Broadcom: What Happened to Its Stock Prices?

From AI Dreams to Profit Reality

In today’s competitive market, companies like Broadcom exemplify how the narrative age is over. Markets now demand execution.

Broadcom (AVGO) and Oracle (ORCL) just posted record-breaking results. The numbers looked perfect on paper. Yet their stocks crashed hard, sending shockwaves through the entire AI infrastructure sector.

What went wrong? The answer reveals a fundamental shift in how markets value AI investments.

The ‘Show-Me’ Revolution Begins

Between 2023 and 2024, markets rewarded AI companies based purely on potential. The narrative alone—”AI will change everything”—was enough to justify soaring valuations. NVIDIA and similar players delivered on expectations, but many others rode high on future promises without proven returns.

Now, in late 2025, the question has changed.

Wall Street is asking: “Does AI investment actually convert to profits?” Analysts coined a new term for this moment: the ‘Show-Me Story’—proof over promises.

Markets want execution. They want evidence. Growth alone isn’t enough anymore.

_{Broadcom Q4 FY25 (Oct. quarter) / Source: appeconomyinsight}

Broadcom’s $73B Problem: When More Revenue Means Less Profit

Broadcom’s order backlog hits an impressive $73 billion. Of that, $53 billion goes to custom AI accelerators (XPUs), with $20 billion for switches and networking gear.

Tech giants like Google, Meta, and ByteDance are using Broadcom to reduce dependence on NVIDIA GPUs. Broadcom delivers complete “full system” solutions—custom chips bundled with network infrastructure.

But here’s the catch: profitability is tanking.

CEO Hock Tan warned investors that AI revenue growth would “lower gross margins due to component pass-through costs.” Translation: Broadcom is becoming a hardware assembler rather than a high-margin chip seller.

Selling complete rack systems to customers like Anthropic increases total revenue but crushes margins. Meanwhile, non-AI semiconductor sales grew just 2% in Q4. AI is carrying the entire company—a dangerous dependency.

Tan refused to provide specific 2026 guidance, saying it’s “difficult to pinpoint exactly how 2026 will unfold.” Markets wanted certainty. Leadership offered caution. The stock reflected that gap.

Wall Street remains optimistic—HSBC maintains a $535 target, BofA raised theirs from $460 to $500—but warns valuation adjustments are inevitable. High expectations are already priced in. Margin pressure is real.

_{Oracle Q2 FY26 (Nov. quarter) / Source: appeconomyinsights}

Oracle’s Dangerous Bet: Cash Gone, Debt Rising

Oracle’s strategy is crystal clear: become the largest AI-focused cloud operator by 2031.

Chairman Larry Ellison projects 40%+ cloud growth, 70%+ infrastructure (IaaS) growth, and 100%+ remaining performance obligations (RPO) growth for fiscal 2026. They’re building 47 multi-cloud data centers and 30 dedicated Oracle cloud facilities in the next 12 months.

The problem? How they’re funding this vision.

Oracle spent $12 billion last quarter alone—$4 billion more than Wall Street expected. Annual capital expenditure targets jumped from $35 billion to $50 billion.

The result: free cash flow (FCF) plunged to negative $10.3 billion in the first half of fiscal 2026. Same period last year? Positive $2.2 billion. Non-current debt now stands at $100 billion with just $19.2 billion in cash.

Cash is vanishing. Growth is funded by debt.

This strategy works when interest rates are low and capital is abundant. But the Fed is slowing rate cuts. Refinancing costs are rising. Credit spreads are widening.

Oracle’s customer base adds another layer of risk. Many key clients are AI startups like OpenAI—well-funded but not yet profitable. If these companies face funding crunches or business model failures, Oracle’s $138 billion RPO could become uncollectible paper.

Management highlighted contracts with Meta and NVIDIA to demonstrate customer diversification, but investor skepticism persists.

KeyBanc’s Jackson Ader downgraded the price target from $350 to $300, noting “Oracle’s capex plan is the most aggressive in the industry relative to its revenue base.” Morgan Stanley added that “investors are losing confidence in Oracle’s ability to convert its massive backlog into sustainable, profitable revenue.”

The Horizontal Expansion Dilemma

Broadcom and Oracle represent a critical trend: horizontal expansion in AI infrastructure. This contrasts sharply with NVIDIA’s vertical integration strategy.

NVIDIA controls the entire stack—GPU hardware, CUDA software, and cloud partnerships. Like Apple with iPhone hardware and iOS software, NVIDIA maintains high margins through ecosystem control.

Broadcom takes a different path. Custom ASICs and networking equipment enable hyperscalers to reduce NVIDIA dependence. But custom chips cost more to develop than general-purpose GPUs, and system sales resemble assembly work with thin margins. Broadcom’s gross margin hovers around 60% but faces downward pressure as AI products grow.

Oracle pursues horizontal expansion in cloud infrastructure. In a market dominated by AWS, Azure, and Google Cloud Platform, Oracle differentiates through AI-specialized cloud services. Oracle Cloud Infrastructure (OCI) loads massive NVIDIA GPU capacity optimized for high-performance computing, with a multi-cloud strategy that partners with existing providers.

But cloud infrastructure is capital-intensive. AWS built operational efficiency over 20 years of scale economics. Oracle just started massive investments. Data center construction and operation require enormous cash, and profit conversion takes time. Oracle’s cash burn illustrates this challenge vividly.

Vertical integration offers high margins and ecosystem dominance. Horizontal expansion, as scale increases, faces growing profitability constraints.

Why High Interest Rates Change Everything

The macro environment no longer tolerates debt-fueled growth.

Post-pandemic, the Fed’s zero-rate policy created ideal conditions for aggressive debt-financed expansion. Borrowing costs were nearly zero.

Today’s reality is different.

Inflation remains above 3%. The Fed has effectively paused its rate-cutting cycle. Oracle’s $100 billion debt means rising refinancing costs and widening credit spreads.

Higher rates force companies to refinance maturing debt at elevated rates. Cash depletion risks credit rating downgrades. Lower ratings increase borrowing costs further—a vicious cycle.

Oracle’s debt-driven growth model only works in low-rate environments. For Oracle and its investors, Fed rate cuts have become the single most critical catalyst for growth. Markets already worry about Oracle’s excessive leverage. If recession brings credit tightening, survival itself could be threatened.

Management pledged to reduce borrowing and maintain investment-grade ratings, but achieving this with negative FCF won’t be easy. Asset sales or equity issuance could service debt but would weaken growth momentum.

Oracle’s strategy is essentially a “burning bridges” approach—all-in with no retreat.

Insight Bridge AI Perspective: Act Two of AI Investment

Broadcom and Oracle reported stellar results but disappointed markets for a simple reason: they’re making money, but spending too much relative to earnings. Margins are shrinking. Cash is disappearing.

The challenge? Debt-fueled growth strategies face headwinds in today’s high-rate environment.

AI investment is entering Act Two. If Act One was a “festival of possibilities,” Act Two is “validation through execution.”

The narrative era is finished. Companies that can’t prove profitability and financial discipline will be left behind. Growth numbers alone won’t cut it anymore.

The market’s message is clear: Show me the money—or get out.

See more insightful news!

Broadcom & Oracle Shock: The ‘Show Me’ Era Has Arrived