buying one technology platform while trimming numerous other holdings.

The company also placed capital directly with Google’s parent during a major fundraising process. Alphabet announced an expanded $84.75 billion equity financing on June 3, after initially proposing a smaller transaction.

Berkshire agreed to purchase $10 billion of Class A and Class C shares through a private placement. The equity financing was intended to support general corporate purposes, including AI infrastructure and global computing capacity.

That transaction changed Berkshire’s role. It was not only buying shares from other investors in the public market. It supplied capital directly to a company pursuing one of the largest infrastructure programs in the technology sector.

The direct investment also connects Berkshire to the hardest question facing large AI platforms. Building data centers, networking systems, custom processors, and energy capacity requires enormous spending before the full economic return becomes visible.

Berkshire has long accepted heavy capital requirements in railroads, utilities, energy, manufacturing, and insurance. Those businesses differ from cloud computing, but they share a need for long-duration planning and disciplined investment.

Abel spent much of his Berkshire career overseeing the conglomerate’s energy operations. That background may make infrastructure-intensive technology easier to evaluate than a consumer application whose popularity can disappear quickly.

Still, the purchase should not be presented as Abel rejecting Buffett’s principles. It more likely reflects a different conclusion reached through the same basic process.

Berkshire has always sought companies with durable advantages, trustworthy management, understandable economics, and opportunities to reinvest capital. The dispute concerns whether a leading AI platform now satisfies those conditions.

The comparison with Apple is revealing. Berkshire’s Apple stake became proof that the company could own technology when brand loyalty and ecosystem economics were sufficiently visible.

The new position goes further because the AI market is more capital-intensive and less settled. Google faces direct challenges from Microsoft, Amazon, Meta, Anthropic, OpenAI, and many smaller model developers.

Search behavior is also changing. Users increasingly ask conversational systems for synthesized answers, potentially reducing the number of traditional result pages and advertising interactions.

Berkshire is therefore betting on a company while its strongest historical franchise faces a credible product transition. That is a more aggressive judgment than buying a mature consumer ecosystem after its economics were already established.

This tension places pressure on both sides. Google must show that AI protects or expands its distribution advantage. Berkshire must demonstrate that its new leadership can distinguish durable infrastructure from expensive technological competition.

The AI Infrastructure Bet Comes With a Cash-Flow Test

The strongest argument for the investment is rapid cloud growth, but the largest risk is the spending required to sustain it.

Google Cloud revenue increased 82% year over year during the second quarter, reaching approximately $24.8 billion. The division benefited from demand for AI infrastructure, data services, enterprise software, and model-based applications.

Companywide revenue increased 24% to $119.8 billion, according to the reported Q2 results. Those figures support the view that AI spending is already connected to material customer demand rather than only internal experimentation.

Yet investors focused heavily on capital expenditures. Management raised its expected 2026 spending range to between $195 billion and $205 billion, adding another layer to an already exceptional infrastructure plan.

The tension appeared immediately after the results. Despite strong operating growth, the stock declined as the market weighed cloud momentum against the larger spending commitment. A quarterly earnings analysis described both the record cloud expansion and investor concern about higher capital requirements.

Capital expenditure purchases long-lived assets such as servers, data centers, networking equipment, and related infrastructure. It does not flow through the income statement in the same way as an ordinary operating expense.

That accounting distinction does not remove the economic risk. Management still must generate future revenue and cash flows sufficient to justify the capital committed today.

AI infrastructure can become less valuable if hardware improves rapidly, model-training methods change, customer demand shifts, or competitors force prices lower. Electricity availability, construction timelines, and supply constraints can also delay returns.

The funding structure adds another complication. Google’s parent historically generated enough cash to finance growth, repurchase shares, and maintain a substantial balance sheet. Issuing a large amount of equity changes that familiar picture.

Equity financing avoids fixed interest obligations, but it dilutes existing shareholders. The burden falls on management to invest the proceeds at returns that outweigh the additional share count.

Berkshire’s participation suggests it accepted that tradeoff. The conglomerate was willing to fund infrastructure expansion while increasing its own ownership through both the private placement and additional purchases.

That does not guarantee the bet will work. Berkshire has made large investments that later produced disappointing results, including Kraft Heinz and several airline positions.

A 13F also cannot show Berkshire’s assumptions about future cloud margins, search monetization, infrastructure utilization, or AI model costs. Investors should not substitute the filing for their own analysis.

The competition is formidable. Microsoft combines Azure with enterprise software distribution and a deep relationship with OpenAI. Amazon Web Services remains a major cloud platform and continues expanding AI capacity.

Meta funds its own vast infrastructure program to support advertising systems, recommendation models, consumer assistants, and open model development. Anthropic and OpenAI continue pushing the model frontier while forming partnerships with cloud providers.

Google’s advantage is that it operates across more layers. It designs custom tensor processing units, runs global data centers, develops Gemini models, controls major consumer distribution, and sells cloud services to outside customers.

That vertical integration can lower costs and accelerate deployment. It can also create organizational complexity and force management to allocate capital among products that sometimes compete internally.

Search presents the most important uncertainty. AI-generated answers can improve the user experience, but they can require more computing resources than traditional queries. They may also change where advertisements appear and how users engage with commercial links.

A successful transition would protect the core advertising business while opening new cloud and subscription revenue. A weaker outcome would combine higher computing costs with pressure on the company’s most profitable franchise.

