Technology · Semiconductors
Nvidia (NVDA)
Nvidia is the fabless designer: high margins, low capex, concentrated customers and large manufacturing commitments, read from its FY2026 10-K.
Selborne Research · · Company Profile
Educational analysis. This profile is not investment advice or a recommendation to buy or sell any security, and it is not personalised.
Snapshot
- FY2026 Revenue
- $215.9B (+65%)
- Data Centre Revenue
- $193.7B (FY2026)
- Gross Margin
- 71.1% (FY2026)
- Operating Margin
- 60.4% (FY2026)
- Capex / Revenue
- 2.8% (FY2026)
- Inventory Days
- 125 (FY2026)
- Supply Commitments
- $95.2B (FY2026)
- Two Direct Customers
- 22% and 14% of revenue (FY2026)
Business Overview
Nvidia is a fabless designer that owns no wafer fabs. FY2026 (year to 25 January 2026) revenue was $215.9B, up 65% from $130.5B. Gross margin was 71.1% and operating margin 60.4%. R&D was 8.6% of revenue, capex 2.8% and stock-based compensation 3.0%. Free cash flow was $96.7B.
Data centre was $193.7B of revenue. Two direct customers were 22% and 14% of revenue. Manufacturing and supply commitments were $95.2B against inventory of $21.4B. Inventory days were 125 (cash conversion cycle 133 days). In FY2023 gross margin fell to 56.9% after about $2.2B of inventory and purchase-commitment charges; FY2025 gross margin was 75.0%.
How the Numbers Read
Capex at 2.8% of revenue is the fabless signature: wafers are bought from foundries, so the capital intensity sits with TSMC and peers. Gross margin in the seventies reflects that structure in a strong demand year. The same structure leaves inventory and purchase commitments as the cycle risk: $95.2B of commitments against $21.4B of inventory.
FY2023's drop to 56.9% gross margin after roughly $2.2B of charges is the inventory-cycle case study inside one company. Customer concentration at 22% and 14% means order timing at a few accounts moves consolidated results. The business-model guide places fabless margins and capex beside foundry and IDM profiles.
How the Model Values This Kind of Company
A fabless designer is valued on units times average selling price. Channel inventory can amplify that: revenue swings harder than end demand when customers fill or drain stock. Cost of goods is largely variable; a down year can carry an inventory write-down.
Capex stays low as a share of revenue and stock pay is charged as a cost. Inventory days rise when the cycle turns down. Terminal value uses mid-cycle free cash flow. A reverse DCF on a fictional company shows what mid-cycle operating margin a typed price requires.
What to Watch in the Financials
Manufacturing and supply commitments versus inventory. $95.2B against $21.4B, 4.4 times as much. The commitments sit off the balance sheet; if demand slows, they become stock the company must take or pay to cancel, as in FY2023.
Inventory days. 125 days. Rising days ahead of slowing revenue growth is the early warning.
Customer concentration. 22% and 14% for two direct customers. Order pauses at either account are material.
Gross margin versus FY2023 and FY2025. 71.1% in FY2026, against 56.9% and 75.0%. Cycle position shows up here before revenue fully turns.
Key Risks
Commitment overhang. $95.2B of manufacturing and supply commitments can force charges if customers destock, as the FY2023 episode showed.
Customer concentration. Two direct customers at 22% and 14% of revenue concentrate both growth and downturn risk.
Channel inventory. Fabless revenue can overshoot and undershoot end demand when customers double-order and then destock.
Foundry dependence. Asset-light economics still depend on foundry capacity and pricing; those costs sit in gross margin, not in capex.
A chip DCF carries the industry cycle through ten explicit years, then strikes the terminal value on what the business earns in a normal year, because the latest margin may sit at a peak or a trough.
- 17 sections, from the business models and the industry cycle to capital intensity, memory, inventory, equipment orders, stock pay and a reverse DCF
- 44 pages
- a chip designer, a foundry, a memory maker and an equipment maker, all fictional
- 4 worked valuations
- listed chip companies on filed margins, capital intensity, inventory days and stock pay, each with its fiscal year
- 6-company screen
The Excel model is the primer's through-cycle DCF live across 13 sheets: a chip designer, a foundry, a memory maker and an equipment maker, all fictional, plus a column for a company of your own in any of five business-model modes. Each mode turns the industry cycle into revenue its own way: units and selling prices, capacity times utilisation times revenue per wafer, bit demand against bit supply, orders and installed-base services, or wafer starts for a materials supplier. Depreciation sits in cost of goods for fab owners, capacity arrives two years after the capex that pays for it, and the terminal value is struck on mid-cycle cash flow. A reverse DCF reads the margin and the growth a price requires, and the sensitivity grids re-run the model in every cell.
See what's in the Semiconductors Sector Primer →£25 PDF · £59 with the Excel model · £159 for all three Technology industries
Learn the Concepts
Understand the valuation frameworks and metrics used in this analysis.
The Semiconductor Inventory Cycle
Chip revenue swings harder than end demand because customers build and drain inventory. WSTS history, Nvidia charges and Texas Instruments capacity.
Capex to Revenue by Chip Company
Ranked capex as a share of revenue for six chip filers. Business model sets the order from foundry and IDM down to fabless and equipment.
Fabless vs Foundry vs IDM vs OSAT
Four chip business models defined by who owns which assets, with margin and capex profiles from Nvidia, TSMC and Intel filings.