Here’s the thing about Broadcom right now. The numbers were actually good. Really good. And the stock still dropped more than 20% from its June peak. That tells you something important about where trader expectations had moved before the Q2 report even landed.
Let’s start with what actually happened.
The Q2 numbers were record-setting across the board. Revenue came in at $22.19 billion, up 48% year-over-year. AI semiconductor revenue hit $10.8 billion for the quarter, up 143% on an annual basis, driven by accelerating demand for custom AI accelerators and networking infrastructure. Free cash flow expanded. The company authorized a $10 billion share repurchase. They declared a quarterly dividend of $0.65 per share.
So why did the stock crater?
CEO Hock Tan didn’t raise the $100 billion AI semiconductor revenue target for 2027. He reiterated it. And at the expectations level AVGO was trading at heading into that report, “reiterate” reads as a miss. Add in some softness on gross margins tied to lower-margin custom AI accelerators, and a report that Google may be shifting more of its next-gen TPU (“TPUv9” / “Triggerfish”) work toward MediaTek, and you get a stock that dropped hard on numbers that, in any other environment, would have been celebrated.
AVGO is now trading near $378, down from a 52-week high of $495. The market cap sits around $1.76 trillion. The 12-month consensus price target from analysts is approximately $493.
The Custom AI Chip Business Is the Story
Broadcom isn’t a traditional chipmaker. What makes it genuinely different is the custom ASIC franchise – the company designs purpose-built AI accelerators for hyperscalers who want to move away from off-the-shelf GPUs for certain workloads. Broadcom has said it has six core custom AI chip customers, but it does not publicly name all of them.
OpenAI’s first custom AI inference chip, called Jalapeño, was developed in partnership with Broadcom. That deal matters structurally – it signals the hyperscaler buildout of proprietary silicon is accelerating, not slowing, and Broadcom is positioned squarely in the middle of it.
Broadcom has discussed an AI backlog of $73 billion over the next ~18 months. For Q3 fiscal 2026, management guided AI semiconductor revenue to be about $16.0 billion, which would represent over 200% year-over-year growth. Q3 total revenue guidance is approximately $29.4 billion. Those are not small numbers.
Slight tangent, but it matters: the reason expectations ran so hot heading into Q2 is that Q1 results were legitimately exceptional. Broadcom said it expected AI semiconductor revenue of $10.7 billion in Q2. When Broadcom delivered the Q2 numbers and held the $100 billion target rather than raising it, the crowded long trade unwound fast.
The Valuation Argument, Honestly
At ~$378, AVGO trades at a very elevated trailing P/E (many data providers currently show it closer to ~98x). That’s a premium valuation. No way around it. The counterargument analysts are making is that forward earnings estimates have been climbing.
Jefferies put out a note yesterday flagging a buying opportunity after the recent pullback and reiterating a Buy rating (with a $550 price target, per reporting). The consensus analyst rating remains overwhelmingly bullish. Average 12-month target sits near $493.
But that doesn’t mean the trade is obvious. Valuation concerns are real when a stock is pricing in flawless execution years into the future. The bear case isn’t that AI spending stops – it’s that hyperscalers eventually bring chip design fully in-house, reducing dependence on Broadcom’s design services. The Google-MediaTek report for the next-gen TPU program is the most concrete version of that risk so far.
Technical Framework
AVGO has retreated from $495 to the $370–$380 zone, which roughly corresponds to where the stock was trading before the initial Q2 guidance beat in March sparked a rally. Watch the 50-day moving average as a reference point for near-term trend health. Volume on the selloff was elevated, which typically suggests institutional repositioning rather than pure retail panic. Holding above the $360–$365 area is key for the bull case to stay structurally intact.
The next major catalyst is Q3 earnings, expected around September 3, 2026. Between now and then, the AI chip delivery cadence and any further news on the Google-MediaTek situation will be the proxies traders watch.
Three Scenarios Worth Modeling
Bull Case: Q3 AI revenue comes in above the $16 billion guide, potentially toward $18 billion, validating the trajectory toward $100 billion in 2027. The stock re-rates toward the mid-$400s as the selloff is confirmed to have been an overreaction. Catalysts include further hyperscaler capex commitments and new custom chip customer announcements.
Base Case: Q3 meets guidance near $16 billion in AI revenue and $29.4 billion in total revenue. The stock stabilizes in the $370–$420 range as the market slowly digests the reset in expectations. The September earnings report becomes the next inflection point.
Bear Case: Google’s shift toward MediaTek for the next-gen TPU program proves to be part of a broader trend of customers reducing Broadcom dependence. Gross margin pressure worsens. The $100 billion 2027 target starts getting questioned. The stock tests $320 or lower if AI capex narratives soften into year-end.
Active Trader Framework
This is a situation where the gap between the fundamental story and the price action is wide enough that positioning discipline matters more than conviction about the outcome. The risk/reward from ~$378 looks asymmetric to bulls given analyst targets, but momentum is clearly against the stock right now. Key levels to monitor: $360 as support, $420 as near-term resistance, $495 as the prior high that now defines full recovery. For options traders, implied volatility may be elevated heading into the September earnings date, which affects the cost of defining risk. Size positions to reflect the uncertainty in the short-term direction, even if the longer-term AI chip thesis remains intact.
The question isn’t whether AI chip demand is real. It clearly is. The question is whether Broadcom’s stock, at roughly $1.76 trillion in market cap, was already pricing in an outcome better than $100 billion in AI semiconductor revenue by 2027 – and whether the September quarter can reset that conversation in the bulls’ favor.
For informational and educational purposes only. Not investment advice. Trading involves risk, including loss of principal.