Is Broadcom Still a Buy After Its Big Run?

Broadcom has reached the stage of its stock-market journey where I start hearing two completely different voices in my head.

The first voice looks at the company’s explosive artificial-intelligence revenue, expanding software business, enormous free cash flow, and growing importance inside modern data centers. It calmly tells me that Broadcom may be one of the best-positioned infrastructure companies of the AI era.

The second voice looks at the stock chart, checks the valuation, and asks whether I have once again arrived at the party after everyone else has eaten the good food.

Both voices have a point.

As of July 24, 2026, Broadcom shares were trading around $392. The stock had already enjoyed a tremendous run as investors rewarded the company for its position in custom AI accelerators, networking products, and infrastructure software. At that price, Broadcom was valued at roughly $1.9 trillion based on its recent share count, and its trailing price-to-earnings ratio was close to 98.

That is not a valuation normally associated with an overlooked bargain sitting beneath a pile of dusty financial statements. Broadcom has been discovered. It has been admired. It has been placed directly beneath the brightest lights on Wall Street.

The question is no longer whether Broadcom is a high-quality business. I believe it is.

The question is whether the quality, growth, and future cash flow are powerful enough to justify buying the stock after the market has already priced in an impressive amount of success.

My answer is yes—but with enough caution to prevent enthusiasm from becoming financial cosplay.

I still consider Broadcom a long-term buy, but I would not treat it as a stock that must be chased at any price. At nearly $400 per share, I see it as a company to accumulate gradually, preferably during volatility, rather than a position I would purchase all at once while assuming the stock market has permanently abolished gravity.

The Company Is No Longer Merely a Chipmaker

When I first encountered Broadcom as an investment, it was easy to think of the company primarily as a semiconductor supplier. It sold specialized chips used in networking equipment, smartphones, broadband devices, storage systems, and other electronic products.

That description is still technically correct, but it now feels like describing a modern airport as a building with chairs.

Broadcom has evolved into a sprawling infrastructure company positioned at the intersection of semiconductors, enterprise software, cloud computing, networking, and artificial intelligence. Its semiconductor products help move and process enormous quantities of data. Its software products help enterprises operate private clouds, mainframes, cybersecurity systems, and mission-critical applications.

In practical terms, Broadcom sells many of the tools that large technology companies and corporations cannot casually replace without creating meetings so unpleasant that everyone involved would rather renew the contract.

The business now has two major segments: semiconductor solutions and infrastructure software. The semiconductor division contains its custom AI accelerators, networking chips, wireless components, storage connectivity products, and other specialized hardware. The software division includes VMware, mainframe software, cybersecurity products, enterprise software, and fibre-channel storage networking.

This combination is important because it gives Broadcom more than one path to growth. AI semiconductor revenue provides the excitement, while infrastructure software can supply recurring revenue and substantial operating profit.

That is a useful arrangement. Excitement attracts investors. Recurring cash flow helps keep them from needing therapy.

The Latest Numbers Are Difficult to Ignore

Broadcom’s fiscal second-quarter 2026 results made the bullish argument considerably stronger.

The company reported quarterly revenue of approximately $22.2 billion, up 48% from the same period a year earlier. Adjusted earnings reached $2.44 per share, while adjusted EBITDA climbed to roughly $15.2 billion, equal to approximately 69% of revenue. Free cash flow was about $10.3 billion, or 46% of revenue.

Those are not ordinary numbers. That is a company converting nearly half of its revenue into free cash flow while growing at a rate normally associated with a much smaller business.

Broadcom is not a promising startup selling an AI-powered toaster and a dream. It is an enormous, established technology company producing tens of billions of dollars in quarterly revenue. Growing that base by 48% is remarkable.

The most important figure was AI semiconductor revenue. It reached $10.8 billion in the quarter, representing year-over-year growth of 143%. Management said that demand for custom AI accelerators and AI networking products drove the result.

Even more striking was the guidance. Broadcom expected third-quarter AI semiconductor revenue of approximately $16 billion, which would represent growth of more than 200% from the prior-year period. Total third-quarter revenue was projected to reach roughly $29.4 billion.

