Key Takeaways
- →Three new moats replace Porter’s: data flywheels, organisational velocity, sovereign stacks
- →91% of enterprises rent their AI capability — only 9% have a sovereign stack
- →Value-on-Intelligence (VOI) captures what traditional ROI misses
- →Allocate 20% of innovation budget to AI-native business model experiments
- →Speed minus trust equals liability — the moats that matter get stronger with every customer
After the strategy offsite I described in Part 1 — the one where a CTO's quiet observation about an eleven-person startup rendered a Five Forces analysis obsolete in real time — the Chief Strategy Officer pulled me aside. 'If the old moats don't hold,' she said, 'what do we actually defend?'
Wrong question. I told her so.
In the era of infinite leverage, pure defense is a losing strategy. When the cost of offense collapses — when an eleven-person company can build what took you 500 employees — the question isn't what you defend. It's what you build that gets stronger with every interaction, faster with every deployment, and harder to replicate the longer you operate it. That requires a different analytical tool entirely.
The Infinite Leverage Strategy Canvas
The canvas I've developed positions business units across two critical dimensions: Leverage Potential (how dramatically AI can scale your operations and collapse marginal costs) and Defensibility Architecture (how sustainable your competitive advantage is against AI-native competitors). The intersection creates four strategic quadrants:
Reading the Four Quadrants
Each quadrant demands a different strategic response
This is the target. AI dramatically scales your operations while proprietary advantages — data flywheels, deep workflow integration, domain-specific models — create moats that widen with every customer interaction. Think Alibaba's New Retail or Bloomberg's financial data terminal. The strategy here is acceleration: pour resources into the flywheel and compound the advantage.
Dangerous territory. AI can scale your business dramatically, but competitors can replicate your advantage within months. Most AI wrapper startups sit here. The strategy is urgent: build proprietary data loops and deep integration before a better-funded competitor copies your approach. You have velocity, not a moat — and velocity without defensibility is a countdown.
Where most large enterprises sit today. Strong competitive positions built over decades of capital investment and regulatory capture, but limited leverage potential because operations still depend on human cognitive labor at scale. The transition from Q3 to Q1 is the central strategic challenge of this decade. Optimize current operations, but invest aggressively in the AI capabilities that unlock leverage.
Neither scalable by AI nor protected from AI-native competitors. Status quo is not viable. This is where hard decisions live — divest, transform the business model entirely, or accept structural decline. The industrial services company's field operations unit landed here on our first assessment. They chose to transform, embedding AI into the service delivery model rather than defending the legacy approach.
Most organizations I assess cluster in Q3. They've built defensible positions through decades of capital investment. But their operations still run on human cognitive labor at scale, which means their leverage potential is capped. The transition from Q3 to Q1 — unlocking leverage while maintaining defensibility — is the strategic challenge that defines this decade.
The Three New Moats
In an economy where intelligence is abundant, three sources of defensibility remain scarce — and therefore valuable.
The Data Moat: Proprietary Flywheels
Foundation models are trained on public data. That makes them powerful but undifferentiated — every competitor has access to the same training corpus. The true advantage lies in closed-loop data systems that generate unique, high-value insights no public model can replicate.
Alibaba's 'New Retail' strategy in China is the clearest example I've seen of a data flywheel operating at scale. By integrating AI, logistics (Cainiao), and digital payments (Alipay) into physical retail (Hema supermarkets), they created a loop where better data produces better experiences, better experiences attract more customers, and more customers generate more data. OC&C Strategy Consultants found that top-performing Hema stores achieve twice the productivity of traditional competitors. That gap widens every quarter — not because Alibaba is spending more, but because the flywheel is compounding.
Anatomy of a Proprietary Data Flywheel
Each rotation widens the competitive gap — competitors cannot shortcut the loop
Capture Unique Data
Collect data that competitors cannot access — proprietary customer interactions, domain-specific workflows, integrated ecosystem touchpoints. Public data trains public models. Your data trains your moat.
Generate Proprietary Insight
Process through domain-specific models that improve with volume and variety. These insights are unreachable by general-purpose models because they depend on data that only exists inside your system.
Deliver Superior Experience
Deploy insights as personalized, context-aware experiences that outperform generic alternatives. The experience gap is what drives adoption — and adoption is what feeds the flywheel.
Compound Through Usage
Every interaction generates more proprietary data. The model improves. The experience gap widens. Competitors fall further behind with every rotation. Time becomes your ally.
