Gross Margin

Gross margin measures how much revenue a company keeps after paying the direct costs of production or delivery (cost of goods sold, or COGS) — hosting and infrastructure costs for software, cloud compute for AI products, materials and fulfillment for physical goods. It's foundational to unit economics, since it determines how much revenue growth actually translates into gross profit that can fund the rest of the business.

SaaS companies typically target gross margins of 70-85%, since software has relatively low marginal delivery costs; AI-native companies with heavy inference and compute costs often run meaningfully lower gross margins (sometimes 40-60%) that investors now scrutinize closely, since compute costs don't naturally decline the way traditional SaaS hosting costs have over time.

Gross margin is the first and most honest test of whether a business model works. Every dollar of revenue at a 40% margin produces less than half the gross profit of the same dollar at 85%, so a low-margin company has to grow much faster to fund the same operating budget. The trend matters as much as the level: margin holding steady while revenue triples shows operating leverage; margin falling as usage scales shows the opposite.

Classification matters. COGS for software typically includes hosting and infrastructure, customer support tied to delivery, third-party data or API fees embedded in the product, and — for AI products — inference and compute. Engineering salaries for building new features are R&D, not COGS; sales, marketing, and overhead aren't COGS either. Misclassifying costs in either direction distorts the number investors will rebuild in diligence.

Compare like with like. A marketplace, a hardware-enabled software business, and an AI API business all have different margin structures than pure software, and benchmarking them against the wrong peer set produces false conclusions.