Broadcom did not merely beat expectations and offer vague comments about “strong momentum.” It effectively showed investors a growth engine that was accelerating from an already enormous base. Broadcom’s second-quarter announcement described record revenue, operating profit, and free cash flow, driven by AI semiconductors and operating leverage.

This is why the stock has run.

Wall Street occasionally behaves irrationally, but Broadcom’s advance was not built entirely on inspirational slides and executives repeating the letters A and I. The underlying revenue and cash flow were very real.

Custom AI Chips May Be Broadcom’s Defining Opportunity

Nvidia remains the company most closely associated with AI computing, and deservedly so. Its graphics processors and software ecosystem have become foundational to the industry.

However, the AI infrastructure market is not automatically a single-company kingdom.

The largest cloud companies are spending staggering amounts of money building data centers. They want more computing power, better energy efficiency, greater control over their systems, and less dependence on one supplier. For specific workloads performed at enormous scale, custom accelerators can offer advantages over general-purpose hardware.

That is where Broadcom becomes extremely interesting.

Broadcom works with hyperscale customers to design custom AI accelerators, sometimes called XPUs or application-specific integrated circuits. These chips can be tailored to the customer’s workloads and infrastructure. Broadcom also supplies the networking technology that connects thousands of accelerators inside massive AI clusters.

This means Broadcom does not need to defeat Nvidia at Nvidia’s own game. It can win by helping large customers build specialized alternatives and by supplying the networking products required regardless of which processors occupy the racks.

I think of Broadcom less as the company selling the star player and more as the company building specialized players, transportation systems, and parts of the stadium.

The opportunity could become enormous because AI performance depends on more than raw computing chips. Data must move quickly between processors, memory, storage, and other systems. A cluster containing thousands of accelerators is only as useful as its ability to communicate without turning every data transfer into the digital equivalent of rush-hour traffic.

Broadcom’s expertise in Ethernet switching, optical connectivity, digital signal processing, and custom silicon gives it an important role across that system.

Management has continued expanding its relationships with major AI customers. The company’s potential customer base has grown beyond its earliest hyperscale partners, suggesting that custom silicon is becoming a broader strategic priority rather than an experiment limited to one or two companies.

If Broadcom can maintain its technological edge and keep winning large custom-accelerator programs, the company’s AI revenue may have years of growth remaining.

That possibility is the main reason I am unwilling to dismiss the stock simply because it looks expensive using trailing earnings.

Networking Is the Less Glamorous Hero

Investors understandably focus on the custom accelerators because they represent enormous contracts and dramatic revenue growth. Networking, however, may be one of Broadcom’s most durable competitive strengths.

The growth of AI clusters creates a simple problem with a complicated solution: more processors must communicate with one another at higher speeds while consuming less power.

Broadcom produces switching and routing technology designed for these demanding environments. Its Tomahawk and Jericho product families help data centers move information between servers and across networks.

This market may not receive the same attention as processors, but it is essential. A warehouse full of advanced chips becomes an extremely expensive collection of heaters if data cannot reach them efficiently.

Broadcom’s position in networking gives it an advantage because the company can benefit as AI infrastructure expands even when customers choose different types of processors. Nvidia GPUs, custom accelerators, and other specialized chips still need networking.

I like businesses that sell the unavoidable layer beneath a technological trend. Individual products may change, but the need to move data does not disappear.

The risk, of course, is competition. Nvidia has been strengthening its own networking products, while other semiconductor companies want a share of this growth. Large cloud customers may also attempt to develop more technology internally.

Broadcom must continue investing heavily in research and development. Its market position cannot be protected by placing a velvet rope around its existing products and charging admission.

So far, the company has demonstrated that it can innovate while maintaining exceptional profitability. That combination is rare and valuable.

It is also already reflected in the stock price.

VMware Changed the Business—and Angered Some Customers

Broadcom’s acquisition of VMware transformed the company into a much larger infrastructure software provider.

VMware’s software is deeply embedded in corporate information-technology environments. It allows companies to virtualize computing resources, manage private clouds, and operate applications across complicated infrastructure.

That installed base gives Broadcom pricing power, recurring revenue, and access to mission-critical workloads. It also gives Broadcom an enormous opportunity to improve VMware’s profitability by focusing on the largest customers and simplifying the product portfolio.

