The $10.8 Billion Quarter: Broadcom Blows Past Estimates But the Stock Drops—Here is Why

 

The $10.8 Billion Quarter: Broadcom Blows Past Estimates But the Stock Drops—Here is Why

**Subtitle:** *From exploding AI revenues to a dividend hike and a $500 billion market cap—why Broadcom’s record-breaking quarter still left Wall Street wanting more.*

**Reading Time:** 8 Minutes | **Category:** Markets & AI

## Introduction: The Hangover After the AI Feast

It was the kind of quarter that most CEOs only dream about. Revenue hit a record $22.2 billion, up 48% from the previous year . Adjusted earnings per share soared 54% to $2.44, crushing analyst estimates . Free cash flow—the lifeblood of any tech company—exploded to $10.3 billion, representing an eye-watering 46% of revenue .

And yet, after the market closed on Wednesday, Broadcom’s stock fell more than 5% in after-hours trading .

The response seemed, on the surface, irrational. How can a company that just posted the best quarter in its history be punished by the market?

The answer lies in the difference between “beating” and “crushing,” and the impossible expectations that the AI boom has created for semiconductor companies. Wall Street is no longer satisfied with 143% year-over-year AI growth . They want 200%. They want 300%. And when a company delivers merely spectacular results instead of earth-shattering ones, the stock gets penalized.

Broadcom’s second-quarter fiscal 2026 report is a masterclass in the strange psychology of today’s market. The company is executing flawlessly. Its custom AI chips—known as XPUs—are powering the largest hyperscalers on the planet. Its backlog stands at an astonishing $73 billion . Its CEO, Hock Tan, has guided for $100 billion in AI chip revenue by 2027 . By any rational measure, this is a company firing on all cylinders.

But the market is not rational right now. It is drunk on AI hype, and it wants more.

In this deep-dive, we will unpack the numbers that matter, explain why the stock dropped despite the beat, and analyze the secret weapon that makes Broadcom different from Nvidia—its sticky, high-margin custom chip business. We will also look at the dividend hike, the Anthropic deal, and what the company’s massive AI backlog means for the rest of 2026.

## Part 1: The Numbers That Matter – A Quarter for the Record Books

Before we talk about why the stock fell, let’s appreciate just how good this quarter was.

### The Earnings Scorecard

Broadcom reported for its second quarter of fiscal 2026, which ended May 3. Here is how the company performed against Wall Street expectations:

| Metric | Q2 2026 Actual | Q2 2025 | Change | Wall Street Expected |

| :--- | :--- | :--- | :--- | :--- |

| **Revenue** | $22.19B | $15.00B | **+48%** | $22.13B |

| **Adjusted EPS** | $2.44 | $1.58 | **+54%** | $2.39 - $2.40 |

| **GAAP Net Income** | $9.31B | $4.97B | **+88%** | N/A |

| **Adjusted EBITDA** | $15.24B | $10.00B | **+52%** | N/A |

| **Free Cash Flow** | $10.26B | $6.41B | **+60%** | N/A |

| **Cash from Operations** | $10.49B | $6.56B | **+60%** | N/A |

*Sources: *

The headline is the double beat. Revenue of $22.19 billion was slightly above the $22.13 billion consensus. Adjusted EPS of $2.44 beat the $2.40 estimate by four cents .

But the real story is buried deeper—in the segment performance.

### The Semiconductor Engine

Broadcom’s semiconductor solutions segment—the heart of its AI business—reported revenue of **$15.01 billion**, up 79% year-over-year and beating estimates of $14.65 billion .

Within that segment, **AI semiconductor revenue** reached **$10.8 billion**, up 143% year-over-year . That means AI now accounts for approximately 72% of Broadcom’s total semiconductor revenue, up from virtually nothing just three years ago.

| Segment | Q2 2026 Revenue | Q2 2025 Revenue | Change |

| :--- | :--- | :--- | :--- |

| **Semiconductor Solutions** | $15.01B | $8.41B | **+79%** |

| **AI Semiconductor** | $10.80B | $4.44B (est) | **+143%** |

| **Infrastructure Software** | $7.18B | $6.60B | **+9%** |

*Sources: *

The infrastructure software segment grew a more modest 9% to $7.18 billion . But that is still solid performance for a mature business, and it provides stability that pure-play chip companies lack.

