Nathan Scharfe

SaaS businesses have one critical baked-in assumption: The marginal cost of another user was close to zero. LLMs are breaking that assumption. In a traditional SaaS model, you paid for infrastructure in chunks. As long as usage stayed within those bounds, an extra user barely moved the needle. Margins improved as you grew. When LLMs become part of the product experience, that logic changes. Costs stop being mostly fixed and start becoming interaction-based. 👉 Every prompt. 👉 Every completion. 👉 Every background agent action. 👉 Every “helpful” suggestion. Each one carries a real, incremental cost. That shifts the economics in ways many SaaS teams haven’t fully internalised yet. In an LLM-driven product: 📈 Costs scale with usage intensity, not just user count 👥 Heavy users are materially more expensive than light users 🏃‍♂️ The link between “active user” and “profitable user” gets fuzzy Two customers on the same plan can now have very different cost profiles, purely based on how they use the AI-powered parts of the product. This creates pressure in unfamiliar places: ❗Pricing models built on uniform usage assumptions ❗“Unlimited” plans that quietly become loss-making ❗Features users love but finance worries about ❗Growth that increases revenue and cost at the same time None of this is an argument against using LLMs. But it is a reason to revisit some deeply embedded assumptions. AI features aren’t “just another feature”. Usage, value, and cost no longer move together. Pricing and product design are now tightly coupled to model economics. The companies that struggle won’t be the ones adopting LLMs. They’ll be the ones still pricing as if marginal cost is zero. When it isn’t. This shift doesn’t usually hurt immediately. It shows up later: 👊 in margin surprises 👊 in awkward pricing changes 👊 in “why is this customer unprofitable?” conversations So here’s the question I keep coming back to: If usage now drives cost, how should pricing evolve to reflect that? Especially where customers want price predicability?

13

5 Comments

I had an absolute blast chatting with Tracy Lee and Danny Thompson on the Leadership Exchange Podcast about what it really takes to build intelligent systems in the AI era. We went deep on: 🤖 The shift from single-model thinking to agentic ecosystems 🧭 Why judgment and evaluation matter more than ever 💼 How AI and Product roles are evolving 🧩 Why data quality and governance quietly make or break AI success 🎯 What the rise of the Chief AI Officer tells us about the future of leadership 🔥 Hot take: AI transformation is 20% tech, 80% culture. If you’ve ever wondered how to evolve your teams, systems, or career in this new AI landscape — this one’s for you. 🎧 Watch here → https://lnkd.in/gABmqvTb What’s your biggest challenge in scaling AI beyond prototypes? #AI #Leadership #Innovation #ProductManagement #DataScience #AIAgents #Engineering

33

2 Comments