Broadcom has never behaved like an executive team that acquires a business and then spends several years organizing inspirational retreats. It cuts costs, concentrates resources, increases efficiency, and expects the acquired company to produce substantial cash.

Investors generally enjoy this approach.

Customers sometimes enjoy it considerably less.

VMware clients and partners have complained about licensing changes, product bundling, price increases, reduced options, and disruptions to previous reseller arrangements. European cloud providers have pushed regulators to examine Broadcom’s VMware practices.

This is not a minor public-relations inconvenience. It is a genuine investment risk.

Software becomes valuable when customers depend on it, but that dependence can create resentment if the supplier pushes too aggressively. In the short term, customers may renew because migration is expensive and disruptive. Over the long term, sufficiently frustrated customers begin searching for alternatives, funding competitors, or rebuilding their systems around different technology.

Broadcom must extract more value from VMware without convincing the entire customer base that escaping has become a sacred mission.

The financial performance has been strong enough to support management’s strategy. Infrastructure software revenue reached approximately $8.3 billion in the second quarter, with VMware Cloud Foundation contributing to the growth. The segment also generated exceptionally high operating margins.

However, I would not assume those profits are immune to customer backlash.

Strong switching costs can protect a software business for years. They are not a license to test the emotional limits of every customer with a technology budget.

The Cash-Flow Machine Is Doing Exactly What I Want

Revenue growth receives most of the attention, but cash flow is where Broadcom becomes especially attractive to me.

Broadcom generated approximately $10.3 billion of free cash flow during its second fiscal quarter. That was about 46 cents of free cash flow from every dollar of revenue.

Many companies describe themselves as cash-generative because they produced enough cash to purchase lunch. Broadcom produces enough quarterly cash to fund major research programs, pay dividends, repurchase shares, reduce debt, and still leave accountants with something substantial to count.

This financial strength gives the company flexibility.

Broadcom pays a dividend, and management has historically increased that dividend as cash flow expanded. The current yield is modest because the stock price has risen so dramatically, but the dividend still adds a shareholder-return component that many fast-growing technology companies do not offer.

The company can also repurchase shares. In the first fiscal quarter of 2026, Broadcom spent approximately $7.85 billion on buybacks, according to its first-quarter regulatory filing.

I have mixed feelings about repurchases when a stock trades at a premium valuation. Buying back undervalued shares can create significant value. Buying back expensive shares can become a very elaborate way of transferring money from the corporate bank account to departing shareholders.

Still, Broadcom’s ability to produce cash at this scale is one of the strongest arguments supporting the valuation. The company is not promising that profits will eventually appear after the business reaches an undefined state called “scale.” The cash is already arriving.

The Balance Sheet Deserves Attention

Broadcom’s cash generation is impressive, but the VMware acquisition left the company with substantial debt.

At the end of the first fiscal quarter, Broadcom reported approximately $68 billion in outstanding borrowings. Cash and equivalents stood at about $14.2 billion, while long-term debt was approximately $63.8 billion.

Those figures would alarm me if Broadcom were a low-margin business facing unpredictable demand and producing inconsistent cash. In that scenario, the debt would resemble a grand piano balanced above the shareholders by a fraying rope.

Broadcom is different because its free cash flow provides significant capacity to service and reduce debt. Its software revenue also adds a recurring component to the business.

Nevertheless, debt reduces flexibility. It increases interest expense, creates refinancing exposure, and makes operational mistakes more expensive. Broadcom cannot assume that AI demand, capital markets, and software renewals will remain favorable under every economic condition.

I would like to see the company continue reducing leverage while maintaining appropriate investment in research and development. A technology company cannot become so focused on financial efficiency that it begins consuming tomorrow’s products to improve today’s margins.

Hock Tan’s record suggests disciplined capital allocation, but even excellent executives operate under the same inconvenient rule as the rest of humanity: previous success does not guarantee permanent infallibility.

Customer Concentration Is a Serious Risk

Broadcom’s growth is tied closely to a relatively small group of enormous customers.

The company disclosed that one semiconductor distributor represented 42% of first-quarter revenue. Broadcom also estimated that its five largest end customers generated approximately half of total revenue during the quarter.

That concentration makes the financial results more vulnerable to individual spending decisions.

