What You'll Learn
- Why Broadcom's Q2 FY2026 earnings report on June 3 could be the most important semiconductor event of the summer
- How AI chip revenue is projected to hit $10.7 billion — a 140% year-over-year jump that dwarfs Nvidia's growth rate
- The six hyperscaler customers fueling Broadcom's $73 billion backlog and what their orders mean for the stock
- Whether VMware's software margins can sustain Broadcom's 60%+ operating profitability as semiconductor cycles shift
Broadcom's Q2 Earnings: The Numbers That Matter
When Broadcom reports its fiscal second-quarter results after the market closes on Wednesday, June 3, Wall Street will be watching one number above all else: AI semiconductor revenue. The company guided for $10.7 billion in AI chip sales for the quarter — a staggering 140% increase from the $4.47 billion it posted in the same period last year. If Broadcom delivers on that target, it will cement its position as the second-largest AI hardware provider behind Nvidia.
The consensus among 37 analysts calls for adjusted earnings of $2.40 per share on total revenue of $22.04 billion. That would represent roughly 51% year-over-year EPS growth on a 43% revenue jump from the $15.4 billion reported in Q2 FY2025. The EPS consensus has climbed 11.1% over the past 90 days, rising from $2.16 to the current $2.40 target — a clear signal that Wall Street is increasingly confident in Broadcom's trajectory.
For context, Broadcom posted Q1 FY2026 revenue of $19.31 billion, up 29.5% year-over-year, with AI semiconductor sales reaching $8.40 billion and growing 106%. The sequential acceleration from $8.4 billion to a guided $10.7 billion in a single quarter underscores the velocity of AI infrastructure spending across the global hyperscaler ecosystem.
The AI Custom Chip Revolution: Why Broadcom Is Different from Nvidia
Unlike Nvidia, which sells standardized GPU platforms, Broadcom operates as the architect of custom AI silicon. The company designs and co-develops application-specific integrated circuits (ASICs) tailored to each hyperscaler's unique workloads. Google's Tensor Processing Units, Microsoft's Maia chips, Meta's MTIA accelerators — all trace back to Broadcom's engineering teams.
This business model creates deep customer lock-in and recurring revenue streams that general-purpose chip designers cannot replicate. As semiconductor supply chains become increasingly geopolitically sensitive, Broadcom's position as a trusted co-design partner to America's largest tech companies becomes even more valuable.
The economics are compelling. For hyperscalers operating at sufficient scale — typically deploying more than 100,000 accelerators — custom chips can deliver 30-50% lower total cost of ownership compared to general-purpose GPUs for specific workloads. Google's TPU, for example, is optimized for the transformer architectures that power its Gemini model family, enabling higher throughput per watt than a comparable Nvidia GPU deployment for that specific use case.
| Metric | Q1 FY2026 (Actual) | Q2 FY2026 (Guided/Expected) |
|---|---|---|
| Total Revenue | $19.31 billion | $22.04 billion |
| AI Semiconductor Revenue | $8.40 billion (+106% YoY) | $10.70 billion (+140% YoY) |
| Non-AI Semiconductor Revenue | $4.10 billion (flat YoY) | ~$4.10 billion (+4% guided) |
| Infrastructure Software Revenue | $6.80 billion (+1% YoY) | ~$7.20 billion (+9% YoY est.) |
| Adjusted EPS | $2.05 | $2.40 |
The Six Hyperscaler Customers Powering a $73 Billion Backlog
Broadcom has confirmed six major custom chip (XPU) customers as of its Q1 FY2026 earnings call: Google, Meta, OpenAI, Anthropic, Apple, and Microsoft. Apple's addition was a new 2026 disclosure not previously public, and it signals that even the world's most secretive hardware company is turning to Broadcom for custom AI silicon.
Google remains the longest-standing partner, with seven generations of co-designed TPUs since 2014. The relationship has deepened into a $46 billion, five-year supply agreement running through 2031, as revealed in an SEC filing. OpenAI signed a multi-year collaboration in October 2025 for 10 gigawatts of custom accelerators, with first deployment targeting the second half of 2026 using both 3nm and 2nm designs.
CEO Hock Tan has stated the company has "line of sight" to more than $100 billion in AI chip revenue in FY2027, backed by a disclosed $73 billion committed customer backlog. Market share estimates now place Broadcom at 70%+ of the custom AI accelerator design services market, up from the 60-80% range Bloomberg Intelligence flagged earlier in 2026.
Counterpoint Research projects that AI server compute ASIC shipments among the top 10 hyperscalers will triple between 2024 and 2027, with the top 10 hyperscalers deploying more than 40 million AI server compute ASIC chips cumulatively during 2024-2028. Broadcom and Marvell together control an estimated 95% of the custom AI ASIC co-design market, but Broadcom's six-customer roster and multi-generational relationships give it a dominant edge.
VMware: The Quiet Profit Engine Behind the AI Story
While AI chips grab the headlines, Broadcom's Infrastructure Software division — anchored by its 2023 VMware acquisition — is the unsung profit engine. The software segment now represents 43% of total revenue and is growing at double-digit rates, driven by rapid adoption of VMware Cloud Foundation (VCF).
