The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Compare NVIDIA with AMD and Broadcom by first defining what each company counts as AI-related revenue, then checking growth, profitability, cash generation, concentration, execution risks, policy exposure and valuation on a like-for-like basis. Their latest reported figures are not directly interchangeable: the companies use different segment boundaries and their quarters ended on different dates.
What do the latest reported figures actually measure?
The figures below are useful starting points, not a ranking of AI-chip sales or market share. NVIDIA reports a broad Data Center business that includes data-center products and systems, including networking. AMD reports a Data Center segment. Broadcom’s AI semiconductor revenue focuses on custom AI accelerators and AI networking. Each measure captures a different slice of activity.
As an Amazon Associate I earn from qualifying purchases.
| Company and period | Reported figures | How to read the scope |
|---|---|---|
| NVIDIA, fiscal Q2 2027; quarter ended July 26, 2026 | $96.2 billion total revenue; $89.0 billion Data Center revenue; 75.0% GAAP gross margin | Data Center is the largest reported business in the release, but it is not a pure AI-chip measure; it includes data-center products and systems, including networking. NVIDIA Q2 FY2027 results and NVIDIA 10-Q. |
| AMD, quarter ended June 27, 2026 | 54% GAAP gross margin | AMD’s filing attributes the increase to the absence of prior-year inventory and related MI308 export-control charges, as well as favorable product mix that included higher Data Center revenue. AMD Q2 2026 10-Q. |
| Broadcom, fiscal Q3 2026; quarter ended August 2, 2026 | $16.7 billion AI semiconductor revenue, up 221% year over year; $29.6 billion consolidated revenue | Broadcom linked AI semiconductor demand to custom AI accelerators and AI networking. This category is not equivalent to NVIDIA’s Data Center segment or AMD’s Data Center segment. Broadcom Q3 FY2026 results. |
Periods matter as much as labels: these latest located quarters ended on July 26, June 27 and August 2, 2026, respectively. Avoid treating them as simultaneous snapshots or using reported segment revenue as a proxy for market share, total addressable market or future demand.
How should investors compare AI exposure and growth?
Start with each company’s definition
Read the segment description and the business lines inside it before comparing revenue. Separate accelerator compute from networking, systems and other products where the company discloses them. If one company gives a narrow AI-semiconductor figure and another gives a broad infrastructure segment, compare the direction and composition of the businesses—not the totals as if they covered identical products.
#1 Best Overall
- Professional AI & Creator Workstation: AMD Radeon AI PRO R9700 GPU with 32GB GDDR6 is engineered for AI development, professional content creation, and compute-intensive workloads.
- Massive 32GB Memory Capacity: 32GB of GDDR6 memory on a 256-bit bus provides ample bandwidth for large AI models, 8K video editing, and complex 3D rendering.
- Advanced RDNA 4 with AI Accelerators: 64 Compute Units with 3rd Gen Ray Tracing and dedicated 2nd Gen AI Accelerators for groundbreaking AI performance and visual computing.
- Professional Blower Cooling: Efficient single blower design exhausts heat directly out of the chassis, ideal for multi-GPU workstation and server configurations.
- Enterprise-Grade Thermal Solution: Vapor chamber heatsink with industrial Honeywell PTM7950 thermal interface material ensures reliable cooling under sustained professional loads.
Match periods and explain what drove the change
For each metric, record the fiscal quarter or year, the period-end date, and whether growth is year over year or sequential. Then check the company’s explanation for mix shifts, acquisitions, supply constraints, one-time charges or unusual prior-year comparisons. AMD’s 54% GAAP gross margin, for example, benefited partly from the absence of prior-year MI308-related charges; that comparison alone does not establish a lasting margin trend.
Longer-term growth can add context, but it should not be substituted for a current-quarter rate. NVIDIA’s fiscal 2026 annual report states $215.9 billion in revenue, up 65% year over year, with Data Center compute revenue up 59% and Data Center networking revenue up 142%. Those are full-year figures, not the growth rates for fiscal Q2 2027. NVIDIA fiscal 2026 10-K.
What does profitability say—and what does it leave out?
Compare GAAP with GAAP, or non-GAAP with non-GAAP, and keep the period attached to every margin. Gross margin shows how much revenue remains after cost of revenue; it does not show the full cost of operating the business. For a fuller view, examine operating margin, research and development, and other operating expenses in the filings.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsRank #2
- NVIDIA Volta GV100 Architecture — 4,608 CUDA Cores, 640 1st-Gen Tensor Cores delivering 14 TFLOPS FP32 and 112 TFLOPS deep learning performance for AI training, inference, HPC, and scientific computing workloads
- 32GB HBM2 ECC Memory — 900 GB/s Bandwidth — High-bandwidth memory on a 4096-bit bus with ECC error correction provides the memory capacity and throughput required for the largest AI models, simulations, and datasets
- PCIe 3.0 x16 Interface — 250W TDP — Standard PCIe Gen3 connectivity with passive cooling designed for enterprise rack server deployment in HPE ProLiant, Dell PowerEdge, and Supermicro platforms with adequate chassis airflow
- NVLink — Scale to 96GB Unified Memory — Connect two V100 GPUs via NVLink at 300 GB/s bi-directional bandwidth to scale GPU memory from 32GB to 96GB for larger AI training and HPC workloads
- Multi-Precision Computing — Supports FP64 (7 TFLOPS), FP32 (14 TFLOPS), FP16 (112 TFLOPS) and INT8 precision modes for flexible deployment across training, inference, and scientific simulation workloads
The current figures above do not form a complete matched peer comparison: NVIDIA’s and AMD’s cited quarter figures are GAAP gross margins, while the cited Broadcom results provide AI semiconductor revenue but no comparable gross-margin figure here. Do not infer a margin advantage from missing or differently defined data. Product mix, pricing, costs and one-time charges can all affect reported margins.