Berkshire’s investment does not resolve that uncertainty. It demonstrates that one of the world’s most closely watched capital allocators considers the potential reward sufficient to accept it.

What the Third-Largest Holding Does Not Prove

Berkshire’s position is a strong signal of conviction, but it is not proof that the AI spending cycle will generate attractive returns.

The first limitation is disclosure timing. Berkshire’s reported holdings are dated June 30, while the filing became public on August 14. The company could have increased, reduced, or hedged portions of its exposure after quarter-end.

The second limitation concerns attribution. Berkshire’s filing aggregates positions held across affiliated managers. It does not state whether Abel, Buffett, Todd Combs, another manager, or a combination of decision-makers directed every purchase.

Buffett remained chairman after Abel became CEO. Investors should therefore avoid presenting the entire position as a solitary decision by either executive without supporting disclosure.

The third limitation concerns ranking. Calling Google’s parent Berkshire’s third-largest holding refers to the reported market value of its disclosed stock positions at quarter-end.

It does not mean the company ranks third among all Berkshire assets. Berkshire owns operating businesses, Treasury securities, cash, foreign investments, and other assets outside the narrow 13F presentation.

Market prices also influence the ranking. A company can move higher in a portfolio because its shares appreciate, because Berkshire purchases more shares, or because other holdings decline or are sold.

In this case, the quarter-to-quarter share increase confirms significant buying. However, future rankings will continue to depend on both trading decisions and market movements.

The fourth limitation is that Berkshire’s investment horizon differs from that of many readers. Berkshire can tolerate volatility, hold through long investment cycles, and evaluate returns over years.

An individual investor copying a delayed filing may purchase at a different valuation and under different market conditions. The disclosure offers evidence about Berkshire’s past action, not a personalized recommendation.

The fifth limitation is concentration risk. A large position can reflect high conviction, but it also makes portfolio results more dependent on a narrower group of companies.

Apple, American Express, and the new third-largest holding together represent a substantial portion of Berkshire’s reported equity value. Their performance can outweigh changes across many smaller positions.

The central skeptical case is straightforward. Google’s parent could continue reporting strong revenue while producing disappointing incremental returns on its infrastructure spending.

Cloud growth can slow as comparisons become harder. AI pricing can decline. Depreciation can rise as newly constructed capacity enters service. Search margins can face pressure from more computationally expensive products.

Regulatory exposure adds another risk. Google remains subject to competition, privacy, advertising, and platform investigations across several jurisdictions. Remedies affecting distribution agreements or product bundling could alter its economics.

The large financing also raises questions about management’s capital discipline. Investors need evidence that new infrastructure will remain highly utilized and support profitable demand, not merely preserve competitive parity.

Berkshire’s involvement strengthens the bullish interpretation because of its reputation for disciplined capital allocation. Reputation, however, is not an operating metric.

The investment will ultimately be judged by Google’s future cash generation, competitive position, and returns on capital. The identity of a prominent shareholder cannot replace those results.

Three Signals Will Test Berkshire’s Alphabet Thesis

The next stage is about execution, with cloud profitability, AI-era search economics, and Berkshire’s future filings providing the clearest evidence.

The first signal is Google Cloud’s growth relative to infrastructure spending. The division’s 82% Q2 expansion makes the current investment cycle easier to defend, but investors need more than one exceptional quarter.

Future reports should show whether cloud revenue, operating income, backlog conversion, and customer demand remain strong as new capacity comes online. Rising utilization would support Berkshire’s apparent view that the buildout serves durable demand.

A sharp slowdown paired with elevated capital expenditure would weaken that interpretation. It would suggest that the company expanded capacity faster than customers could absorb it profitably.

The second signal is the economics of AI-enhanced search. Management must demonstrate that conversational answers can preserve user engagement and commercial intent without permanently compressing margins.

Watch advertising growth, traffic-acquisition costs, query volume, and management’s discussion of AI-serving expenses. Those indicators will show whether the company can modernize its core franchise while retaining its financial quality.

Improving monetization would strengthen the case that Google’s distribution remains a defensible advantage. Weak advertising growth alongside higher computing costs would challenge the idea that the transition is economically controlled.

The third signal is Berkshire’s next portfolio disclosure. Another increase would suggest that the conglomerate’s conviction survived developments after June 30.

An unchanged position would still represent a major long-term commitment. A meaningful reduction would raise questions about valuation, portfolio limits, or a reassessment of the AI investment cycle.

Readers should also separate market reactions from operating evidence. A short-term share-price move will not determine whether the underlying thesis is working.

Berkshire’s Q2 action already answers one question. The conglomerate no longer treats leading technology platforms as categorically outside its circle of competence.

The unresolved question is harder. Can a company spending at unprecedented scale on AI infrastructure produce the durable, predictable returns Berkshire traditionally demands?

Developers and enterprise buyers should watch the same three indicators. Cloud capacity affects service availability, model performance, product pricing, and the range of tools companies can deploy.

Knowledge workers should watch how AI changes search itself. The transition will influence how information is discovered, summarized, attributed, and converted into commercial activity.

Berkshire’s investment makes this more than another quarterly portfolio story. It places one of history’s most conservative capital-allocation institutions behind an unusually aggressive infrastructure expansion.

The next few quarters will reveal whether that pairing represents a durable new model for Berkshire or an expensive test of its post-Buffett judgment. Watch the operating evidence, then compare it with what Berkshire reports owning next.