Broadcom (AVGO.US) is set to release its fiscal 2026 third-quarter earnings after the US market closes on September 2, which will be the early morning of September 3 Beijing time. While the company's financial results are almost certain to be exceptionally strong, Morgan Stanley has pointed out that the true question is whether simply being "strong" will be enough. This is the situation heading into the third quarter. The core issue is not whether Broadcom can deliver good numbers, but whether it can clear the high bar privately set by some investors, which involves raising its fiscal 2027 AI revenue expectations to over $150 billion. In comparison, Morgan Stanley's own forecast stands at $120 billion. The risk of short-term volatility lies precisely in the gap between these two figures. It is worth remembering that the second quarter served as a clear warning: when market expectations are overly high, even very robust results may fail to satisfy investors.

Morgan Stanley has maintained its "Overweight" rating on Broadcom. However, the title of its research report is particularly noteworthy: "The main risk heading into earnings is expectations, not fundamentals." The report shows that Morgan Stanley's forecasts for the quarter ending in July are highly detailed and largely align with Wall Street consensus estimates. It projects revenue of $29.4 billion, representing an 84.3% year-over-year increase and a 32.5% sequential rise. AI-specific revenue is forecast at $16 billion, up 48% sequentially, which includes $10.8 billion from custom ASIC business and $5.2 billion from AI networking. The gross margin forecast is 74.0%, slightly above the Wall Street consensus of 73.5%. Earnings per share are projected at $3.24, marginally above the consensus estimate of $3.22.

By any historical standard, this would be a remarkable quarter. However, Morgan Stanley believes investors should take a moment to examine the following data. Broadcom had already recorded $10.8 billion in AI chip revenue in the second quarter of fiscal 2026, a 143% year-over-year increase. According to the company's Q2 earnings report, CEO Hock Tan had provided third-quarter guidance at that time, indicating that AI revenue would "grow over 200% year-over-year to reach $16 billion." Therefore, the $16 billion AI revenue figure for the third quarter does not represent a surprise; it is merely the performance baseline that management had already established. The report notes that for the fourth quarter ending in October, Morgan Stanley forecasts revenue of $34.8 billion, a 93.4% year-over-year increase, with AI revenue accelerating another 32% sequentially to reach $21.2 billion. That figure is where the potential for genuine positive surprises lies.

There is also a more important point of discussion regarding the fiscal 2027 debate that truly determines Broadcom's valuation, and it is completely unrelated to this quarter's performance. Broadcom had previously set its fiscal 2027 AI revenue guidance at a level "well above" $100 billion. In the last quarter, management displayed growing confidence and signaled that this growth momentum would continue "well into 2028." According to the report, Morgan Stanley projects fiscal 2027 AI revenue at approximately $120 billion. However, some investors' expectations have gradually risen to $150 billion or even higher. The gap between $120 billion and $150 billion is not merely a debate over forecasting methods; it represents a divergence in valuation logic. At $120 billion, Broadcom appears reasonably valued at current share price levels, but at $150 billion, the stock would seem very inexpensive. If the guidance implied by the September 2 announcement points to $120 billion, then investors expecting $150 billion would likely be disappointed, regardless of how strong the actual quarterly results are. Even Morgan Stanley acknowledged this dynamic in its report: "The underlying business can continue to perform exceptionally well, yet still fail to surpass the most aggressive expectations." This is a politely worded warning to investors about the downside risk in this situation.

The TPU Supplier Question is another unresolved issue worth addressing: is Google diversifying its custom chip suppliers away from Broadcom? Recent reports have indicated that MediaTek is involved in the Tensor Processing Unit (TPU) project, AMD is participating in TPU v10, and Marvell Technology has entered into a warrant agreement with Google. These developments have been interpreted as signals that the hyperscaler is expanding its ecosystem. According to the report, Morgan Stanley's position remains unchanged. Supplier diversification does exist, but Broadcom's first-mover and established advantages are extremely solid. The firm expects Broadcom to retain approximately 80% of its long-term TPU market opportunity even if Google introduces alternative suppliers. The analyst's framework is: "If anything, the fact that multiple semiconductor companies are positioning around the TPU further highlights the enormous scale of this market opportunity." This is a very important shift in perspective. When multiple leading chip companies are competing for a share of the same customer's custom chip project, it does not mean the market space is shrinking; rather, it indicates that the market is large enough that everyone wants a piece of it. The report indicates that in Morgan Stanley's ranking of AI computing stocks, Broadcom holds the second-highest position, just behind its top pick, Nvidia (NVDA.US).

The Industry Context Beyond Broadcom's Quarterly Results is also worth noting. Based on FactSet data as of August 28, 2026, the semiconductor industry just reported a 142% year-over-year increase in second-quarter 2026 profits. In fact, the sector was the largest contributor to overall profit growth in the IT industry. If the semiconductor segment were excluded from the IT industry calculations, the combined profit growth rate for the IT sector would plummet from 75.3% to 38.3%. This illustrates the dominant position of the chip industry cycle. And Broadcom sits at the very core of this cycle. Based on data through the end of August, Broadcom's stock has risen 7.40% year-to-date, while the S&P 500 index has gained 12.28% over the same period. Over the past year, the stock has delivered a return of 25.44%, and over the past three years, a return of 316.02%. The stock's underperformance relative to the broader market this year reflects the sell-off that followed the better-than-expected second-quarter earnings report. This perfectly exemplifies the "expectations risk" that Morgan Stanley is again warning about ahead of the third-quarter results. The business is exceptionally strong, and the bar is also set very high. These two things can be true simultaneously. And on September 2, investors will learn which force ultimately prevails.