If a major hyperscaler delays deployment, changes an internal design, shifts orders to a competitor, or reduces capital spending, Broadcom could feel the effect quickly. Large customers also possess negotiating leverage because losing one relationship could remove billions of dollars from the revenue forecast.

The same customers creating Broadcom’s extraordinary growth can therefore create extraordinary volatility.

I do not view concentration as a reason to avoid the stock entirely. Serving the world’s largest technology companies naturally creates large account sizes. These customers choose Broadcom because its products address highly complex problems, and replacing those products is not always simple.

Still, I refuse to treat announced demand as a permanent law of nature.

AI infrastructure spending is currently enormous, but capital-expenditure cycles can change. Companies may eventually decide they built capacity faster than applications can produce returns. Technical architectures may evolve. Custom accelerators could face delays. A product transition could shift revenue between quarters.

The higher the valuation climbs, the less patience the market will have for any of these normal business complications.

The Valuation Is Where My Enthusiasm Meets Adult Supervision

Broadcom’s business performance is exceptional.

Its valuation is demanding.

At roughly $392 per share, investors are paying for years of powerful growth, expanding AI demand, successful execution, continued customer wins, durable networking leadership, VMware profitability, and enormous free cash flow.

That may all happen.

The uncomfortable part is that much of it must happen.

A premium valuation is not automatically evidence that a stock is overpriced. Great companies often look expensive because traditional metrics measure past earnings while the market values future cash flow.

Broadcom’s trailing price-to-earnings ratio is also distorted by acquisition-related accounting expenses, including amortization of intangible assets. Adjusted earnings and free cash flow can offer a clearer picture of underlying operating performance.

Even after making those adjustments, however, Broadcom is not cheap.

Investors are paying a substantial multiple of expected earnings and cash flow. That leaves little room for ordinary disappointment.

A company can report impressive growth and still watch its stock decline if the growth is slightly below expectations. This is the strange world of premium stocks, where excellent results can be punished for failing to be excellent enough.

The stock does not need Broadcom to fail in order to fall. It only needs investor expectations to cool.

That is why I separate the company from the entry price. Broadcom can remain an outstanding business while the stock experiences a 20% or 30% correction. Those ideas do not contradict each other.

They frequently travel together.

My Bull Case

My bullish argument begins with the possibility that AI infrastructure spending remains strong for many years.

If hyperscale companies continue designing custom accelerators, Broadcom can capture a meaningful share of that market. Its networking portfolio gives it another way to participate in the expansion of AI clusters. These products address technically demanding problems, creating barriers to entry and strengthening customer relationships.

VMware and the broader software portfolio contribute recurring revenue, high margins, and diversification beyond semiconductors. Broadcom’s operating discipline could continue improving the profitability of those assets.

The company’s free cash flow supports debt reduction, dividends, repurchases, and continued investment.

Under a strong scenario, Broadcom could grow earnings rapidly enough for today’s valuation to become more reasonable over time. The stock may continue appreciating even if its valuation multiple gradually declines, provided earnings expand faster than the multiple contracts.

That is the outcome long-term investors want: the business grows into the price.

If Broadcom’s AI semiconductor revenue reaches the enormous figures management and investors anticipate, the company could become one of the defining infrastructure suppliers of this technology cycle.

In that world, purchasing the stock near $392 may look expensive only in retrospect’s rearview mirror, where every successful investment was apparently obvious.

My Bear Case

The bearish case begins with expectations.

Broadcom’s valuation assumes tremendous AI growth. If hyperscale spending slows, custom-chip programs face delays, or a major customer changes suppliers, revenue estimates could fall quickly.

Competition may intensify. Nvidia, Marvell, and other semiconductor companies want larger positions in custom silicon and networking. Major cloud companies have the resources to develop more internal capabilities.

Broadcom could also face margin pressure as more revenue comes from complete AI systems and lower-margin hardware. The company has warned that sales of AI racks and XPU-based systems could reduce gross margins even if they increase operating profit.

VMware customer dissatisfaction could create longer-term erosion. Regulatory action could constrain licensing practices. Debt remains substantial. Customer concentration magnifies volatility. Trade restrictions, tariffs, supply-chain disruptions, or geopolitical tensions could interfere with semiconductor production and sales.

