Broadcom is emerging as a core AI infrastructure supplier, with record revenue, surging AI semiconductor sales, and powerful operating leverage. The main debate is whether its premium valuation and heavy leverage are justified by the growth runway.
Investment thesis: Broadcom (AVGO) is becoming a core supplier of AI infrastructure through custom accelerators, Ethernet networking, optical connectivity, and VMware-based private cloud software. Fiscal Q2 2026 revenue reached a record $22.2B, up 48% year over year, while AI semiconductor revenue rose 143% to $10.8B. Management guided for Q3 revenue of $29.4B, up 84%, including $16.0B of AI semiconductor revenue.
The business has strong operating leverage. Q2 operating margin reached 67.3%, adjusted EBITDA was 69% of revenue, and free cash flow was $10.3B. The counterweight is valuation and leverage: trailing P/E is 60.4x, EV-to-revenue is 24.6x, debt is $64.9B, and net debt is about $49.0B. A Buy rating fits a medium-term, moderate-risk investor who can accept semiconductor volatility in exchange for exposure to a powerful AI spending cycle.
The report's fair value estimate of $500 reflects strong AI growth, a high-margin software base, and Broadcom's customer relationships, offset by customer concentration, supply-chain dependence, debt, and the possibility that AI investment eventually cools.
Broadcom Inc. (AVGO), headquartered in Palo Alto, California, designs semiconductor devices and supplies infrastructure software worldwide. Founded in 1961, the company had approximately 33,000 employees and was led by President and CEO Hock Tan in the supplied company information.
Broadcom reports two operating segments: Semiconductor Solutions and Infrastructure Software. Fiscal 2025 revenue was $63.9B, split between $36.9B from Semiconductor Solutions and $27.0B from Infrastructure Software. That mix gives AVGO more recurring software exposure than a pure semiconductor company, while its semiconductor portfolio captures the faster growth in AI data-center infrastructure.
Semiconductor Solutions generated $15.0B in Q2 fiscal 2026 revenue, up 79% year over year. AI semiconductors contributed $10.8B, or 49% of total company revenue, while non-AI semiconductors generated $4.2B, up 6%. Semiconductor operating margin was approximately 62%, up 460 basis points year over year.
The AI semiconductor portfolio includes custom XPUs, Ethernet switches, PCI Express switches, SerDes, optical DSPs, lasers, network interface controllers, and fabric solutions. Broadcom expects Q3 semiconductor revenue of $20.5B, up 124%, with AI semiconductor revenue of $16.0B, up more than 200%, and non-AI semiconductor revenue of approximately $4.5B, up 12%.
Infrastructure Software produced $7.2B in Q2 revenue, up 9% year over year. Annual recurring revenue grew 17%, software gross margin was 93%, and software operating margin was approximately 79%, up 310 basis points. Q3 software revenue guidance of approximately $8.9B represents 31% year-over-year growth.
This segment includes VMware Cloud Foundation, private AI, application networking and security, mainframe software, cybersecurity, AIOps, automation, and Fibre Channel storage management. Its revenue growth is slower than AI semiconductors, but its margins and recurring revenue characteristics improve the stability of the overall company.
Get AI research on any stock
Instant reports, daily intelligence, and an AI analyst in your pocket.
Broadcom's flagship growth platform is its custom AI silicon and networking stack. Q2 bookings for AI semiconductors exceeded $30B against $10.8B of shipments, a substantial demand signal tied to XPUs and networking. Management expects fiscal 2026 AI semiconductor revenue of $56B, up approximately 180% from fiscal 2025, and expects fiscal 2027 AI semiconductor revenue to exceed $100B.
The networking layer is especially important because AI clusters need high-speed links inside racks, between racks, and across data centers. Broadcom is shipping the Tomahawk 6 100-terabit Ethernet switch, is taping out a 200-terabit successor, and supplies Jericho3 and Jericho4 fabric solutions for large hyperscaler deployments.
VMware Cloud Foundation 9.1 is the flagship software update. It supports GPUs and CPUs from NVIDIA, AMD, and Intel and is designed to run AI, Kubernetes, and traditional virtualized workloads on a common private-cloud environment. That positioning connects VMware directly to enterprise AI adoption without requiring Broadcom to sell the underlying compute hardware.
