Broadcom Market Analysis
AI Revenue +143%, Q3 Guidance +200% — AVGO Down 15% After Earnings on “Chips Only” Pivot
Key Questions
Why did Broadcom stock fall 15% after record Q2 earnings?
Broadcom reported Q2 FY2026 revenue of $22.2 billion — up 48% year-on-year — with AI semiconductor revenue of $10.8 billion growing 143% year-on-year. Despite Q3 guidance of $29.4 billion and AI revenue expected to grow over 200% year-on-year to $16 billion in Q3, the stock fell approximately 15–17% from its June 1 all-time high of $459.97. Two factors drove the selloff: Hock Tan announced Broadcom would offer “chips only” rather than the complete integrated AI systems previously promised to customers, and the company did not raise its full-year AI semiconductor revenue guidance beyond the previously stated $56 billion — below some analyst expectations of $60 billion-plus.
Who are Broadcom’s six AI customers and why does it matter?
Hock Tan confirmed on the Q2 earnings call that Broadcom has six core custom chip customers, including Anthropic, Google, Meta, and OpenAI — with two additional unnamed customers whose identities are subject to NDA. These six are driving all of Broadcom’s AI semiconductor revenue growth. The Meta partnership has been extended through 2029. $6 billion in new AI orders were booked from the two additional customers as of the earnings call. Long-term supply deals are in place for multi-gigawatt AI compute, with significant new deployments beginning in fiscal 2027.
What is the Broadcom AI revenue outlook through 2027?
Broadcom reaffirmed full-year FY2026 AI semiconductor revenue guidance at $56 billion — approximately 180% growth year-on-year. For FY2027, management reiterated expectations of more than $100 billion in AI semiconductor revenue, supported by multi-year contracts and major hyperscaler customer orders. Q3 FY2026 AI guidance of $16.0 billion represents over 200% growth year-on-year. The Apollo/Blackstone partnership to deploy over 20 gigawatts of AI compute capacity has a first tranche valued at $35 billion being rolled out by Apollo.
Key Facts
- Current price: ~$380–$400 (−15% from ATH post-earnings)
- 52-week ATH: $459.97 (June 1, 2026)
- YTD performance: +40%
- Q2 FY2026 revenue: $22.2B (+48% y/y)
- Q2 AI semiconductor revenue: $10.8B (+143% y/y)
- Q2 adjusted EBITDA: $15.2B (69% margin)
- Q2 EPS: $2.44 (vs $2.40 estimate)
- Q3 FY2026 revenue guidance: $29.4B (+84% y/y)
- Q3 AI guidance: $16.0B (+200%+ y/y)
- FY2026 AI guidance: $56B (~180% growth)
- FY2027 AI target: >$100B
- AI customers: 6 (Anthropic, Google, Meta, OpenAI + 2)
- Apollo/Blackstone deal: $35B first tranche, 20GW AI compute
- Analyst consensus target: $525–$550
Broadcom occupies a category that the semiconductor industry did not have a name for five years ago: the custom AI accelerator foundry for hyperscalers who need chips that Nvidia does not build. Google’s TPUs, Meta’s MTIA, and the custom silicon programs of OpenAI and Anthropic all require a partner with the design and manufacturing expertise to translate architectural requirements into production silicon at scale. Broadcom is that partner — and in Q2 FY2026, the revenue from that position grew 143% year-on-year to $10.8 billion.
AVGO trades near $380–$400 as of July 2026 — down approximately 15% from its June 1 all-time high of $459.97, after the Q2 earnings call delivered record results but accompanied a “chips only” strategic pivot and an unchanged AI revenue guidance that disappointed investors expecting an upward revision. The stock remains up approximately 40% year-to-date. Q3 FY2026 guidance of $29.4 billion with AI revenue expected to exceed $16 billion — over 200% growth — is among the most extraordinary near-term guidance figures in semiconductor history. The live chart below reflects the current AVGO share price in real time.
The “Chips Only” Pivot and What It Actually Means
The most consequential strategic announcement from Q2 was not the revenue figures — it was Hock Tan’s statement that Broadcom would offer “chips only” rather than the complete integrated AI systems the company had previously indicated it would provide. This reversal from the full-stack AI system approach disappointed investors who had priced in higher-margin, stickier system-level contracts. The practical effect: Broadcom’s value proposition becomes more purely tied to the quality and performance of its custom silicon rather than its ability to deliver turnkey AI infrastructure solutions. That narrows the addressable opportunity per customer but also focuses execution on Broadcom’s clearest competitive strength.
The Meta partnership extension through 2029 is the clearest evidence that the chips-only model is commercially viable at the highest level. Meta — which has one of the most sophisticated in-house AI infrastructure teams in the world — is extending its Broadcom custom chip relationship by years, suggesting the chips-only value proposition is compelling even when the customer has the capability to integrate systems independently. The Apollo/Blackstone partnership to deploy 20 gigawatts of AI compute — with a $35 billion first tranche being rolled out by Apollo — adds a financial infrastructure dimension that connects Broadcom’s chip supply to the capital markets funding the buildout.