How can investors assess cash generation and concentration?
Check cash conversion and investment needs
Revenue growth does not by itself show whether sales are converting into cash or how much investment is needed to sustain them. For each company and the same reporting period, review cash from operations, capital expenditures, free cash flow, working-capital movements and material commitments. Use the company’s own definition of free cash flow and reconcile it to the underlying cash-flow statement rather than assuming every company calculates it identically.
Look for reliance on a small set of customers or programs
Review disclosed major customers, end-market exposure and dependence on a small number of accelerator programs or large buyers. Also consider whether products can serve workloads beyond a narrow set of deployments. Company filings are the place to check these details; the figures above do not provide a matched customer-concentration comparison across NVIDIA, AMD and Broadcom, so they cannot establish which business is more diversified.
Rank #3
- Built for Running LLMs Locally: RDNA 4, 128 AI Accelerators, up to 1,531 TOPS (INT4) for fast inference and fine-tuning
- 32GB GDDR6 VRAM for Large AI Models: 256-bit, up to 640GB/s bandwidth, run large language and multi-modal AI models without offloading
- Multi-GPU Scaling for Local AI Clusters: PCIe 5.0 and 2-slot design support dense multi-GPU builds for local AI training and inference clusters
- Diecast Shroud and Backplate: Wave-pattern design cuts memory temperature by up to 16%, keeping clocks steady during long AI training runs
- Phase-Change GPU Thermal Pad: Delivers superior thermal conductivity for consistent performance and longevity under heavy AI loads
Which execution and supply risks can change the outcome?
AI-chip demand becomes revenue only if products can be made, delivered and deployed. Consider fabrication capacity, advanced packaging, high-bandwidth memory, power availability and data-center construction. A shipment outlook can be affected by bottlenecks at any of these stages, while customer plans may shift if infrastructure is delayed.
Free tools Windows power users keep installed
One-click scans. No signup required.
NVIDIA’s filing specifically says availability of land, power, shell and capital can affect customers’ infrastructure buildout and NVIDIA’s financial performance. NVIDIA 10-Q. Treat customer deployment capacity as part of the investment case, not as a detail separate from chip demand.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How should investors weigh export-control risk?
Export controls can alter which products a company may sell, where it can sell them, and whether inventory or plans need to change. NVIDIA’s fiscal 2026 filing said it was effectively foreclosed from China’s data-center compute market at fiscal year-end. AMD has reported MI308-related charges associated with U.S. export controls. NVIDIA fiscal 2026 10-K and AMD Q2 2026 10-Q.
Rank #4
- 24GB GDDR7 ECC Memory: handles large AI, 3D and rendering files smoothly
- Powerful CUDA Compute - 8,960 CUDA cores for fast graphics and computing power
- AI & Ray Tracing Boost - Tensor of the 5th generation and RT cores of the 4th generation
- PCIe 5.0 x16 interface - fast data connection with modern systems
- 4 × DisplayPort 2.1 - Multi-monitor support for professional workflows
These are company- and product-specific disclosures at particular dates, not a permanent forecast of policy or sales. Reassess them as rules and company disclosures change rather than assuming that a past restriction will remain unchanged.
How do you compare stock valuations fairly?
Operating growth is not the same as investment return: a strong business can still be priced for demanding expectations. To compare valuations, use market capitalization or enterprise value from the same date and pair it with the same forward period for earnings, sales and free cash flow. Check how estimates are formed and whether the companies’ fiscal calendars align; then ask what growth and margins the price appears to require.
The reported operating results above do not establish which stock is cheaper or more attractive. Without synchronized share prices and consensus estimates, a present-day valuation ranking would not be supported. Update price and estimate inputs together before drawing that conclusion.
Quick Recap
A practical comparison checklist
- Write down the exact segment or revenue definition, and identify whether it includes networking, systems or non-AI products.
- Align fiscal periods and period-end dates; label each growth rate as year over year or sequential.
- Compare the same accounting basis for margins, and note unusual charges or mix effects.
- Review operating margin, cash from operations, capital expenditures, free cash flow and working capital for matched periods.
- Check customer and product concentration disclosures, as well as manufacturing and deployment constraints.
- Track export-control exposure by product, geography and disclosure date.
- Use same-date valuation inputs and a consistent forward estimate horizon; do not infer valuation from growth alone.
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.
