Key Takeaways
- →AI training costs are collapsing at 70% annually — steeper than Moore’s Law
- →Lean AI companies achieve 15-25x revenue per employee vs traditional SaaS
- →Porter’s Five Forces lose predictive power when intelligence cost approaches zero
- →Your competitors are now any organisation that can deploy AI against your customers’ needs
- →The war of attrition is over — what’s scarce is organisational clarity to deploy intelligence
Last quarter, I sat in on a strategy offsite for a $2 billion industrial services company. The Chief Strategy Officer had projected Porter's Five Forces framework onto the screen — the same framework I've seen in every strategy room for twenty years. Barriers to entry: high. Supplier power: moderate. Competitive rivalry: manageable. The board nodded along.
Then the CTO raised her hand. 'Three months ago, a startup with eleven people launched an AI system that does 40% of what our field services team does. They're pricing at one-tenth our cost. They didn't exist a year ago.' Silence.
The Five Forces analysis on the screen said the company was well-positioned. The market was saying something different.
The strategic frameworks that have guided executive decisions for forty years — most notably Michael Porter's Five Forces — were forged in an era defined by scarcity. Capital was finite. Labor was a significant variable cost that scaled linearly with output. Information was constrained by geography and institutional access. We are exiting that era, and the strategy tools built for it are losing their predictive power.
Artificial intelligence is introducing a new economic reality: the era of infinite leverage. Not hyperbole — an economic definition. Infinite leverage describes the condition where the capacity to execute cognitive tasks — analysis, creation, decision-making — decouples from human labor hours, and the marginal cost of intelligence asymptotically approaches zero.
Efficiency gains from AI are table stakes — every competitor gets those. The real strategic question is whether your organization can learn faster than the market shifts beneath it. That's what demands a rewrite of the playbook.
The Economics of Infinite Leverage
The evidence for this phase change isn't subtle. It shows up in three convergent trends — and the numbers are getting harder to explain away.
The cost curve is collapsing
Research from Ark Invest, applying Wright's Law to AI training, shows that the cost to train a large language model to GPT-3-level performance decreased by approximately 70% annually between 2016 and 2022. If those trends persist — and the data so far says they will — frontier model training costs could see a 10,000× reduction by the end of the decade.
This isn't incremental. This is a phase change in the economics of intelligence — steeper than Moore's Law for semiconductors, faster than the cost decline in solar energy. The implications compound in ways that traditional strategy frameworks were never designed to capture.
The rise of the Lean AI Company
ElevenLabs hit $100M+ in annual recurring revenue with fewer than 50 employees — a voice synthesis company that would have needed 500 engineers a decade ago. Cursor is building in the same mold: an AI-native code editor reaching comparable valuation milestones with a team that fits in a single conference room. These aren't outliers. They represent a 15–25× improvement in Revenue Per Employee compared to the SaaS benchmarks of the last decade.
When 42% of your cost structure is cognitive labor, and the cost of cognitive output is collapsing, the organizational model that got you here won't get you where you need to go. The constraint on growth shifts from 'how many people can we hire' to 'how fast can we deploy intelligence.'
The productivity evidence
A landmark 2023 study by Harvard Business School and Boston Consulting Group — 'Navigating the Jagged Technological Frontier' — quantified what many of us had observed anecdotally: consultants using advanced AI completed tasks 25.1% faster and produced 40% higher quality results. Not marginally better. Fundamentally different economics of cognitive output.
Three terms matter for what follows. Infinite leverage is the simplest to explain: it's what happens when you can scale cognitive work at near-zero marginal cost, constrained by compute and data rather than headcount. Agentic AI is harder to grasp until you see it — systems that autonomously execute multi-step workflows, adjusting strategy mid-stream without waiting for human approval at each step. Then there's machine speed competition, which is what every executive should be losing sleep over: OODA loops — observe, orient, decide, act — running at algorithmic speeds while your planning cycle is still quarterly.
Why Porter's Five Forces Lose Their Predictive Power
Porter's framework endured for four decades because its logic was sound within an economy of scarcity. Each force measured a different dimension of resource-based competition. In the era of infinite leverage, every dimension breaks — but in different ways that demand different responses.