The Six Customers: Concentration and Visibility
Broadcom’s AI revenue is entirely generated by six customers. Hock Tan confirmed Anthropic, Google, Meta, and OpenAI are among them — with two additional customers under NDA whose orders totaled $6 billion as of the Q2 call. That concentration is simultaneously the biggest risk and the biggest visibility advantage in the semiconductor space. On the risk side: if any of these six customers reduces custom chip orders, Broadcom’s AI revenue trajectory changes materially. On the visibility side: hyperscaler AI infrastructure programs operate on multi-year planning cycles, and Broadcom has long-term supply deals in place for multi-gigawatt compute deployments starting in FY2027. The $100 billion FY2027 AI revenue target is not speculative — it is supported by booked orders from named customers with signed multi-year contracts.
Infrastructure software — primarily VMware, acquired in 2023 — generated $7.18 billion in Q2, up 9% year-on-year but slightly below the $7.32 billion StreetAccount estimate. The VMware integration has proceeded largely on schedule, and the software segment provides a high-margin recurring revenue base that reduces Broadcom’s dependence on the AI semiconductor cycle for overall profitability. Adjusted EBITDA margins of 69% in Q2 reflect both the semiconductor mix shift toward high-value AI chips and the software segment’s inherent operating leverage.
The Competitive Dynamics That Determine the Long-Term Case
Broadcom’s custom silicon approach is structurally differentiated from Nvidia’s GPU model. Nvidia builds general-purpose accelerators optimized across a wide range of workloads. Broadcom builds chips designed specifically for a single customer’s architecture — delivering superior performance-per-watt for that customer’s specific workloads at the cost of narrower applicability. As hyperscalers become more sophisticated about their AI infrastructure requirements, the willingness to invest in custom silicon increases — which is why Google, Meta, and OpenAI are all maintaining or expanding Broadcom relationships even as they also purchase Nvidia GPUs.
The competitive risk from Marvell, Intel Foundry, and emerging custom silicon design firms is real but not yet material at scale. Broadcom’s 20-plus year relationships with its largest customers, its depth of co-design expertise, and its manufacturing partnerships with TSMC create a moat that is difficult to replicate on any timeframe shorter than five to seven years. The “chips only” pivot simplifies Broadcom’s competitive positioning — it is explicitly not competing with Nvidia on systems integration, allowing it to focus entirely on the custom silicon segment where it has genuine differentiation.
MatrixPro24 Analytical View
Broadcom at $380–$400 in July 2026 is trading at a 15% discount to its June 1 all-time high on results that included 143% AI revenue growth, Q3 guidance of $29.4 billion, and a FY2027 target of over $100 billion. The “chips only” strategic pivot and unchanged $56 billion FY2026 AI guidance — below some expectations — are the specific triggers for the selloff, not any deterioration in the business. Seeking Alpha’s PEG ratio of 0.68 at the post-earnings price suggests the stock is genuinely attractively valued relative to its growth rate — a rare condition for any semiconductor company in the current cycle.
The analyst consensus at $525–$550 implies 31–37% upside from the post-earnings price. Cantor Fitzgerald’s $525 target and the broader analyst community’s buy consensus reflect confidence that the $100 billion FY2027 AI revenue target is achievable given the booked order visibility. The key risk is customer concentration: six customers generating all AI revenue creates a scenario where any reduction in custom chip spending by even one major account would materially impact the growth trajectory. At $380–$400, the market is pricing some probability of that risk materializing — which is precisely what creates the valuation gap between the current price and the analyst consensus.
Three variables to watch through Q3: the Q3 FY2026 results in September as the first test of whether AI revenue actually hits the $16 billion guidance — a figure that would represent the largest single-quarter AI chip revenue in Broadcom’s history, the FY2027 customer order trajectory as the leading indicator for whether the $100 billion target has enough booked support to be credible, and any update on the two unnamed AI customers whose $6 billion in orders represents the most significant new demand signal from the Q2 call.
Sources
- Broadcom Investor Relations
- CNBC — Broadcom Q2 2026 Earnings
- Motley Fool — Q2 2026 Earnings Transcript
- Yahoo Finance — AVGO
- Seeking Alpha — Broadcom Analysis
- U.S. Securities and Exchange Commission (SEC)
- Federal Reserve
- TIKR Financial Data
About MatrixPro24 Editorial Team
Published by MatrixPro24 Editorial Team. Read our Editorial and Content Policy to understand our compliance and brand publishing standards.
MatrixPro24 Editorial Team publishes independent, data-driven market analysis covering cryptocurrencies, stocks, and commodities.
Disclaimer
This analysis is for informational purposes only and does not constitute financial advice. Price data referenced as of July 11, 2026. Past performance is not indicative of future results. Always conduct your own research before making investment decisions.
