Broadcom (AVGO) reports fiscal Q3 2026 earnings after the close on Wednesday, Sept. 2. The numbers will almost certainly be extraordinary. The real question, and Morgan Stanley spelled it out, is whether extraordinary is enough.
That’s the setup heading into Q3. Not “Can Broadcom deliver?” but whether it can clear a bar that some investors have quietly set at more than $150 billion in AI revenue for fiscal 2027. Morgan Stanley models $120 billion.
The gap between those two numbers is where near-term volatility lives. Remember, Q2 was a great reminder that very strong results are not necessarily enough when expectations are high.
The firm maintains its Overweight rating on AVGO, according to a note shared with me at TheStreet. But it’s worth reading the note’s headline: “The main risk into the print is expectations rather than fundamentals.”
Also Read: Broadcom Inc. Latest News and Stories
What Morgan Stanley models for Broadcom’s Q3F26, and what it would take to impress
The July quarter estimates from Morgan Stanley are detailed and sit roughly in line with Street consensus, according to the note.
- Revenue is modeled at $29.4 billion, up 84.3% year over year (YoY) and 32.5% quarter over quarter (QoQ).
- AI revenue specifically is modeled at $16.0 billion — up 48% sequentially — split between $10.8 billion in custom ASIC revenue and $5.2 billion in AI networking.
- Gross margin estimate sits at 74.0%, slightly above the Street’s 73.5%.
- EPS estimate is $3.24, a touch ahead of consensus at $3.22.
That’s an impressive quarter by any historical measure. But take a second to look at this other one.
Broadcom already posted AI semiconductor revenue of $10.8 billion, up 143% YoY in Q2 fiscal 2026, and CEO Hock Tan guided Q3 AI revenue to grow “over 200 percent year-over-year to $16.0 billion,” according to Broadcom’s Q2 report.
Related: BMO sees writing on the wall for Broadcom stock after earnings
So $16 billion in Q3 AI revenue isn’t really upside. I see it as the floor management already set.
For October quarter guidance, Morgan Stanley models revenue of $34.8 billion, up 93.4% YoY, with AI revenue accelerating another 32% sequentially to $21.2 billion, according to the note.
Now that’s the number where real upside surprise could live.
The 2027 debate that’s actually driving Broadcom’s valuation
A more important conversation sits here, and it’s not about this quarter at all.
Broadcom previously guided fiscal 2027 AI revenue “well above” $100 billion, according to Reuters. Management sounded increasingly confident last quarter, including signals that growth should persist “well into 2028.”
Morgan Stanley models roughly $120 billion for fiscal 2027, according to the note. Some investor expectations have drifted to $150 billion or higher.
More AI Stocks:
- Wells Fargo strongly resets Marvell stock target before earnings
- Marvell’s $120 billion deal with Google has fine print
- Analyst rethinks Nvidia stock ahead of earnings
I don’t see the spread between $120 billion and $150 billion as just a forecasting debate. It’s a valuation debate. At $120 billion, AVGO looks reasonably priced at current levels. At $150 billion, the stock looks cheap.
If Sep. 2 guidance implies $120 billion, investors expecting $150 billion will be disappointed, regardless of how good the quarter actually was.
Even Morgan Stanley mentioned this dynamic in the note: “The underlying business can continue to perform exceptionally well without necessarily exceeding the most aggressive expectations.”
That’s a politely worded warning about setup risk.
The TPU supplier debate and why Morgan Stanley isn’t worried
The other overhang question worth addressing is whether Google is diversifying its custom chip suppliers away from Broadcom.
Recent reports have pointed to MediaTek participation in Tensor Processing Unit (TPU) programs, AMD involvement in TPU v10, and Marvell‘s warrant agreement with Google as signals that the hyperscaler is broadening its ecosystem.
Morgan Stanley’s position is unchanged, according to the note. Supplier diversification is real, but Broadcom’s incumbency advantage is very strong. The firm expects Broadcom to retain roughly 80% of its longer-term TPU opportunity, even as Google adds alternative suppliers.
The analyst’s framing was that “If anything, the number of semiconductor companies positioning around TPU underscores the scale of the opportunity.”
That’s an important reframe. When multiple major chip companies are racing to win a slice of a single customer’s custom silicon program, it doesn’t mean the opportunity is shrinking. It means the opportunity is large enough that everyone wants in.
AVGO sits as Morgan Stanley’s second-ranked AI compute name, behind only Nvidia (NVDA), which the firm calls its top pick, according to the note.
The sector backdrop that makes the Q3F26 print matter beyond Broadcom
There’s one more piece of context worth noting. The semiconductor sector just reported 142% YoY earnings growth in Q2 2026. In fact, this sector was the largest contributor to Information Technology sector earnings growth, based on FactSet data as of Aug. 28, 2026.
If semiconductors were excluded from the IT sector calculation, blended IT earnings growth would fall from 75.3% to 38.3%. That’s how dominant the chip cycle has been.
Broadcom is at the center of that cycle. AVGO shares were up 7.40% year to date compared to the S&P 500’s 12.28% gain as of the close of August, according to Yahoo Finance.
The stock has returned 25.44% over the past year and 316.02% over three years. Year-to-date underperformance relative to the index reflects the post-Q2 sell-off that followed an earnings beat. It’s a perfect illustration of the expectations risk Morgan Stanley is flagging again heading into Q3.
The business is exceptional. The bar is high. Those two things can both be true simultaneously, and on Sept. 2, we will find out which one matters more.
Related: Morgan Stanley: Broadcom bears are wrong about Google TPU
