Broadcom’s Accelerating AI Revenue Meets Heightened Expectations and a Sharp Market Repricing

Broadcom Inc. (AVGO) reported another quarter of exceptional top-line expansion driven by its expanding role in artificial intelligence infrastructure, only for investors to punish the stock for what they viewed as insufficiently aggressive forward guidance. The disconnect highlights a classic late-cycle dynamic in high-growth technology sectors: operational delivery that continues to accelerate can still trigger valuation compression when it fails to exceed already euphoric expectations.

In fiscal Q2 2026 (ended approximately May 3), Broadcom posted record consolidated revenue of $22.187 billion, up 48% year-over-year. Semiconductor Solutions revenue reached $15.009 billion (+79% YoY), while Infrastructure Software (primarily VMware) grew a more modest 9% to $7.178 billion. The standout figure was AI-related semiconductor revenue of $10.8 billion, representing 143% year-over-year growth and comfortably beating internal forecasts. This segment now constitutes the dominant driver of Broadcom’s semiconductor mix.

Adjusted EBITDA rose 52% to $15.244 billion (69% of revenue), demonstrating powerful operating leverage even as the product mix shifts heavily toward higher-volume but lower-gross-margin AI accelerators and networking silicon. Non-GAAP diluted EPS of $2.44 beat consensus estimates around $2.39–$2.40. Free cash flow remained robust at $10.262 billion.

Guidance and the Market’s Interpretation

Management guided Q3 FY2026 consolidated revenue of approximately $29.4 billion (+84% YoY) and AI semiconductor revenue of $16.0 billion (more than 200% YoY growth). This implies a very large sequential step-up from the $10.8 billion recorded in Q2. Non-GAAP operating margin is expected to remain stable near 67%.

On the surface, these figures reflect continued acceleration rather than any deceleration in underlying demand. CEO Hock Tan emphasized that momentum in custom AI accelerators (XPUs) and AI networking “continues,” supported by six core hyperscaler and AI lab customers including Google, Meta, Anthropic, and OpenAI. He highlighted planned deployments scaling to more than 10 gigawatts of compute capacity in 2027, with further expansion expected in 2028.

Yet the stock fell approximately 14–15% in after-hours and subsequent trading—the sharpest single-day decline in well over a year. Several factors converged to produce this reaction:

• The Q3 AI revenue guide of $16 billion came in slightly below certain Street models (around $16.36 billion in one Visible Alpha compilation).
• Management reiterated rather than raised its long-term target of AI semiconductor revenue “in excess of $100 billion” for fiscal 2027. Some investors had anticipated an upward revision given the strength of Q2 and the implied full-year 2026 run-rate (approximately $56 billion, or roughly 180% growth).
• Infrastructure Software growth of only 9% missed estimates and underscored ongoing integration or legacy headwinds in the VMware business.
• Total Q2 revenue slightly missed consensus ($22.19 billion actual vs. ~$22.27 billion expected).

In a market that had already priced in near-perfection for AI enablers—Broadcom shares were up roughly 40% year-to-date before the print—even modest shortfalls relative to elevated expectations produced rapid de-risking.

Context: Hyperscaler Capex Remains in Full Acceleration Mode

The market reaction should not be conflated with an actual slowdown in AI infrastructure spending. Aggregate capital expenditure plans from the leading hyperscalers continue to scale aggressively. Recent guidance points to roughly $725–750 billion in combined capex for 2026 across Amazon ($200 billion), Microsoft ($190 billion), Alphabet ($175–190 billion range), Meta ($115–145 billion), and Oracle, representing year-over-year growth of 60–77% in several cases.

These figures dwarf prior cycles and reflect sustained conviction that training and inference workloads will continue to absorb capacity at scale. Power availability, not demand, remains the more frequently cited constraint in earnings commentary. Broadcom’s own visibility into multi-gigawatt deployments and a deep custom-ASIC backlog aligns with this macro picture.

Analytical Assessment: Growth Trajectory vs. Expectation Reset

The data do not support a narrative of AI demand faltering. AI semiconductor revenue growth is accelerating on both a year-over-year and sequential basis, margins are holding despite mix shift, and customer concentration—while real—now spans multiple frontier labs and hyperscalers with multi-year visibility. Networking silicon (Tomahawk/Jericho-class switches and related optics) benefits from the same buildout and provides a more diversified exposure than pure GPU plays.