### The Profitability Picture

Perhaps the most impressive numbers are the profitability metrics. Broadcom’s adjusted EBITDA margin hit **69%** of revenue—meaning that for every dollar the company brought in, it kept 69 cents as operating profit .

Free cash flow of $10.26 billion represents **46% of revenue** . To put that in perspective, most mature software companies would kill for a 20% free cash flow margin. Broadcom is generating nearly half of its revenue as free cash.

That cash is being returned to shareholders. The company paid $3.09 billion in dividends during the quarter and declared a quarterly dividend of **$0.65 per share**, payable on June 30, 2026 .

### The Guidance

For the third quarter, Broadcom guided for revenue of approximately **$29.4 billion**, up 84% year-over-year . That is ahead of the analyst consensus of $28 billion.

AI semiconductor revenue is forecast to reach **$16 billion** in Q3, representing growth of more than 200% year-over-year .

So, if the numbers were so strong, why did the stock drop?

## Part 2: The Disconnect – Why the Stock Fell on Good News

The answer lies not in what Broadcom reported, but in what investors *expected* it to report.

### The “Buy-Side” vs. “Sell-Side” Gap

There is a dirty secret in Wall Street earnings analysis. The “consensus” estimates you see on Yahoo Finance or Bloomberg—the numbers that determine whether a company “beats” or “misses”—come from sell-side analysts. These are the research departments at investment banks like Goldman Sachs, Morgan Stanley, and UBS.

But there is another set of numbers that matter more: the **buy-side expectations**. These are the numbers that large institutional investors—mutual funds, hedge funds, pension funds—are actually using to make trading decisions. And those numbers are often much higher than the published consensus.

For Broadcom, the buy-side expectation for Q2 AI semiconductor revenue was approximately **$11.3 billion** . The actual number was $10.8 billion—a miss of about $500 million relative to what large investors were hoping for.

For Q3 guidance, the buy-side target was approximately **$18 billion** for AI semiconductor revenue. Broadcom guided to $16 billion .

| Metric | Reported | Buy-Side Expectation | The Gap |

| :--- | :--- | :--- | :--- |

| **Q2 AI Revenue** | $10.8B | $11.3B | **-$0.5B** |

| **Q3 AI Guidance** | $16.0B | $18.0B | **-$2.0B** |

*Sources: *

The stock did not fall because Broadcom performed poorly. It fell because the AI hype machine had inflated expectations to an unrealistic level.

### The “Show Me” Phase of AI

Broadcom’s story illustrates a broader shift in the AI investment landscape. We have entered what analysts call the “show me” phase.

In 2024 and 2025, investors were willing to buy AI stocks based on potential. Any company with even a tangential relationship to AI saw its stock soar. The revenue didn’t matter. The profits didn’t matter. The story mattered.

That era is over.

Today, investors want to see actual numbers. They want proof that the AI demand is translating into revenue. And when a company reports numbers that are merely “great” rather than “transcendent,” the stock gets punished.

Broadcom is the victim of its own success. Because it has grown AI revenue so quickly—from zero to a $40 billion annual run rate in three years—investors now expect that pace to continue indefinitely. When it merely doubles instead of tripling, the market reacts negatively.

**The Human Touch:** For the retail investor who bought Broadcom at $400 expecting a smooth ride to $600, the after-hours drop is a gut check. The company is not broken. The AI boom is not over. But the days of effortless 10x returns are behind us. From here, the gains will be harder won.

### The Options Market Signal

Traders had priced in a roughly **9% post-earnings swing** in Broadcom’s stock price, with near-the-money straddles pointing to a trading band of about $439 to $525 by Friday expiration .