Broadcom's advantage comes from combining specialized intellectual property with deep customer integration. Its AI products are not a single chip. They form a connected portfolio spanning XPUs, Ethernet, SerDes, optical components, PCI Express, switches, and routing fabrics. This increases the value of each design win because a customer can source several critical infrastructure components from one supplier.
The company has also secured multi-generation relationships with Google and Meta, while agreements involving Anthropic and OpenAI extend its exposure to AI model developers. Broadcom said it has purchase orders totaling $6B from the newer customer programs and is creating an AI XPV platform with Apollo, Blackstone, and other investors to deploy more than 20 gigawatts of compute capacity through 2028.
The software moat is different but complementary. VMware, mainframe, cybersecurity, and enterprise infrastructure products sit close to mission-critical workloads. The 2025 10-K identifies switching costs, customer relationships, intellectual property, and competition from internal customer resources as important factors in the business.
Broadcom operates as a fabless semiconductor company and depends on external foundries, packaging providers, and other suppliers. That model limits ownership of manufacturing assets but creates dependence on advanced wafer and packaging capacity at a time when AI demand is rising quickly.
Q2 inventory was $4.3B, and days of inventory on hand rose to 86 from 68 in Q1 as Broadcom secured supply for the second half of fiscal 2026. Management said supply for 2026, 2027, 2028, and 2029 was being secured and that customers had been approaching the company for incremental capacity.
The higher inventory position supports the $16.0B Q3 AI semiconductor revenue guide, but it also raises working-capital exposure if customer delivery schedules change. The 10-K separately identifies supplier concentration, manufacturing interruptions, trade restrictions, and semiconductor cyclicality as material business risks.
The semiconductor market is entering an unusually large AI-led expansion. Gartner forecasts worldwide semiconductor revenue of $1.32T in 2026, up 64%, followed by $1.55T in 2027. Gartner also expects AI semiconductors to represent approximately 30% of total semiconductor revenue in 2026 and hyperscaler AI infrastructure spending to rise more than 50%.
Broadcom is positioned in the infrastructure layer of that market. Its Q2 AI revenue of $10.8B and Q3 guide of $16.0B show direct participation in custom accelerators and the networking required to connect them. The company's disclosed customer programs include 10 gigawatts of OpenAI deployment through 2029, a multi-generation Meta partnership, and additional TPU-related programs with Google and Anthropic.
The non-AI market is also improving. Broadcom reported Q2 non-AI semiconductor bookings above $6B and cited strength in broadband, server storage, and enterprise networking. That recovery provides a second growth leg, although seasonal wireless weakness remained visible during the quarter.
Like what you're reading?
Get full access to AI-powered research reports, market analysis, and portfolio tools.
Broadcom serves hyperscalers, AI model developers, enterprise technology customers, telecommunications companies, storage providers, and industrial customers. The Q2 transcript identifies Google, Anthropic, OpenAI, and Meta among six core AI customers. These relationships involve multi-generation products, large deployment commitments, and significant engineering coordination.
Google has a long-term agreement covering multiple generations of TPUs and AI networking. Anthropic is tied to more than 1 gigawatt of Broadcom TPU-based compute in 2026 and another 5 gigawatts beginning in 2027. OpenAI has a 1.3-gigawatt 2027 commitment within a larger 10-gigawatt agreement, while Meta is expected to deploy 3 gigawatts through the end of 2028.
The customer mix creates both strength and risk. Large customers can fund ambitious infrastructure programs and support design wins, but the 2025 10-K identifies the loss of a significant customer and changes in customer demand timing as material risks. VMware expands the customer base into enterprises, with private-cloud products aimed at organizations running AI and traditional workloads together.
Broadcom competes with NVIDIA (NVDA), AMD (AMD), Marvell Technology (MRVL), Intel (INTC), Qualcomm (QCOM), Cisco Systems (CSCO), and other semiconductor and infrastructure providers. The competitive field changes by product: NVIDIA is strongest in GPUs and AI platforms, AMD competes across CPUs, GPUs, and rack-scale systems, and Marvell competes in custom silicon and connectivity.
