AVGO at the Crossroads: Why Wall Street is Selling a Record Quarter — and Why That's the Opportunity

By Balaji N, CFP · True Value Research · June 7, 2026

The Paradox: Record Everything, Falling Stock

On June 3, 2026, Broadcom Inc. reported the most exceptional quarter in its corporate history. Revenue of $22.2 billion — up 48% year-over-year and ahead of every estimate on the Street. Non-GAAP earnings of $2.44 per share, up 54% from a year ago. Free cash flow of $10.3 billion in a single quarter — a 46% conversion rate that most technology companies can only dream about. And a Q3 revenue guide of $29.4 billion — an 84% year-over-year acceleration implying the company will generate more revenue in a single quarter than it did in all of fiscal 2021.

And yet, AVGO shares currently trade at approximately loading..., down in a session where the broader NASDAQ has cratered 4.18% and the S&P 500 is off 2.64%. The question every institutional desk is asking is not "why are the results good?" — they obviously are. The question is: why is the stock falling, and is this a trap or a gift?

Having covered semiconductors and infrastructure software for over a decade, I will give you the honest answer Wall Street will not say plainly: the stock is falling for five reasons, none of which change the long-term thesis.

Five Reasons the Stock Is Falling — None of Them Are the Business

1. The Market Is in Risk-Off Mode, and AVGO Is a High-Beta Name

Let's start with the obvious and most significant driver. The NASDAQ is down over 4% today. In a session like this, no stock — no matter how good the earnings — escapes the gravitational pull of macro fear. Broadcom carries a market capitalisation of approximately $1 trillion, which means every institutional portfolio manager who needs to raise cash is selling the winner. AVGO, having outperformed dramatically through the AI buildout cycle, is a natural source of liquidity. This is not a stock-specific signal. It is portfolio mechanics.

2. "Sell the News" After a Pre-Earnings Run

Broadcom's AI narrative has been one of the most well-telegraphed stories on the Street for 18 months. Investors who were paying attention knew the Q2 print would be exceptional. The stock price had likely already moved to price in much of the good news before the earnings release. When the results confirm what was already expected — even if they exceed the bar — profit-taking from traders who bought the rumour is a rational, mechanical response. This is not capitulation from long-term holders. It is position rebalancing from short-duration players.

3. Gross Margin Compression Is Triggering Margin Models

This is where the bears have their most legitimate concern, and it is worth taking seriously. Non-GAAP gross margins came in at 77.1% — down 230 basis points year-over-year from 79.4% in Q2 FY2025. Quant models that screen for margin deterioration flagged this immediately. The worry is mechanically sound: if Broadcom is growing revenue at 48% but its gross margins are compressing, then the incremental dollar of AI semiconductor revenue carries lower profitability than historical segments. The concern is that as custom AI accelerators (XPUs) become a larger share of the mix, the structurally higher-margin software and networking businesses are diluted.

4. The Inventory Surge Is Raising Red Flags

Inventory jumped from $2.27 billion in November 2025 to $4.33 billion by May 2026 — a 90.7% increase in just six months. On the surface, this is consistent with management building ahead of the enormous demand they see coming. The CEO explicitly guided to $16 billion of AI semiconductor revenue in Q3 alone, up from $10.8 billion in Q2. If that guidance is accurate, you need inventory to fulfil it. But Wall Street's risk models don't see it that way. A 91% inventory build in six months, without complete visibility into the demand from each hyperscaler customer, triggers supply-chain risk flags — the same flags that preceded painful corrections in semiconductor cycles throughout history. The bears are asking: what if one or two hyperscalers slow their capex? Who absorbs $4.3 billion in inventory?

5. The Q3 Guide Creates an Impossibly High Bar

This is perhaps the most counterintuitive bear argument: the guidance was too good. A Q3 revenue target of $29.4 billion — 32% above Q2's already record $22.2 billion — is a sequential step-up that the Street has never seen from a company of this scale. When management sets a bar that high, any stumble — a hyperscaler shifting a shipment by 4–6 weeks, a foundry yield issue, a China-related export restriction — instantly becomes a $1 billion miss on what was already a stretched consensus estimate. The market is pricing in execution risk. It is not saying the business is bad. It is saying the bar is terrifying.

Why the Bears Are Looking at the Wrong Scoreboard

Every one of the five concerns above is real. None of them is wrong, exactly. But they all share a common flaw: they are measuring Broadcom against yesterday's business, not tomorrow's. Let me walk through why the bull case has never been stronger, and why a pullback to loading... in a market selloff is one of the most interesting entries in the AI infrastructure trade.

The AI Accelerator Thesis Is Compounding, Not Decelerating

Broadcom is not NVIDIA. That distinction matters enormously and is consistently missed by investors who lump every semiconductor name into a single "AI chip" bucket. NVIDIA designs and sells general-purpose GPU clusters. Broadcom designs custom AI accelerators (XPUs) built to the exact specifications of individual hyperscaler customers — Google's TPUs, Meta's MTIA chips, Amazon's Trainium. These are not interchangeable commodity chips. They are bespoke silicon integrated deeply into each customer's training and inference stack.

This distinction creates something extraordinary: switching costs so high that hyperscalers cannot easily exit. Once Google has built its training infrastructure around a custom Broadcom XPU, the cost of redesigning, retaping, and requalifying a replacement chip is measured in billions of dollars and multiple years. Broadcom is not just a semiconductor supplier. It is an embedded infrastructure partner.

The numbers validate this: AI semiconductor revenue went from approximately $4.4 billion in Q2 FY2025 to $10.8 billion in Q2 FY2026 — a 143% increase in 12 months. And Q3 guidance implies $16 billion — another 48% step-up in a single quarter. The compounding rate here is not slowing. It is re-accelerating.