The critical question for any AI strategy: are you building a data flywheel, or consuming someone else's? If your competitive advantage depends entirely on a third-party model API, you're renting your strategic position from a landlord who can change the terms at any time.
Speed as Strategy: Agentic Organizational Velocity
The second moat isn't a thing you own — it's a capability you build. The shift from AI-assisted work to AI-autonomous work, from tools that help humans decide faster to agentic systems that execute complex workflows on their own, is the next frontier of competitive advantage.
I call this Organizational Velocity: the speed at which a company can sense market signals, make decisions, and deploy responses. In the pre-AI era, velocity was measured in quarterly planning cycles and monthly reports. In the era of infinite leverage, velocity runs at machine speed.
The Harvard Business School and Boston Consulting Group study from Part 1 surfaced a distinction that matters here: the highest-performing organizations don't just add AI to existing workflows. They cultivate what the researchers called 'Cyborg' capabilities — human-AI collaboration where the combined system outperforms either alone. The humans set strategy and exercise judgment. The AI executes at machine speed. Neither works as well without the other.
An AI-native organization can complete four OODA loops in the time a traditional competitor completes one. Over twelve months, that's not a gap. It's a chasm that no amount of capital can bridge — because the advantage compounds with every cycle.
Why Sovereignty Matters: Verticalized AI Stacks
The third moat is about control. Relying solely on centralized foundation model providers is a strategic vulnerability — and it's getting more dangerous, not less, as geopolitical tensions rise and regulatory frameworks fracture across jurisdictions.
The United Arab Emirates provides the most compelling national example. The UAE's Technology Innovation Institute developed Falcon — a leading open-source large language model that reduces dependence on foreign technology while creating capabilities tailored to regional needs, including advanced Arabic language processing. France followed with Mistral. China has DeepSeek. Each represents a sovereign bet that intelligence infrastructure is too strategic to import entirely.
The principle scales down to enterprises. For mission-critical applications, large organizations need to verticalize their AI stacks — proprietary models built on open-source foundations, domain-specific fine-tuning they control, and the ability to switch providers without rebuilding their intelligence layer from scratch.
Nine percent. That's the share of enterprises I've assessed that have anything resembling a sovereign AI capability. The other 91% are building their competitive strategies on infrastructure they don't control. Some of them are doing it deliberately, with eyes open. Most are doing it by default, because it's easier. Easier isn't a strategy.
The Strategic Playbook: From Optimization to Transformation
Moving to the Strategic Apex demands a decisive shift — from viewing AI as an optimization layer on existing processes to recognizing it as a force that changes what's possible. Three fronts, each requiring a different kind of action.
Recalibrate How You Measure AI Investment
Traditional ROI metrics systematically undervalue AI investments because they measure linear returns in a non-linear system. The compounding value of a data flywheel, the velocity advantage of agentic systems, the optionality of owning your intelligence stack — none of these show up in a standard ROI calculation.
I propose Value-on-Intelligence (VOI) as the primary measure: the economic value generated per unit of deployed intelligence. VOI captures what traditional ROI misses. Track it quarterly. Publish it alongside traditional financial metrics. Make it a board-level conversation, not a technology team internal metric.
Investment mandate: Allocate a minimum of 20% of innovation budget to AI-native business model experiments (Q1 initiatives on the canvas), distinct from optimization efforts (Q3). Set a target: 30% of total revenue from AI-enhanced or AI-native offerings within 36 months. This forces the structural shift from incremental efficiency to business model transformation.
Restructure for Machine Speed
This is the hardest part — not because the organizational changes are complex, but because they require giving up control patterns that executives have spent careers building.
Organizational Architecture: Industrial Era vs. Intelligence Era
| Dimension | Industrial Era | Intelligence Era |
|---|---|---|
Structure Functional silos with hierarchical reporting. Decisions flow up, execution flows down. Cross-functional AI Pods organized around customer outcomes — engineering, product, and domain expertise working as a single unit with shared metrics. | Decision Speed Quarterly planning cycles. Monthly reviews. Weekly status updates. Decisions take weeks to reach execution. Machine-speed OODA loops where AI recommends and humans oversee. The target: cut insight-to-action cycle time in half within 18 months. | AI Integration AI as tool: humans decide, AI assists. Helpful but still limited to human processing speed. AI in the loop — recommending and executing within guardrails while humans set strategy and handle the judgment calls that machines shouldn't make. The last company I assessed had reduced exception handling by 70% in six months. |
Authority Model Consensus-driven. Multiple approval layers. Risk mitigation through oversight — which also means risk mitigation through slowness. Pod autonomy with guardrails and direct authority to ship. Monitor outcomes automatically instead of gatekeeping inputs manually. |
Navigate the Regulatory Fracture
The divergence in AI governance across jurisdictions creates strategic complexity that most organizations are dangerously underestimating. I sat in a board meeting last month where the general counsel admitted they hadn't mapped which of their AI systems would be classified as 'high-risk' under the EU AI Act. They had eighteen months of exposure they hadn't measured.