The options activity showed a bullish tilt, with the 500 call drawing more than 16,000 contracts, along with additional upside interest clustered around the 480, 485, 510, 530, 550 and 600 strikes .

So far, the move has been to the downside—but the 9% potential swing is still in play. The coming days will determine whether the dip is a buying opportunity or the start of a deeper correction.

## Part 3: The Secret Weapon – Broadcom’s Custom Chip Moat

To understand why Broadcom is positioned for long-term success, you have to understand the difference between what it does and what Nvidia does.

### ASICs vs. GPUs – The Great Semiconductor Divergence

Nvidia dominates the market for **GPUs**—general-purpose processors that can be used for a wide range of AI workloads. These chips are sold off the shelf. Any cloud provider, any startup, any enterprise can buy them.

Broadcom dominates a different market: **ASICs** (Application-Specific Integrated Circuits), which the company calls “XPUs.” These are custom-designed chips built for specific customers and specific workloads.

| Chip Type | Nvidia | Broadcom |

| :--- | :--- | :--- |

| **Product** | GPUs (off-the-shelf) | ASICs/XPUs (custom-designed) |

| **Customers** | Anyone | Hyperscalers only |

| **Switching Costs** | Low | Extremely high |

| **Margins** | Very high | High (but different profile) |

| **Design Time** | N/A | Years |

*Sources: *

Once a hyperscaler like Google, Meta, or Anthropic commissions Broadcom to design a custom chip, the switching costs are enormous. The design process takes years. The data center is built around that specific architecture. Changing suppliers would require a complete rebuild.

This is the **stickiness** of Broadcom’s business. It is arguably more durable than Nvidia’s moat.

### The Customer Roster – A Who’s Who of Tech

Broadcom’s custom chip customers read like a list of the most valuable companies on Earth:

- **Alphabet (Google):** Broadcom has designed at least seven generations of Google’s Tensor Processing Units (TPUs), starting in 2014 .

- **Meta Platforms:** The MTIA accelerator runs on Broadcom silicon, with a two-year roadmap laid out .

- **Anthropic:** Signed a multi-year collaboration in October 2025 for chips targeting deployment in the second half of 2026 .

- **OpenAI:** Has a multi-year agreement with Broadcom for custom AI chips .

- **ByteDance:** Uses Broadcom’s custom silicon for its massive-scale operations .

- **Apple:** Widely expected to be next, reportedly partnering with Broadcom on a custom AI server chip codenamed “Baltra” .

This is not a customer base. It is a roll call of the AI revolution.

### The Anthropic Shift – Lower Revenue, Higher Margins

One of the most interesting stories from the quarter involves Anthropic. Earlier expectations had Broadcom shipping full server racks to the AI lab behind Claude. Full racks bundle the chip with memory, networking, and other hardware, which adds a lot of low-margin revenue.

That order has now shifted to a **chip-only arrangement**. It brings in roughly 25% of the originally expected revenue—but at far higher margins .

This is why Susquehanna analyst Christopher Rolland raised his price target to $490 even while trimming his AI revenue estimates . He is paying up for cleaner, higher-margin earnings rather than raw revenue scale.

The takeaway: Broadcom is not just chasing top-line growth. It is prioritizing profitability. And in a market that is increasingly focused on earnings, that is the right strategy.

### The $73 Billion Backlog

Perhaps the most underappreciated number in the entire earnings report is the backlog.

Citi projects Broadcom’s total AI revenue will reach **$115 billion** in fiscal 2027 and **$180 billion** in fiscal 2028 . That would represent roughly 81% of the company’s total sales.

Underpinning that outlook is a **$73 billion AI backlog**—larger than Broadcom’s entire fiscal 2025 revenue base—with delivery scheduled over approximately the next 18 months .

This is not hype. This is contracted revenue. This is already locked in.

The $73 billion backlog means that even if new orders slow down, Broadcom has visibility into massive growth for the next year and a half. That is the kind of certainty that most semiconductor companies can only dream of.