Enterprise demand for hybrid cloud flexibility has pushed subscription adoption higher, lifting software gross margins to 86% and pushing overall operating margins above 60%. Management has guided for FY2026 software revenue exceeding $30 billion, with operating income from this segment expected to reach $17 billion. VMware's total contract value bookings reached $9.2 billion in Q1 alone, with annual recurring revenue up 19%.
The strategic logic is clear: as enterprises scale generative and agentic AI workloads, operational complexity increases demand for virtualization and private cloud control planes. Broadcom has bundled VCF with AI data infrastructure, creating a dual-revenue engine that stabilizes performance beyond semiconductor cycles. Cost synergies from the integration continue to boost profitability, and the transition from 8,000+ VMware SKUs to a subscription-first model is generating predictable recurring revenue.
Networking Chips: The Hidden Growth Driver
Broadcom's Jericho3-AI Ethernet switch represents the company's aggressive push into AI data center networking — a segment where it's directly challenging Nvidia's InfiniBand dominance. The Jericho3-AI BCM88890 delivers 28.8 terabits per second of switching capacity, supporting up to 18x 800GbE or 72x 200GbE ports.
As AI clusters scale to 500,000 or a million processors, networking is expected to grow from 5-10% of data center chip spending to 15-20%, according to CEO Hock Tan. Broadcom's Ethernet-based approach offers a cost and flexibility advantage over InfiniBand for many hyperscaler deployments, particularly as the industry moves toward Ultra Ethernet for AI workloads.
The Tomahawk series, especially the Tomahawk 6, remains the most widely deployed data center switching silicon globally. Combined with Broadcom's custom ASIC design services, the networking business creates a vertically integrated AI infrastructure stack that few competitors can match.
Analyst Sentiment and Price Targets
Wall Street remains overwhelmingly bullish on Broadcom ahead of the June 3 report. The stock is pressing against its 52-week high of $442.36, and multiple analysts have raised price targets in recent weeks:
| Firm | Price Target | Rating |
|---|---|---|
| Citi | $500 | Buy |
| Morgan Stanley | $470 | Overweight |
| Susquehanna | Raised (ahead of Q2) | Positive |
| Oppenheimer | Bullish | Outperform |
For the full fiscal year, analysts expect AVGO to report EPS of $9.95, up 76.7% from $5.63 in fiscal 2025. The full-year revenue forecast has climbed from $86 billion to $94.7 billion, reflecting the accelerating AI semiconductor ramp and continued VMware software momentum. With operating margins near 41% and LTM revenue of $64 billion, Broadcom's scale advantages are compounding.
Risks and What Could Go Wrong
Despite the bullish setup, several risks could pressure Broadcom's stock after the report. First, AI semiconductor concentration is a double-edged sword. If any major hyperscaler pushes out orders or reduces deployment timelines, Broadcom's revenue guidance could disappoint. The company has historically provided detailed financial targets during earnings calls, and any signs of demand normalization or deployment delays could materially impact the stock.
Second, Nvidia's growing networking ambitions pose a competitive threat. Nvidia has been expanding beyond GPUs into Ethernet and InfiniBand networking, directly competing with Broadcom's switching business. While Broadcom's Jericho3-AI offers compelling performance, Nvidia's bundled GPU-plus-networking strategy could win share in greenfield AI data center deployments.
Third, geopolitical risk remains elevated. U.S.-China trade tensions, potential export restrictions on advanced AI chips, and the broader semiconductor supply chain disruptions highlighted by the EU's emergency chip seizure powers could affect Broadcom's access to key manufacturing capacity at TSMC.
Finally, the stock is priced for perfection. Trading near its 52-week high with an EPS consensus that has already been revised upward 11%, any miss on revenue or guidance could trigger a sharp correction. Historically, AVGO's earnings date notices have produced average moves of about -0.93%, but a guidance miss in the current AI euphoria could be far more severe.
What Investors Should Watch on June 3
The most critical data point will be Q3 FY2026 guidance. If Broadcom raises its AI semiconductor revenue target above the current trajectory, it would validate the $100 billion FY2027 AI revenue ambition and likely send the stock higher. Conversely, any flattening of the AI growth curve — even if Q2 numbers beat — could spark concern about the sustainability of hyperscaler spending.
Investors should also listen for updates on the OpenAI deployment timeline. First shipments of custom accelerators are targeted for H2 2026, and any acceleration or delay would signal the pace of the next wave of AI infrastructure buildout. Updates on the Google TPU supply agreement and new customer additions would also be material.
On the software side, watch for VMware Cloud Foundation adoption metrics and any commentary on enterprise AI workload migration. If VCF subscription growth accelerates, it would reinforce the thesis that Broadcom's software business is entering a secular growth phase alongside the AI semiconductor boom.
The earnings call is scheduled for 5:00 PM ET on June 3. With the stock near all-time highs and AI spending showing no signs of slowing, Broadcom's Q2 report could set the tone for the entire semiconductor sector heading into the summer.
For investors holding Broadcom or considering a position, the June 3 report represents a high-stakes moment. The AI custom chip revolution is still in its early innings, and Broadcom sits at the center of it. But at current valuations, the margin for error is razor thin.
Last Updated: June 01, 2026 | Source: Broadcom Inc. (Official Earnings Release), MarketBeat, Alphabet Street, 247 Wall Street