Broadcom (AVGO) delivered what would normally be considered an outstanding earnings report. Revenue rose 48% year-over-year to a record $22.2 billion, light of expectations, adjusted earnings beat expectations, free cash flow topped $10 billion, AI semiconductor revenue surged 143%, and management issued upside guidance for the July quarter. Yet the stock suffered one of its sharpest post-earnings declines in years, falling from nearly $490 ahead of the report to as low as $409 this morning. The reaction has sparked a debate across Wall Street: is this the beginning of the end for the AI trade, or merely a healthy reset after an extraordinary run?

The answer is probably somewhere in the middle.

To understand the market's reaction, investors first need to appreciate the expectations that had built into Broadcom heading into the report. Shares had rallied from roughly $289 in early April to nearly $490 before earnings. Much of that advance occurred over the previous month as investors extrapolated the blowout results seen from Dell Technologies (DELL), Hewlett Packard Enterprise (HPE), Nvidia (NVDA), and numerous AI infrastructure names. By the time Broadcom reported, investors were no longer looking for a beat. They were looking for a spectacular beat.

Instead, they received something closer to a solid beat.

Broadcom reported adjusted earnings of $2.44 per share versus expectations of roughly $2.40 while revenue of $22.19 billion came in modestly ahead of consensus. AI semiconductor revenue reached $10.8 billion, slightly above management's prior forecast of $10.7 billion. The company also guided fiscal third-quarter revenue to approximately $29.4 billion, comfortably above Wall Street estimates.

Under normal circumstances, those results would have been celebrated.

The problem was that the most important number in the report disappointed relative to investor expectations. Broadcom guided third-quarter AI semiconductor revenue to $16 billion. While that represents more than 200% year-over-year growth, investors had been expecting closer to $17 billion or even $18 billion. Equally important, management did not raise its fiscal 2027 AI revenue framework of more than $100 billion despite enormous optimism surrounding AI infrastructure spending.

That left investors wondering whether the pace of AI spending growth may be starting to normalize.

Another issue centered on margins. One of the more important revelations from the conference call was that AI is actually becoming a margin headwind for Broadcom. Management indicated gross margins will decline to approximately 74% in the third quarter from 77.1% in the second quarter as AI semiconductor products become a larger percentage of revenue. While operating margins are expected to remain stable due to scale and operating leverage, investors received a reminder that not all AI revenue is equally profitable.

This represents a subtle but important shift in the narrative.

For much of the AI boom, investors have assumed that more AI revenue automatically translates into higher profitability. Broadcom's results suggest the reality may be more nuanced. Custom AI accelerators generate tremendous revenue growth, but they carry lower margins than parts of Broadcom's software business and networking portfolio. As AI becomes a larger share of the revenue mix, investors may need to recalibrate some of their profit assumptions.

The conference call also introduced new questions around execution risk. Management guided AI semiconductor revenue to $16 billion next quarter while maintaining a full-year AI revenue target of $56 billion. At the same time, CEO Hock Tan noted that bookings exceeded $30 billion during the quarter compared to $10.8 billion of shipments.

That sounds bullish on the surface. However, analysts immediately began asking whether some deployments are being pushed further into the future.

Tan acknowledged that many customers must coordinate power availability, infrastructure deployment, networking equipment, and other components before systems can be delivered. That raises a concern that has existed beneath the surface of the AI trade for months: demand is clearly enormous, but execution remains complicated.

If deployments begin slipping by quarters rather than accelerating smoothly, investors could become much less willing to pay premium valuations for AI infrastructure stocks.

The reaction from analysts has been fascinating. Despite the stock's decline, many firms actually raised their price targets following the report. Bank of America raised its target to $530. Mizuho lifted its target to $530. Benchmark raised its target to $545.

Those firms generally agree that the long-term AI story remains intact.

Analysts point to Broadcom's six major AI customers, including Google, Meta Platforms (META), OpenAI, and Anthropic. They also note that management continues to forecast AI semiconductor revenue exceeding $100 billion by fiscal 2027 while visibility reportedly extends into 2028.

The issue is not demand.

The issue is whether expectations got too far ahead of reality.

From a technical perspective, Broadcom is approaching a critical area. The stock has effectively erased the gains generated during the four sessions preceding earnings. More importantly, shares are now testing the same $400-$410 region that provided support during the first half of May. The 50-day moving average sits near $396, creating a major technical support zone that many traders will be watching closely.

The broader question facing investors is whether this marks the end of the AI trade itself.

At this stage, that conclusion appears premature.

There are simply too many important catalysts still ahead. Ciena (CIEN) earnings, Apple's (AAPL) WWDC keynote, Oracle (ORCL) earnings, Adobe (ADBE) earnings, the highly anticipated SpaceX IPO, and Micron Technology (MU) earnings all have the potential to reshape investor sentiment over the coming weeks. AI spending trends remain robust, hyperscalers continue investing aggressively, and the largest technology companies are still prioritizing AI infrastructure.

What may be ending, however, is the period where investors could blindly buy any AI-related stock and expect immediate gains.

Options positioning became heavily skewed toward bullish outcomes over the past several months. Call activity dramatically outweighed put activity as investors embraced a near one-way market. Broadcom's reaction serves as a reminder that expectations matter. Even strong results can trigger selling when optimism becomes excessive.

As a result, investors should expect volatility to increase. Demand for downside protection is likely to rise, particularly given how inexpensive puts have remained relative to historical standards. That could create periodic pressure across technology stocks even if the long-term AI story remains intact.

Ultimately, Broadcom's earnings report does not signal the end of the AI trade. What it may signal is the end of the easy phase of the AI trade. The next stage will require companies to consistently execute, justify elevated valuations, and deliver on increasingly ambitious expectations. For traders, that likely means bigger swings, more opportunities on both sides of the market, and a far more challenging environment than the nearly straight-line advance that characterized much of the past several months.

In other words, the AI boom may still be alive and well. But after Broadcom's earnings report, the ride is about to get a lot more interesting.