Driving Intelligence to Zero: Reshaping Enterprise Strategy
Over the last two years, I’ve witnessed a dramatic shift in the AI landscape – one that I believe is going to force organizations across every industry to rethink their strategic roadmap. The cost of “intelligence” itself – the raw computational power and increasingly sophisticated models underpinning AI – has been on a precipitous decline, even as performance continues to climb. In many cases, if you don’t need the absolute bleeding edge, you can opt for a previous generation model at a fraction of the cost, sometimes close to zero. This reality, vividly illustrated in the accompanying graph, is fundamentally changing the way enterprises deploy and monetize AI.
The Graduate-Level Google-Proof Q&A test (GPQA) is a series of very hard multiple-choice problems designed to test advanced knowledge. PhDs with access to the internet get 34% right on this test outside their specialty, 81% inside their specialty. The cost per million tokens is the cost of using the model (Gemini Flash Thinking Costs are estimated). Professor Ethan Mollick – please see his original post here:
What does this cost curve mean for an enterprise in my opinion? Quite simply, companies need to stop thinking about this as paying for raw tokens or CPU cycles. They need to think about this as paying for outcomes – whether it’s automating customer support, streamlining contract workflows, or, in the case of healthcare and life sciences, accelerating clinical trial document analysis and expediting drug discovery. The value isn’t in the isolated model; it’s in the integration of that model into a broader, outcome-driven system. As AI performance converges and the cost of even older models becomes negligible, the pressing question for decision-makers is no longer “Which model should we adopt?” but rather “How can we embed AI into our operations to drive meaningful, measurable impact?”
The way I see it is that this shift transforms AI from a futuristic add-on into a strategic imperative. When the cost of intelligence plummets, every function – from finance and operations to R&D and HR – has the potential to harness these capabilities. The true competitive edge now lies in how deeply and creatively a company integrates AI into its value chain. It’s no longer about buying a cutting-edge model; it’s about building an ecosystem that turns that intelligence into a tangible outcome. For instance, in healthcare, the ability to process vast datasets for faster, more accurate patient insights can mean the difference between merely surviving regulatory pressures and truly innovating in patient care.
Moreover, this cost curve offers significant implications for margins and reinvestment. As the price of raw AI continues to drop dramatically – plummeting along a log-scale axis – vendors and enterprises should be able to realize substantial margin expansion. However, in today’s hyper-competitive landscape, these efficiency gains should not just bolster the bottom line. Instead, they should be reinvested into growth initiatives, enhancing customer experiences, or developing next-generation products. This is not about saving – it is about growing. In sectors like life sciences, where rapid iteration and stringent compliance are the norm, these reinvestments can translate into accelerated innovation cycles and a much higher return on R&D spending.
Enterprises have always grappled with the core versus context dilemma – this was a constant theme of clients when I ran my strategy consulting firm – determining what to build in-house versus what to buy off-the-shelf. With AI’s rapid commoditization, this question takes on renewed urgency. Organizations must discern which processes truly differentiate them, where AI can amplify that competitive advantage, and where standard, commoditized solutions will suffice. In a world where access to intelligence is nearly free and that intelligence continues to improve, the differentiator becomes the proprietary layers – data integration, business logic, and user experience – that transform mere automation into a strategic moat.
In practical terms, I am suggesting that enterprises should double down on integration, adopting outcome-based metrics that go beyond traditional ROI calculations. As AI’s cost per token becomes negligible, the focus shifts from the technical minutiae to the real-world impact: increased conversion rates, reduced time-to-insight, and accelerated product development cycles. Companies need to develop a clear strategy for when to upgrade models – balancing the marginal gains of bleeding-edge technology against the near-zero cost of previous-generation models. Ultimately, it’s about matching model capabilities to specific use cases, ensuring that every dollar saved on intelligence is reinvested to further embed AI into the core fabric of the business.
From what I am seeing we are at a moment where the abundance of low-cost intelligence is not just about automating tasks, but about fundamentally reshaping enterprise strategy. The organizations that thrive will be those that view AI not as a peripheral technology, but as a cornerstone of their competitive strategy – a tool that supercharges productivity, fuels innovation, and creates value that is, quite frankly, priceless. I wish I had this in some of my previous companies! With intelligence trending toward zero cost, you must harness it strategically to unleash its full transformative potential.