Then there is the simplest risk: the stock is expensive.

If the market decides that Broadcom deserves a lower earnings multiple, shareholders could experience significant losses even while the company continues growing.

The bear case does not require a corporate catastrophe. It requires reality to be merely good while the stock price expects magnificent.

How I Would Approach the Stock

If I did not already own Broadcom, I would not invest my entire intended position at today’s price.

I would divide the amount into several portions and purchase gradually. That approach would give me exposure if the stock continues rising while preserving capital for inevitable volatility.

For example, I might begin with one-quarter of my intended position, add another portion after the next earnings report, and reserve the remainder for meaningful pullbacks. The exact percentages would depend on my portfolio, risk tolerance, time horizon, and existing technology exposure.

I would be especially interested during declines caused by broad market weakness rather than deterioration in Broadcom’s competitive position.

I would be less eager to buy a decline caused by the loss of a major customer, collapsing AI guidance, or clear evidence that VMware’s revenue strategy is damaging long-term retention. A falling price is not automatically a bargain. Sometimes the market is sending an invoice.

I would also watch position size. Broadcom’s size, growth, and momentum can create the temptation to treat it as a guaranteed winner. No stock deserves that status.

A concentrated position may produce wonderful results when the thesis works. It can also turn an earnings announcement into a personal medical event.

My Price Expectations

Predicting a precise stock price is an excellent way to make uncertainty look neatly organized.

Still, I can establish reasonable ranges.

My base case assumes Broadcom continues generating powerful AI growth, software remains profitable, and the valuation gradually becomes less extreme as earnings expand. Under that scenario, I could see the stock trading between $425 and $475 over the next 12 to 18 months.

My bullish scenario assumes AI revenue continues exceeding expectations, new custom-silicon customers ramp successfully, and investors maintain a premium valuation. That could push the stock toward $500 or higher.

My bearish scenario assumes AI growth slows, customer spending becomes less predictable, or the market reduces valuation multiples across the semiconductor sector. Under those conditions, Broadcom could fall toward $300 to $330 without the business becoming permanently impaired.

A more severe disappointment could send it lower.

This range is wide because the future is wide. Anyone offering a precise target without acknowledging uncertainty is not removing risk. They are decorating it.

My Final Verdict: Buy, but Do Not Chase Blindly

After weighing the growth, cash flow, competitive strengths, valuation, debt, and customer concentration, I still view Broadcom as a long-term buy.

But my conviction belongs to the business more than the current price.

Broadcom has developed a remarkable combination of assets. It occupies essential positions in custom AI accelerators, high-speed networking, enterprise software, storage connectivity, wireless chips, and private-cloud infrastructure. It is growing at an extraordinary rate for a company of its size, and it converts an impressive portion of revenue into cash.

Those qualities justify a premium.

They do not justify an unlimited price.

At nearly $400 per share, Broadcom demands patience and disciplined position sizing. I would accumulate gradually, expect sharp volatility, and maintain a time horizon measured in years rather than the distance between two earnings reports.

I would also monitor AI revenue growth, new hyperscale customers, networking share, software retention, free cash flow, debt reduction, and gross-margin trends. If those fundamentals remain strong, temporary stock-price weakness could create attractive opportunities.

If the fundamentals weaken while management begins describing every problem as “timing,” I would become considerably less charitable.

Broadcom’s run has made the decision harder, but it has not destroyed the investment case. The company is not rising merely because investors discovered a fashionable ticker. It is producing the revenue, profits, and cash flow required to support a serious long-term argument.

The challenge is emotional.

Buying a stock after a huge run feels uncomfortable because I immediately imagine that everyone else knows something I discovered embarrassingly late. Refusing to buy can feel equally uncomfortable if the company keeps growing and the stock continues climbing.

The solution is not to predict the perfect moment. I cannot.

The solution is to respect both the company and the valuation.

Broadcom may still be one of the strongest ways to invest in the infrastructure beneath artificial intelligence. I am willing to buy that opportunity.

I am simply unwilling to pretend that a great company has made risk obsolete.

Broadcom is still a buy in my view—just not the kind I would chase with both hands, closed eyes, and the comforting belief that trillion-dollar companies now grow to the sky without ever encountering weather.