**The Human Touch:** For the Broadcom employee working on these custom chips, the backlog means job security. The company is not laying off. It is hiring. It is expanding. The AI boom is not a fad—it is a multi-year buildout that is just getting started.

## Part 4: The Bigger Picture – Broadcom’s Role in the AI Infrastructure Buildout

To understand where Broadcom fits in the AI ecosystem, you have to look beyond the chip itself.

### The Networking Story

AI training clusters require thousands—sometimes hundreds of thousands—of chips to work together. The connections between those chips are just as important as the chips themselves.

Broadcom is a leader in **AI networking**—the silicon that links thousands of GPUs and XPUs inside a data center. This business is growing rapidly alongside the custom chip business, and it provides another layer of revenue that is less visible but equally important .

### The Software Pivot

Broadcom’s infrastructure software segment—which includes the VMware acquisition—grew 9% to $7.18 billion . This segment is often overlooked, but it provides stability and diversification.

When the semiconductor cycle inevitably turns, the software business will provide a cushion. That is why Broadcom trades at a premium to pure-play chip companies.

### The VMware Integration

The VMware acquisition has been controversial. Some customers have complained about price hikes and licensing changes. But financially, the integration is working. The software segment is generating steady cash flow, and the operating leverage is improving.

CEO Hock Tan has a reputation for ruthless efficiency. He buys companies, integrates them, and extracts synergies. The VMware acquisition is following that playbook.

## Part 5: What This Means for Your Portfolio

Broadcom’s earnings report has implications for three different types of investors.

### For the Long-Term Investor

If you are investing for the next five years, the after-hours drop is irrelevant. Broadcom is executing perfectly. The custom chip moat is widening. The backlog is massive. The management team is proven.

At current levels—around $480 per share—Broadcom trades at a forward P/E of approximately 37 . That is expensive by historical standards, but cheap relative to its growth rate. The PEG ratio (price/earnings-to-growth) is under 1, suggesting the stock is reasonably valued given the growth trajectory.

The key risk is customer concentration. Broadcom’s AI business is heavily dependent on a handful of hyperscalers. If one of them—say, Google or Meta—decides to bring chip design in-house, it could hurt the business. But given the design cycles and switching costs, that risk is minimal in the near term.

### For the Trader

The options market is pricing in continued volatility. The 9% post-earnings swing is still in play, and the stock could trade in a wide range over the coming days .

The bullish tilt in options activity suggests that some large traders are betting on a rebound. The 500 call drew more than 16,000 contracts, and there was significant interest at the 510, 530, 550, and 600 strikes .

If you are trading this stock, be prepared for whipsaw. The fundamentals are strong, but sentiment is fickle.

### For the Income Investor

Broadcom declared a quarterly dividend of $0.65 per share, payable on June 30, 2026 to shareholders of record as of June 22 .

That works out to a dividend yield of approximately 0.5% at current prices—not huge, but growing. The company has a history of increasing its dividend annually, and the massive free cash flow generation suggests that trend will continue.

### The Analyst Consensus

Despite the after-hours drop, the analyst community remains overwhelmingly bullish on Broadcom.

| Firm | Rating | Price Target |

| :--- | :--- | :--- |

| **UBS** | Buy | $490 |

| **Citi** | Top Pick | $500 |

| **Goldman Sachs** | Buy | $500 |

| **Susquehanna** | Positive | $490 |

| **Oppenheimer** | Buy | $450 |

| **Evercore** | Buy | $582 |

*Sources: *

The consensus rating is a **Strong Buy**, with an average price target of approximately **$480** . That suggests the stock is fairly valued at current levels—but there is upside if the AI momentum continues.

**The Human Touch:** For the investor who bought Broadcom at $300 a year ago, the 80% gain is life-changing. For the investor who bought at $480 yesterday, the drop is painful. But over the long term, the company’s trajectory is clear. The AI buildout is just getting started, and Broadcom is one of the few companies with a front-row seat.