AI Networking: The Hidden Crown Jewel

Most of the attention on Broadcom's AI story focuses on custom silicon. The real sleeping giant is AI networking. As AI clusters scale from thousands to hundreds of thousands of accelerators, the interconnect — the fabric that allows GPUs and XPUs to communicate — becomes as important as the compute itself. Broadcom's Tomahawk and Jericho product lines are the de facto standard for hyperscale AI networking. Every major AI training cluster being built today runs on Broadcom switching silicon.

This business carries structurally higher margins than custom silicon and enjoys even stronger competitive positioning — there are no credible alternatives at scale. As cluster sizes explode from 100,000-chip installations today toward projected 1-million-chip "AI factories" by 2028, the networking revenue per cluster grows super-linearly. More chips in a cluster means exponentially more switch ports, more cables, and more networking silicon. Broadcom is the only company positioned to supply this at scale.

Infrastructure Software: The Recurring Revenue Engine Nobody Talks About

Broadcom's $61 billion acquisition of VMware in November 2023 was controversial. The Street worried about integration risk, customer churn, and whether the software business could survive Broadcom's aggressive licensing restructuring. Two and a half years in, the answer is unambiguous: the software business is a recurring cash machine.

Infrastructure software generated $7.18 billion in Q2 FY2026 — up 8.8% year-over-year, and critically, the revenue is predominantly subscription-based. The VMware portfolio — vSphere, vSAN, NSX, Aria — represents the compute and networking virtualisation layer that 95% of Fortune 500 companies run their data centres on. These customers cannot leave. Their entire IT architecture depends on VMware. As AI drives enterprises to modernise their infrastructure, VMware's private cloud solutions become more relevant, not less.

The Operating Leverage Story Is Just Beginning

Here is the number I want every investor focused on: in Q2 FY2026, Broadcom grew revenue by 48% while growing total non-GAAP operating expenses by just 3%. Every marginal dollar of revenue above the cost base falls almost entirely to the bottom line. Non-GAAP operating margins expanded 200 basis points year-over-year to 67.3%. Adjusted EBITDA margins reached 69%.

At Q3's guided $29.4 billion revenue and 67% non-GAAP operating margins, that implies approximately $19.7 billion of operating income in a single quarter. I estimate annualised FCF approaching $45–50 billion by Q4 FY2026 — which would make Broadcom one of the highest free cash flow generating companies in the world, on par with Apple in its peak years.

Five Reasons to Buy or Hold AVGO Here

1. You Are Buying Record Free Cash Flow at a Discount

At approximately loading... per share, investors are buying a business generating roughly $41–50 billion of annualised free cash flow — a 4–5% FCF yield. For a company growing revenue at 80%+ and expanding margins, that is cheap relative to the growth trajectory. Apple and Microsoft trade at similar or lower FCF yields with a fraction of the growth. The gross margin compression argument deserves a direct rebuttal here: the decline from 79.4% to 77.1% reflects mix shift, not pricing power deterioration. 77.1% gross margins on $22.2 billion of revenue generates $17.1 billion of gross profit in a single quarter — a record. The percentage is lower; the dollars are higher and accelerating. Selling on margin percentage while ignoring margin dollars is a category error.

2. The AI Capex Cycle Has Years of Runway Left

Every major hyperscaler — Google, Meta, Amazon, Microsoft — has committed to multi-year, multi-hundred-billion-dollar AI infrastructure programmes. These are not discretionary budget line items. They are strategic bets on the future of their core businesses, already contracted and in-flight. The AI capex supercycle has a minimum of 3–5 years of runway from where we stand today. The revenue Broadcom will report in FY2028 is being contracted in design cycles right now.

3. Custom Silicon Is Structural Revenue, Not Cyclical

XPUs are not a commodity purchase. They are 2–3 year design-to-tape-out cycles, tightly integrated into each hyperscaler's software stack. The revenue visible in FY2026 was contracted in FY2024. Once Google or Meta builds a training cluster around a Broadcom custom accelerator, the cost of redesigning and requalifying a replacement is measured in billions of dollars and multiple years. This is not cyclical exposure to a procurement budget. It is structural lock-in. The inventory build of $4.3 billion reflects this reality — Broadcom is building ahead of demand that is already committed, not speculative.

4. The Dividend Provides Real Capital Return

At $0.65 per share quarterly ($2.60 annualised), Broadcom pays a real and growing dividend. At loading... per share that is a 1.3% yield — modest in isolation, but meaningful when combined with buybacks and set against a company generating $10 billion of free cash flow per quarter. As FCF grows into the $45–50 billion annualised range by Q4 FY2026, dividend growth has significant headroom. Investors waiting for the "right entry" on AVGO have consistently paid more for less.

5. Hock Tan Has Never Missed His Own Guidance

This is not a trivial point. Every quarter where Tan has issued forward guidance, Broadcom has met or exceeded it. The Q3 guide of $29.4 billion is aggressive by any conventional measure. But it is the same CEO who guided to $10.8 billion of AI semiconductor revenue in Q2 and then described it as "above our forecast." The execution track record here is exceptional. Investors who have taken Tan at his word over the past decade have been rewarded consistently. Betting against the guidance is betting against one of the most operationally disciplined management teams in enterprise technology.

Important Disclosures & Disclaimer. This article is produced by True Value Research for informational and educational purposes only. It does not constitute investment advice, a solicitation to buy or sell any security, or a guarantee of future performance. All financial data is sourced from Broadcom Inc.'s official Q2 FY2026 press release (June 3, 2026). Forward EPS estimates, price targets, and probability assessments are the analyst's own projections and are inherently uncertain. Past performance is not indicative of future results. Readers should conduct their own independent research and consult a qualified financial advisor before making any investment decision. Not investment advice.