The Five Forces Under Infinite Leverage
How each force changes when the marginal cost of intelligence approaches zero
Porter assumed high capital costs protect incumbents. That assumption died quietly in 2023. A Lean AI Company can achieve global scale with minimal capital — open-source models, cloud infrastructure, and a team small enough to fit in a WeWork office. ElevenLabs and Cursor didn't need billion-dollar war chests. They needed proprietary data and learning velocity. Here's what this means for the $2 billion industrial services company from my opening: their capital moat is shrinking. Not slowly — rapidly. A startup that can't compete on headcount can build an AI system that makes headcount irrelevant, serving each customer with personalized intelligence that manual processes can't match at any scale. The new barrier to entry isn't how much you've spent. It's how fast you learn from every customer interaction.
Porter's forces still exist. But they were built to measure competition when intelligence was expensive and locked inside people's heads. That's not the world we're operating in anymore — and using the old calibration is like navigating with a map drawn before the geography changed.
“The war of attrition is over. Intelligence is cheap now — what's scarce is the organizational clarity to deploy it before your competitors do.”
Is Your Industry Susceptible?
Not every industry hits the infinite leverage wall at the same speed. The velocity depends on five characteristics — score your industry 1–5 on each:
Infinite Leverage Susceptibility Diagnostic
Score each criterion 1–5, then sum for your composite assessment
Is the core product or service primarily digital? Can it be delivered through software without physical transformation? Financial services, media, education, and professional services score 4–5. Manufacturing, construction, and agriculture score 1–2. The higher the digitization, the faster AI can substitute or enhance the core offering.
What percentage of operating costs is tied to cognitive labor — analysis, writing, decision-making, customer communication? If more than 60% of your costs are people doing thinking work, you are sitting in the blast radius of infinite leverage. The donut chart above shows a typical enterprise at 42%. Some professional services firms run above 70%.
How expensive is it to serve one additional customer with your current operating model? High marginal costs — where each new customer requires proportionally more human effort — signal maximum disruption potential. AI collapses marginal costs for cognitive work toward zero. If your business model is built on selling human hours, the economics are about to break.
Can your business create a closed-loop data system where every customer interaction improves the product? Industries with rich behavioral data (e-commerce, fintech, healthcare) have natural flywheel architectures. Industries with sparse, episodic customer interactions (construction, real estate) need to manufacture them. The flywheel is the new moat — more on this in Part 2.
How rapidly are the core tasks in your industry being automated or augmented by AI? Not adjacent tasks — the core revenue-generating activities. If your competitors are already deploying AI against your primary value proposition, you are in a sprint, not a marathon. Check what startups funded in the last 18 months are targeting in your space.
Above 20, the disruption is already underway — your industry sits in the high-leverage zone and the economics are shifting now. The uncomfortable middle, 12 to 20, is where most executives I work with find themselves: sensing the shift but uncertain of its speed. Below 12 means the disruption is real but further out. Every industry will get there eventually. The question is whether you're building the moats now — or reading about the companies that did, in a case study two years from now.
What Replaces Porter?
If the Five Forces lose their predictive power, what fills the strategic vacuum? Not another defensive framework — defense is a losing strategy when the cost of offense collapses. What executives need is an offensive framework.
In Part 2 of this series, I introduce the Infinite Leverage Strategy Canvas — a 4-quadrant positioning tool for mapping business units across two dimensions: leverage potential and defensibility architecture. I'll cover the three new moats that matter — proprietary data flywheels, agentic organizational velocity, and sovereign AI stacks — with case studies from Alibaba's New Retail strategy, the UAE's sovereign AI initiative, and a 90-day executive action plan for making the transition.
That industrial services company? After the strategy offsite, the CSO put the Five Forces slide away. We spent the next two days mapping their business units on the Infinite Leverage Strategy Canvas instead. Three units went into the 'transform or divest' quadrant. Two had flywheel potential nobody had identified. The strategic conversation changed completely — from 'how do we defend our position' to 'how fast can we accelerate.'
Read Part 2: The New Moats — Building Defensibility in the Era of Infinite Leverage for the Infinite Leverage Strategy Canvas, three new moat architectures, and a 90-day executive action plan. For evaluating specific AI initiatives within this new landscape, the AI Use Case Canvas provides the 12-block evaluation framework.
Get Weekly Thinking
Join 2,500+ AI leaders who start their week with original insights.
Senior AI strategist helping leaders make AI real across four continents. Forbes Technology Council member, IEEE Senior Member.