Broadcom's distinction is its concentration in custom accelerators, Ethernet networking, optical interconnect, and infrastructure software. It does not need to displace NVIDIA's general-purpose GPU platform to benefit from AI spending. Instead, its products can supply the custom compute and data movement surrounding GPU and XPU clusters.
In software, VMware Cloud Foundation competes with Microsoft (MSFT), Red Hat, and other private-cloud and hybrid-cloud platforms. VMware's installed base creates switching friction, while Broadcom's Q3 software guidance of $8.9B shows that the business is being monetized alongside the AI semiconductor ramp. Competition remains serious because VMware customers can evaluate alternative platforms when pricing, product packaging, or support terms change.
AI infrastructure spending is the dominant macro force for AVGO. Gartner forecasts a 64% increase in global semiconductor revenue in 2026, while hyperscaler AI infrastructure spending is expected to rise more than 50%. Those figures support the near-term demand environment reflected in Broadcom's 84% Q3 revenue growth guide.
Supply-chain localization is also reshaping the sector. The Semiconductor Industry Association reported more than 100 U.S. chip projects across 28 states, representing over $500B of private investment and an expected tripling of U.S. chipmaking capacity by 2032. Broadcom benefits from greater infrastructure investment, but its fabless model still leaves it exposed to foundry, packaging, and geopolitical bottlenecks.
Trade restrictions and export controls remain direct risks because Broadcom sells advanced semiconductors internationally. Deloitte identifies geopolitical fragmentation, supply-chain resilience, and talent shortages as major semiconductor industry issues. Gartner also highlights memory price volatility, including projected 2026 DRAM price growth of 125% and NAND price growth of 234%, which can redirect customer budgets and delay non-AI demand.
Broadcom carries $64.9B of debt and about $49.0B of net debt, so the balance sheet is solid but still a real constraint on flexibility.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →Q2 revenue hit a record $22.2B and operating margin reached 67.3%, showing exceptional earnings power as AI semiconductors surged 143% to $10.8B.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →Management guided Q3 revenue to $29.4B, with AI semiconductor revenue expected to jump to $16.0B and total semiconductor revenue to $20.5B.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →Broadcom trades at 60.4x trailing earnings and 24.6x EV-to-revenue, leaving little room for disappointment despite the growth surge.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →The report’s fair value estimate is $500, reflecting strong AI demand, VMware’s recurring software base, and the offset from debt and customer concentration.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →Broadcom has moved beyond the traditional semiconductor cycle. Its Q2 fiscal 2026 results show a company scaling custom AI silicon and networking at exceptional speed while maintaining operating margin above 67% and quarterly free cash flow above $10B. VMware adds a high-margin software base that broadens the earnings model.
The investment case is strongest when viewed as a compound platform: AI accelerators create growth, networking increases content per cluster, and infrastructure software adds recurring revenue and margin support. The risks are equally concrete. Broadcom carries approximately $49.0B of net debt, relies on external manufacturing and advanced packaging, faces intense competition from NVIDIA, AMD, Marvell, Microsoft, and Red Hat, and depends on a small group of very large AI customers.
A Buy rating is justified for a medium-term investor who values durable infrastructure exposure and can withstand sharp price swings. The $500 fair value estimate recognizes Broadcom's leadership without assuming that every AI spending forecast arrives on schedule. In markets, that distinction is often the difference between owning a strong business and overpaying for a perfect story.
Get AI-powered research reports, daily market intelligence, and a personal analyst in your pocket.
Get Full Access →Not ready to subscribe? ·
Cancel anytime
Broadcom Inc. (AVGO) drops after a strong fiscal Q3 earnings beat, as investors focus on a Q4 revenue guide that fell short of lofty AI expectations. The company still posted rapid AI growth, but the stock is repricing around valuation and future acceleration.
Broadcom Inc. (AVGO) drops after Reuters reported Marvell will help Google develop custom AI chips, raising fresh competition concerns. The selloff comes despite Broadcom’s strong quarterly growth, including surging AI semiconductor revenue and a recent earnings beat.
AVGO’s 5.9% selloff looks like an expectations reset, not a demand collapse. With AI semiconductor revenue up 143% and Q3 guidance at $16 billion, we see the Sept. 3 earnings report as a catalyst for the bullish case.