iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Broadcom’s AI semiconductor revenue rose 221% year over year to $16.7 billion in Q3 FY2026. That is a striking result, but it is not a 221% increase in Broadcom’s total revenue—and it does not prove that custom chips are a bigger business than GPUs. The figures show rapid growth in demand for AI accelerators and networking, while leaving the market-wide comparison unresolved.
What Broadcom’s 221% figure measures
The 221% increase applies to Broadcom’s AI semiconductor revenue in Q3 FY2026, compared with the same quarter a year earlier. Broadcom reported $16.7 billion in that category. Its consolidated revenue for the quarter was $29.6 billion, a separate company-wide figure. The two numbers describe different scopes, so the AI growth rate should not be presented as Broadcom’s overall revenue growth. Broadcom’s Q3 FY2026 results
Broadcom attributes the increase to demand for custom AI accelerators and AI networking. The company does not provide a full revenue split between those categories in the cited results. As a result, $16.7 billion is not a disclosed custom-chip-only figure; it includes the company’s reported AI semiconductor revenue.
How Broadcom’s AI revenue changed across recent quarters
Broadcom’s reported AI semiconductor revenue accelerated through Q1–Q3 FY2026. The table pairs each quarter’s revenue with Broadcom’s reported year-over-year growth rate.
#1 Best Overall
- ✅Powered by 26 Tera-Operations Per Second (TOPS) Hailo-8 AI Processor. 2.5W typical power consumption
- ✅Scalable, enabling simultaneous processing of multi-streams & multi-models
- ✅Enabling real-time, low latency and high-efficiency AI inferencing on the edge devices
- ✅Supports TensorFlow, TensorFlow Lite, ONNX, Keras, Pytorch frameworks
- ✅Supports Linux and Windows. Supports the temperature range of -40°C to 85°C
| Fiscal quarter | AI semiconductor revenue | Year-over-year growth |
|---|---|---|
| Q4 FY2025 | Not stated in the cited release summary | 74% |
| Q1 FY2026 | $8.4 billion | 106% |
| Q2 FY2026 | $10.8 billion | 143% |
| Q3 FY2026 | $16.7 billion | 221% |
Broadcom also reported that Q3 AI semiconductor revenue grew 54% sequentially from Q2. These are company-reported results, not an independent estimate of the entire AI-chip market. Sources: Q1 FY2026, Q2 FY2026, Q3 FY2026, and Q4 FY2025 results.
Why custom accelerators and networking are growing
Broadcom CEO Hock Tan has described rising demand for both custom AI accelerators and AI networking as drivers. In the Q1 FY2026 results release, he said AI revenue grew 106% year over year to $8.4 billion, above the company’s forecast, “driven by robust demand for custom AI accelerators and AI networking.” In the Q2 release, he said AI semiconductor revenue of $10.8 billion grew 143%, also above forecast, “driven by increasing demand for custom AI accelerators and AI networking.” Q1 release · Q2 release
Rank #2
- High-Performance AI Processing: The MX3 is designed to handle the most demanding AI computer vision workloads, delivering exceptional performance and efficiency.
- Flexible Integration: The MX3 can be easily integrated into your existing systems via its M.2 M-key form factor and support for Linux operating systems.
- Energy Efficient: The MX3 is designed to provide high performance while minimizing power consumption.
- Comprehensive Software Development Kit (SDK): The MX3 is supported by a comprehensive SDK that simplifies development and deployment.
- Hardware compatability: The MX3 is compatible with the PCI-SIG M.2 M-key 2280 Specification. It can be used with the Raspberry Pi 5 with a M-key 2280 HAT.
Those statements point to two related parts of AI infrastructure: chips designed for particular customers or workloads, and networking equipment that helps connect data-center systems. Broadcom’s disclosures do not quantify how much of its AI semiconductor revenue came from each. The growth therefore supports the conclusion that demand across its AI semiconductor business is increasing, but not a precise estimate of custom accelerators alone.
Does the growth show custom chips could become bigger than GPUs?
It shows that Broadcom’s AI semiconductor business is scaling quickly. It does not establish that custom chips as a market are larger than GPUs, or that they will become larger. Broadcom’s quarterly releases report the company’s own revenue and management’s explanation; they do not provide market-wide revenue totals for custom accelerators and GPUs.
Nor do the reported growth figures establish that custom chips outperform GPUs, cost less, use less power, or suit more workloads. Those comparisons depend on the workload and system being evaluated, and the cited results do not supply comparative performance, power, cost, availability, or software-ecosystem data. Custom accelerators may be an important and rapidly growing part of AI infrastructure without replacing GPUs across the market.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to read Broadcom’s outlook alongside reported results
Company forecasts are guidance, not completed results. In its Q2 FY2026 release, Broadcom expected Q3 AI semiconductor revenue of $16.0 billion. The subsequent Q3 release reported $16.7 billion. The forecast and the reported result are distinct figures from different releases; the later result exceeded the earlier estimate. Q2 outlook · Q3 result
Rank #4
- ✅Powered by 26 Tera-Operations Per Second (TOPS) Hailo-8 AI Processor. 2.5W typical power consumption
- ✅Scalable, enabling simultaneous processing of multi-streams & multi-models
- ✅Enabling real-time, low latency and high-efficiency AI inferencing on the edge devices
- ✅Supports TensorFlow, TensorFlow Lite, ONNX, Keras, Pytorch frameworks
- ✅Supports Linux and Windows. Supports the temperature range of -40°C to 85°C
For readers evaluating the headline claim, the most defensible takeaway is narrower: Broadcom reported unusually fast growth in AI semiconductor revenue, driven by demand it associates with custom accelerators and AI networking. That is evidence of a growing business at Broadcom, not proof of a coming custom-chip victory over GPUs.
Quick Recap
Best Value
- DEEPX DX-M1M NPU: Powered by the DEEPX DX-M1M neural processing unit, purpose-built for efficient on-device AI inference workloads.
- COMPACT M.2 2242 FORM FACTOR: Fits the standard M.2 2242 slot, making it easy to integrate into embedded systems, edge devices, and compact computing platforms.
- EDGE AI ACCELERATION: Designed to accelerate deep learning inference at the edge, enabling real-time AI applications without relying on cloud connectivity.
- RADXA AICORE MODULE: The Radxa AICore DX-M1M delivers a plug-and-play AI compute solution ideal for robotics, smart cameras, and industrial automation.
- WARRANTY AND ORIGIN: Backed by a 1-year manufacturer warranty and crafted with quality components for reliable long-term performance in demanding environments.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