The Regulatory Trilemma: Three Governance Models
Each model creates different constraints and opportunities
Risk-based compliance, enforced with real teeth. The EU AI Act classifies AI systems by risk category — unacceptable, high, limited, minimal — with stringent requirements escalating by tier. Fines reach up to 7% of global turnover. Transparency, explainability, and human oversight are mandated for high-risk applications. The strategic read: treat the EU AI Act as the likely global floor, not a regional requirement. The GDPR precedent is instructive — what started as EU regulation became the de facto global standard. Build for compliance now. For the operational framework, see the Governance Playbook and AI GDPR Compliance Guide.
The strategic response is dual-track: ensure compliance with the EU AI Act as the global floor while simultaneously using innovation-friendly jurisdictions for rapid experimentation. Organizations that treat governance as a constraint will be outmaneuvered by those that treat it as a capability — because compliance-ready AI systems can deploy globally. Non-compliant ones are stuck.
The Next 90 Days: Executive Action Plan
Strategy without a timeline is aspiration. Here's the plan I walk clients through:
90-Day Infinite Leverage Transition Plan
From strategic assessment to deployed AI pods
Map Your Position (Weeks 1–2)
Plot every business unit on the Infinite Leverage Strategy Canvas. This is a C-suite exercise, not a technology team exercise — bring the CFO, the heads of business units, and the general counsel into the room. Identify where you're over-investing in Q3 optimization and under-investing in Q1 transformation. Score your industry against the susceptibility diagnostic from Part 1.
Audit Data Assets (Weeks 3–4)
For each major business unit, answer three questions: What data do we generate that competitors cannot access? Is there a closed loop where product usage improves the product? What would it take to build one? This audit almost always reveals valuable data assets that nobody has connected to the AI strategy — the raw material for flywheels that don't exist yet.
Recalibrate Investment (Weeks 5–6)
Take a revised budget proposal to the board. Commit to the 20% innovation allocation for Q1 experiments. Define VOI metrics for every AI initiative. This is the hardest conversation — you're asking the board to redirect resources from profitable Q3 optimization toward uncertain Q1 transformation. Come with the data from Weeks 1–4 to make the case.
Launch AI Pods (Weeks 7–9)
Identify three high-impact use cases for agentic AI. Staff cross-functional pods — not committees, not working groups, pods with delivery authority. Each pod needs a clear customer outcome, a 90-day delivery target, and direct authority to deploy without navigating traditional approval chains. Focus on Cyborg integration: human judgment paired with machine execution.
Global Strategy Review (Weeks 10–12)
Map the regulatory trilemma to your geographic footprint. Identify which AI systems would be classified as high-risk under the EU AI Act. Assess supply chain vulnerabilities across model providers and compute infrastructure. Develop sovereign capabilities for mission-critical applications. This review should produce a 12-month regulatory readiness roadmap.
The New Strategic Imperative
The era of infinite leverage is here. The IMF estimates that 40% of jobs globally will be affected by AI — and that estimate came before the current generation of agentic systems. The disruption potential is unprecedented. So is the opportunity.
If you're still running the Porter playbook — defending market position through capital accumulation and information asymmetry — the organizations that understood this phase change two years ago are already operating in Q1. They built the flywheels. They rewired for velocity. They own their intelligence stack. The gap is widening, and it compounds with every cycle.
The strategy question has changed. 'How do we defend our position?' doesn't have a useful answer anymore — not when an eleven-person startup can materialize from a different industry and take 40% of your value proposition overnight. The useful question is 'How fast can we accelerate?' And the honest answer, for most organizations I assess, is 'not fast enough yet.' That's where the work starts.
“Speed minus trust equals liability. The moats that matter now aren't the ones you've already built — they're the ones that get stronger every time a customer uses your product.”
Start with the 90-day plan above. Map your business on the Strategy Canvas. If you need calibration from cross-industry pattern matching, book an advisory engagement. For specific AI initiative evaluation, use the AI Use Case Canvas. For governance readiness, start with the 5-Pillar AI Readiness Assessment.
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.