## Frequently Asked Questions (FAQ)

**Q: Did Broadcom beat or miss earnings?**

A: Broadcom beat analyst expectations on both revenue and earnings. Revenue of $22.19 billion beat the $22.13 billion consensus, and adjusted EPS of $2.44 beat the $2.39-$2.40 estimate . However, the stock fell because buy-side expectations for AI revenue were higher than the actual reported numbers .

**Q: How much AI revenue did Broadcom report?**

A: Broadcom reported AI semiconductor revenue of **$10.8 billion** in the second quarter, up 143% year-over-year. For the third quarter, the company guided to $16 billion in AI revenue, representing growth of more than 200% .

**Q: What is Broadcom’s dividend?**

A: Broadcom declared a quarterly dividend of **$0.65 per share**, payable on June 30, 2026 to shareholders of record as of June 22, 2026 .

**Q: Who are Broadcom’s main custom chip customers?**

A: Broadcom’s custom chip customers include Alphabet (Google), Meta Platforms, Anthropic, OpenAI, and ByteDance. Apple is widely expected to be the next major customer .

**Q: What is the difference between Nvidia and Broadcom?**

A: Nvidia sells off-the-shelf GPUs that can be used for any AI workload. Broadcom designs custom chips (ASICs or XPUs) for specific customers and specific workloads. Broadcom’s custom chip business has much higher switching costs and is arguably more defensible .

**Q: What is Broadcom’s AI backlog?**

A: Broadcom has a **$73 billion AI backlog**—contracted revenue scheduled for delivery over approximately the next 18 months. This is larger than the company’s entire fiscal 2025 revenue base .

**Q: Is Broadcom stock a buy after the drop?**

A: (Disclaimer: Not financial advice.) The analyst consensus is a Strong Buy, with price targets ranging from $450 to $582. The company has a massive backlog, sticky customer relationships, and industry-leading margins. However, the stock is trading at a forward P/E of approximately 37, which is expensive by historical standards. Long-term investors may see the dip as a buying opportunity, but traders should be prepared for continued volatility.

**Q: What is Hock Tan’s 2027 AI revenue target?**

A: Broadcom CEO Hock Tan has guided for AI chip revenue **in excess of $100 billion** by 2027, with supply capacity locked in through 2028 .

## Conclusion: The Long Game

We started this article with a paradox: a record-breaking quarter followed by a stock drop. We end with a reminder that the market is not always rational in the short term.

Broadcom is executing perfectly. The AI custom chip business is growing at an astonishing rate. The backlog is massive. The margins are industry-leading. The management team is proven.

And yet, because the buy-side expectations were inflated—because the AI hype machine had priced in perfection—the stock fell.

**For the Long-Term Investor:**

Ignore the noise. Broadcom is one of the best-positioned companies in the AI buildout. The custom chip moat is widening, not shrinking. The $73 billion backlog provides visibility. The free cash flow generation is breathtaking.

**For the Short-Term Trader:**

The options market suggests continued volatility. The bullish tilt in call activity indicates that some large traders are betting on a rebound. But be prepared for whipsaw.

**For the Income Investor:**

The dividend is safe and growing. At $0.65 per share quarterly, the yield is modest, but the growth trajectory is strong.

**The Bottom Line:**

Broadcom just delivered one of the best quarters in its history. The stock fell because investors wanted transcendent, not merely spectacular. But over the long term, the company’s trajectory is clear. The AI buildout is just getting started, and Broadcom is one of the few companies with a front-row seat.

The $10.8 billion AI quarter is a milestone. The $16 billion AI quarter is coming. And the $100 billion AI year is on the horizon.

The story is not over. It is just beginning.

---

**#Broadcom #AVGO #AI #Semiconductors #Earnings #Dividend #Investing #TechStocks**

---

*Disclaimer: This article is for informational purposes only. It does not constitute financial advice. Stock markets are volatile; past performance does not guarantee future results. Always consult a licensed professional before making investment decisions.*