Average Selling Price is the mean revenue per unit across all chips sold in a product line or category. ASP is a critical business metric that determines revenue and profitability for semiconductor companies.

Why ASP Matters

Revenue = ASP × Volume. A company can grow revenue by increasing ASP (selling more valuable chips), increasing volume (selling more units), or both. The semiconductor industry constantly balances these two levers.

ASP by Product Category

• Microprocessors (CPUs): $50-500 (consumer), $2,000-15,000 (server/data center) • GPUs: $200-1,500 (consumer), $10,000-40,000 (data center AI) • Memory (DRAM): $2-10 per chip, but sold in modules at $20-200 • Analog/Mixed-Signal: $0.10-5.00 (high volume, low ASP) • Automotive chips: $1-50 (MCUs, sensors, power) • AI Accelerators: $10,000-40,000 (NVIDIA H100/H200 class)

ASP Trends

AI is driving ASP up: Data center GPUs and AI accelerators have dramatically increased the average ASP of the semiconductor industry. NVIDIA's data center ASP exceeds $10,000 per chip. Commoditization drives ASP down: Mature products face price erosion as competition increases and manufacturing costs decline. Product mix: Companies manage ASP by shifting product mix toward higher-value, higher-margin products.

ASP vs. Margin

High ASP doesn't always mean high profit. What matters is ASP minus cost per chip. A $30,000 GPU with $10,000 in manufacturing cost has better margin than a $1 chip with $0.90 in cost, even though the percentage margins are similar.