Nvidia trades at 28x forward earnings with 75% data-centre gross margin. AMD trades at 32x with 55%. The multiple gap does not reflect the margin gap.
At current prices, Nvidia trades at roughly 28x forward earnings and AMD at roughly 32x. The relative multiple implies AMD should grow faster than Nvidia, which is mathematically true from AMD's smaller base, and should command superior margins on the incremental revenue, which it does not.
Our view is that AMD's multiple is too high relative to Nvidia's given the 2000 basis point gross-margin gap and the customer-concentration differential. Nvidia is the better risk-reward at the current spread.
Nvidia generated approximately $130 billion of revenue in fiscal 2025 at a 75% gross margin. Data centre represents 88% of the mix. Operating margin sits at 62%. Free cash flow margin is 48%.
The company holds roughly 80% share in merchant AI accelerators, essentially 100% share in training workloads above a certain complexity, and a growing share of inference. The CUDA software moat is the structural advantage that has not weakened in six years despite multiple challenger frameworks.
TickerXray Report
Get the complete AMD report with all 12 quantitative models, AI-generated investment thesis, and real-time data.
AMD's 2024 revenue was approximately $27 billion, with data centre contributing $12.6 billion and growing rapidly. Gross margin for the group is 53%, with data centre running at approximately 55 to 58%. Operating margin is 24%.
MI300 has been a credible Hopper competitor in specific inference workloads. The MI325X and MI350 roadmap narrows the gap on training. But the software ecosystem (ROCm) remains meaningfully behind CUDA, and the enterprise sales motion lags Nvidia's by roughly two years.
Silicon performance. The MI350 closes most of the raw performance gap to Blackwell on single-chip benchmarks. Nvidia retains a decisive lead on multi-chip scale-out, where the NVLink fabric and InfiniBand integration create system-level advantages that AMD's Infinity Fabric does not yet match. Edge to Nvidia.
Software ecosystem. CUDA has 15 years of ecosystem depth. ROCm has improved meaningfully in the last two years but remains the number two choice for most ML teams. The gap is roughly 18 to 24 months in developer tools and framework support. Edge to Nvidia.
Customer concentration. Nvidia's top four customers represent roughly 46% of revenue. AMD's top four customers represent roughly 32%. This is the one dimension where AMD has an advantage. Both companies face customer-concentration risk, but AMD's is slightly less acute. Edge to AMD.
Margin trajectory. Nvidia operates at 75% gross margin, AMD at 53%. The gap is roughly 2200 basis points and has been stable for two years. AMD's gross margin expansion at steady-state mix is capped roughly 500 basis points above current levels. It does not close the gap. Edge to Nvidia, decisively.
Nvidia at 28x forward earnings on $7.25 EPS implies $203 per share. AMD at 32x forward earnings on $5.50 EPS implies $176 per share. The question is whether AMD's higher multiple is justified.
The pro-AMD case rests on growth: AMD's data centre revenue is growing at roughly 40 to 60% year on year, while Nvidia's is growing at roughly 35 to 45%. On incremental revenue alone, AMD has the more impressive growth print.
The problem is that AMD's incremental gross profit dollars are 55% of incremental revenue, while Nvidia's are 75%. Pound for pound of incremental revenue, Nvidia creates 36% more gross profit. On a multiple basis, that mattered. AMD's premium to Nvidia does not reflect that difference.
Custom silicon adoption. If Amazon's Trainium, Google's TPU, or Meta's MTIA capture 15% or more of hyperscale compute spend by 2027, Nvidia's customer concentration becomes a liability. Same dynamic penalises AMD to a lesser degree given its lower starting share.
ROCm inflection. A genuine software-ecosystem breakthrough at AMD (for example, credible Pytorch-first parity with CUDA) could narrow the gap faster than we expect.
Macro capex cuts. Hyperscaler capex growth below 10% year on year in 2026 would compress both names. The data-centre revenue impact is roughly proportional.
On current multiples and growth trajectories, Nvidia is the better risk-reward. The 28x multiple on a 75% gross-margin franchise with 80% share is cheaper per dollar of durable gross profit than AMD's 32x on a 53% gross-margin franchise.
Our preference order: Nvidia first, AMD second. We would buy AMD meaningfully below $150, we are buyers of Nvidia below $180. Holding both makes sense for diversified portfolios that want AI silicon exposure; choosing one means choosing Nvidia.
Full forensic analysis of AMD
+ 6 more models included
150,000+ stocks covered
Global coverage across 60+ exchanges. Every report includes all 12 quantitative models and AI analysis.
View plansEvery report runs 12 quantitative models and generates an AI investment thesis. From Piotroski scores to manipulation detection -- get the full picture in seconds.
12 forensic models
Piotroski, Altman, Beneish, DuPont & more
AI investment thesis
Synthesized outlook on every stock
Manipulation detection
Spot red flags before they hit the news
150,000+ tickers
Global coverage across 60+ exchanges
Expected return
Forward return projections for every stock
Real-time data
Live prices, insider trades, news sentiment
Free accounts get 1 report per month. Pro gets unlimited.
Our earlier 'Five Trillion' narrative argued Nvidia was being underpriced for the AI capex cycle. Since publication, free cash flow has reached $96.7 billion, the multiple has held, and Wall Street's biggest AI bull just got louder. We update the thesis.
A $5.2 trillion market cap with revenue still compounding above 70% year over year is a profile the public market has never seen at this scale. The signals are telling a specific story.
NVDA's market cap has tripled in two years on FCF that grew from $3.8B to $96.7B. Most of the bear arguments are about the multiple, not the cash. Here's what the consensus is pricing wrong.