However, the episode illustrates several structural features of the current AI infrastructure cycle:

1. Diminishing marginal returns to guidance beats. After multiple quarters of upward revisions and parabolic stock performance, the bar for positive surprise has risen dramatically. Reiterating a $100 billion+ 2027 target, even while delivering 143–200%+ growth, registers as conservatism rather than confidence in some quarters.

2. Software drag and segment divergence. The Infrastructure Software business, while stable in absolute terms, is growing far more slowly than semiconductors. This creates a modest but visible drag on overall growth rates and highlights execution risks in the VMware integration.

3. Valuation and sentiment fragility. AI-related names have enjoyed extraordinary multiple expansion on the thesis of multi-year hyper-growth. When growth remains strong but the rate of acceleration or the quantum of upward revision falls short of the most bullish models, capital rotates quickly. The selloff rippled into other semiconductor names, underscoring sector-wide sensitivity.

4. Lumpiness inherent in custom ASIC programs. Large custom chip programs involve lengthy co-development cycles and phased deployments. Revenue recognition can therefore appear lumpy even when underlying demand pipelines remain robust.

Outlook and Implications

Broadcom remains exceptionally well positioned at the intersection of custom AI silicon and high-performance networking—two of the highest-conviction areas of the infrastructure stack. The path to well over $100 billion in annual AI semiconductor revenue by fiscal 2027 is still intact according to management, supported by expanding customer engagements and gigawatt-scale deployments.

That said, the market has signaled that future quarters will be scrutinized not merely for beats but for evidence that growth remains in an accelerating phase rather than transitioning to a high-but-stable plateau. Investors will watch subsequent hyperscaler capex updates, power and data-center build timelines, and any commentary on enterprise inference ROI closely. Margin stability amid rising AI mix and continued software execution will also matter.

For a 20-year veteran observer, this pattern is familiar: periods of extraordinary technological adoption frequently produce sharp sentiment swings even while the underlying capital expenditure wave remains in its early-to-middle innings. Broadcom’s results demonstrate that the wave is still cresting; the stock’s reaction demonstrates that pricing and expectations can crest ahead of it.

The analytical takeaway is therefore nuanced rather than binary. There is no fundamental evidence of an AI infrastructure slowdown in either Broadcom’s numbers or aggregate hyperscaler spending plans. There is, however, clear evidence that the market is repricing the AI enabler complex for a world in which growth remains exceptional but perhaps less parabolic than the most aggressive 2024–2025 models assumed. That distinction will likely define positioning in the sector over the next several quarters.

References

Broadcom Inc. (2026, June 3). Broadcom Inc. announces second quarter fiscal year 2026 financial results and quarterly dividend. https://investors.broadcom.com/news-releases/news-release-details/broadcom-inc-announces-second-quarter-fiscal-year-2026-financial

CNBC. (2026, June 3). Broadcom stock plunges on weak software sales, unchanged AI chip forecast for the year. https://www.cnbc.com/2026/06/03/broadcom-avgo-earnings-report-q2-2026.html

Reuters. (2026, June 4). Broadcom set to shed $300 billion in value as AI results fail to impress. https://www.reuters.com/business/broadcom-tumbles-revenue-miss-clouds-ai-boom-bets-2026-06-04/

Seeking Alpha. (2026, June 4). Broadcom plunges even as Q2 results, guidance top Wall Street’s forecast. https://seekingalpha.com/news/4600327-broadcom-slides-even-as-q2-results-guidance-top-wall-streets-forecast

Yahoo Finance / Goldman Sachs analysis (via secondary reporting). (2026, June). Hyperscaler capex forecasts and aggregate spending estimates for 2026. Multiple contemporaneous reports citing Amazon ($200B), Microsoft ($190B), Alphabet ($175–190B range), and Meta ($115–145B range) guidance.

Bloomberg. (2026, June 3–4). Coverage of Broadcom shares sliding on AI outlook disappointment (multiple dispatches).

Futurum Group. (2026, February; updated context). AI Capex 2026: The $690B infrastructure sprint. Contextual hyperscaler spending analysis.