Broadcom Shares Drop 20.6% From June High Despite $30 Billion AI Bookings

Broadcom stock fell to $393.02 even as AI semiconductor bookings topped $30 billion in the latest quarter. Investors are weighing strong demand, margin pressure, and customer concentration risks.

Broadcom shares fell to $393.02, down 3.40% in the latest session, leaving the stock 20.6% below its June 3, 2026 high of $495.00. The decline came despite one of the strongest demand signals in the AI semiconductor market: more than $30 billion in quarterly AI chip bookings against $10.8 billion of shipments.

The selloff highlights a sharp disconnect between operating momentum and market sentiment. Broadcom remains one of the largest AI infrastructure companies by market value at $1.88 trillion, but investors have become more cautious as Treasury yields rise, semiconductor valuations reset, and concerns build around customer concentration and margin compression.

At the same time, the company continues to post exceptional profitability. Trailing 12-month revenue stands at $75.465 billion, with gross margin of 68.35%, net margin of 38.85%, and EBITDA of $42.063 billion, underscoring why Broadcom remains central to the AI investment debate.

Key Facts

  • Broadcom traded at $393.02, down 3.40%, versus a 52-week high of $495.00 and a 52-week low of $273.00.
  • AI semiconductor bookings exceeded $30 billion in the quarter, compared with $10.8 billion of AI shipments.
  • Second-quarter AI semiconductor revenue rose 143% year over year to $10.8 billion, representing roughly 49% of total revenue.
  • Total quarterly revenue reached a record $22.2 billion, up 48% year over year, while non-GAAP EPS came in at $2.44.
  • Broadcom expects third-quarter AI semiconductor revenue of $16.0 billion and total revenue of about $29.4 billion.

Broadcom AI Bookings and Valuation

The core issue for investors is not whether Broadcom’s AI business is growing. It is. The question is how much of that growth should be valued today when a meaningful portion of demand is tied to deliveries stretching into 2027 and 2028. With the 10-year Treasury yield near 4.60%, markets are assigning less value to long-duration growth stories, even when order visibility is unusually strong.

Broadcom’s latest quarter showed that demand for custom AI accelerators and networking products remains intense. AI semiconductor revenue reached $10.8 billion, up 143% from a year earlier, while total semiconductor revenue climbed to $15 billion. Management also pointed to multi-year agreements and order visibility extending into 2028, with an expectation that annual AI semiconductor revenue could exceed $100 billion starting in fiscal 2027.

Yet the stock has struggled since the company declined to raise that $100 billion target after second-quarter earnings. That decision appears to have amplified doubts about how quickly the current order book will translate into recognized revenue. It also sharpened attention on two risks: dependence on a small number of major hyperscale customers and the lower gross margin profile of custom AI chips versus Broadcom’s software business.

Broadcom is being valued as if its AI order book may not fully convert, even though bookings are already running at nearly three times quarterly shipments.

Why margins and customer mix matter

Broadcom’s AI growth is arriving with a trade-off. Gross margin in the second quarter fell 230 basis points year over year to 77.1%, and the company guided to approximately 74% in the third quarter. The pressure is largely mix-driven: custom ASICs and TPU-related products carry lower margins than infrastructure software, even when unit economics remain healthy.

Customer concentration adds another layer of uncertainty. Google remains Broadcom’s largest custom silicon customer, and reports of MediaTek gaining a role in Google’s AI chip ecosystem have fueled concern that Broadcom’s dominant position may face more pricing pressure over time. For investors, that means future earnings quality matters almost as much as future revenue growth.

Even so, Broadcom is not relying on a single theme. The company announced a long-term strategic commitment with Standard Chartered to support the bank’s global infrastructure modernization through a private cloud foundation. It also has a renewed supply agreement with Apple through 2031 worth more than $30 billion, covering wireless components and likely custom chips tied to artificial intelligence initiatives.

Implications for Investors

For portfolio managers, Broadcom presents a more complicated AI exposure than many pure semiconductor momentum names. On one hand, the stock trades at 24.37 times forward earnings, a level that looks relatively restrained given triple-digit AI revenue growth, record profitability, and a backlog that suggests demand remains far ahead of current shipments. That valuation may appeal to investors seeking AI exposure with substantial free cash flow and an established dividend.

On the other hand, the market is clearly demanding proof that bookings will convert into revenue without materially eroding margins. Investors should watch three markers closely over the next two quarters: whether AI revenue reaches the guided $16.0 billion in the third quarter, whether gross margin stabilizes near the 74% range, and whether management provides greater clarity on the timing of backlog conversion into fiscal 2027 and beyond.

There is also a competitive angle worth monitoring. If large customers such as Google, OpenAI, Anthropic, and Apple continue expanding custom silicon programs with Broadcom, the company could reinforce its position as a leading non-GPU AI infrastructure supplier. If those customers diversify more aggressively to other partners, the stock may continue to trade at a discount despite strong headline growth.

Broadcom remains one of the most closely watched names in the AI infrastructure trade because its results cut across valuation, customer concentration, and margin sustainability. The next stage for the stock will likely depend less on raw demand and more on management’s ability to show that its massive AI backlog can become profitable, durable revenue on schedule.

Ultima Markets