40 Top Scholars Issue Joint Declaration: AI Is Reshaping Supply Chain Management at Five Fundamental Layers
When 40 of the world’s most distinguished operations management scholars from MIT, Stanford, Wharton, INSEAD, and Kellogg issue a joint vision statement, the supply chain industry has every reason to pay close attention. Published in December 2025 on SSRN, Supply Chain Management in the AI Era—led by Maxime C. Cohen (McGill), Tinglong Dai (Johns Hopkins), and Georgia Perakis (MIT Sloan), with co-authors including David Simchi-Levi (MIT), Hau L. Lee (Stanford), Gad Allon (Wharton), and Christopher S. Tang (UCLA)—proposes a systematic five-layer framework addressing a fundamental question: How exactly is AI rewriting the rules of supply chain management?
The significance of this statement extends beyond its academic weight—it represents the collective consensus of the global operations management (OM) community. More importantly, it highlights a critical blind spot: AI is not only transforming supply chains, but AI itself is becoming an unprecedented mega supply chain. The global scramble for GPU chips, the energy consumption of data centers, and the geopolitical risks surrounding computing infrastructure are supply chain problems that have only recently entered mainstream awareness.
The Five-Layer Framework: A Comprehensive Deconstruction from Intelligence to Infrastructure
The paper’s core contribution is its Five-Layer Framework for AI-supply chain interaction, spanning from micro to macro, from technical to humanistic dimensions. These layers form an interconnected system where breakthroughs at any level send shockwaves up and down the stack.
Layer 1: Intelligence. This is the most intuitive application domain. AI dramatically improves demand forecasting accuracy through machine learning, enables real-time risk signal parsing from global news and social media via NLP, and automates quality inspection through computer vision. The authors highlight that Generative AI (GenAI) is opening entirely new possibilities—from auto-generating procurement documents to intelligent contract review, multi-scenario simulation, and real-time decision recommendations. McKinsey estimates that AI could save global supply chains $1.2–2 trillion annually in demand forecasting and inventory optimization alone. However, the authors warn that AI’s “black box” nature may introduce new systemic risks—when all companies use similar AI models, a collective forecasting bias could trigger an “AI bullwhip effect,” amplifying supply-demand oscillations rather than dampening them.
Layer 2: Execution. AI is redefining how supply chains physically operate. Amazon has deployed over 750,000 robots across its warehouses; the global autonomous trucking market is projected to reach $160 billion by 2030; drone delivery has mature applications in medical logistics (e.g., Zipline’s blood delivery network in Africa). The deeper transformation isn’t just “replacing humans with machines”—it’s real-time optimization and dynamic adaptation. Logistics routes are no longer fixed plans but continuously adjusted optimal solutions responding to real-time traffic, weather, and order changes. This demands a fundamental shift in supply chain IT architecture from batch processing to stream processing.
Layer 3: Strategy. This layer has the most profound impact on senior executives. AI is transforming strategic-level decisions: network design (where to locate factories and warehouses), supplier selection (risk assessment and multi-sourcing strategies), and make-or-buy decisions. Digital Twins have evolved from concept to practical tool—companies can now simulate thousands of supply chain configurations in virtual environments, completing in hours what previously took months. AI is also catalyzing entirely new business models: Supply Chain as a Service (SCaaS), algorithmic pricing, and predictive maintenance subscriptions are blurring the traditional boundaries of supply chains.
Layer 4: Human. The most easily overlooked yet arguably most critical layer. AI won’t eliminate supply chain jobs, but will fundamentally restructure required skill sets. The paper predicts that by 2030, over 60% of supply chain roles will require some form of AI collaboration capability. “Data translators”—professionals who bridge business needs and technical implementation—will become the scarcest talent. The authors particularly emphasize the importance of “Human-in-the-Loop” design: AI recommends, humans decide—at least in supply chain, where the cost of errors is extremely high.
Layer 5: Infrastructure. This is the paper’s most original contribution. AI itself is becoming one of the world’s most complex supply chains. Training a large language model requires thousands of GPU chips, with over 80% of high-end AI chips designed by NVIDIA, manufactured by TSMC, and shipped through a handful of ocean routes. A Taiwan Strait crisis or a TSMC capacity bottleneck could stall global AI development. Data center energy consumption is equally alarming—by 2026, global data centers are projected to consume over 1,000 TWh of electricity, equivalent to Japan’s total national consumption. The paper calls on OM scholars to integrate these issues into their research agenda.
Academic Self-Examination Meets Practitioner Critique
Despite its landmark significance, the vision statement has drawn sharp criticism. Lokad founder Joannes Vermorel, in a commentary titled “Why Practitioners Are Right to Ignore This ‘AI Era’ Vision for Supply Chain,” raised three core objections:
First, the emptiness of “supra-economic goals.” The paper repeatedly emphasizes that supply chains must achieve “sustainability, resilience, and equity” but never provides operationally quantifiable definitions. Vermorel argues: “Scarcity doesn’t disappear just because we invoke these words. Every pallet, man-hour, and coin devoted to one objective is withheld from another.” Supply chain management’s ultimate metric is risk-adjusted return on capital—not hollow adjectives.
Second, the academic-practice disconnect. While logically coherent, the five-layer framework offers virtually no actionable guidance for a supply chain manager who needs to make concrete decisions on Monday morning. It tells you “AI matters” but not “what to do tomorrow.”
Third, the absence of economic objectives. Profit, capital productivity, and risk-adjusted returns are barely discussed directly. Vermorel considers this a fatal omission: “Supply chain is not moral philosophy; it is applied economics—survival depends on the ledger.”
From Theory to Practice: What This Means for Practitioners
Setting academic debates aside, this 40-scholar joint statement sends several unmistakable signals to supply chain practitioners:
Signal 1: AI’s impact on supply chains is structural, not instrumental. This isn’t simply about “building a better forecasting model with AI.” The five-layer framework clearly shows that AI is simultaneously changing how supply chains perceive (Intelligence), operate (Execution), decide (Strategy), employ people (Human), and exist (Infrastructure). Any company optimizing at only one layer risks being swept aside by systemic transformation.
Signal 2: The AI supply chain itself will become a new focal point of global competition. The Infrastructure layer discussion reveals a profound insight: the ability to control the AI supply chain—chips, compute, energy, data—will become a core element of 21st-century competitiveness for both nations and corporations.
Signal 3: The talent crisis is more urgent than the technology crisis. The biggest bottleneck isn’t algorithms or hardware, but people who can understand AI and embed it into business decisions. Companies investing in team AI capability building now will gain significant competitive advantages in 3–5 years.
Six Actionable Recommendations for Supply Chain Professionals
1. Start with forecasting, but don’t stop there. Demand forecasting is AI’s most mature supply chain application and the highest-ROI starting point. But quickly extend to anomaly detection, dynamic pricing, and automated replenishment to build a complete intelligent decision loop.
2. Build a “data balance sheet” mindset. AI effectiveness depends on data quality. Conduct quarterly “health checks” on core supply chain data—demand, inventory, supplier performance, logistics tracking—assessing completeness, timeliness, and accuracy. Without quality data, even the most advanced AI is futile.
3. Monitor AI supply chain geopolitical risks. If your business relies on AI capabilities (e.g., cloud-based supply chain optimization), assess the concentration risk of your AI infrastructure. Where do your AI vendor’s GPUs come from? Which country hosts their data centers? These questions are increasingly critical amid intensifying US-China tech competition.
4. Invest in “data translators,” not just data scientists. Most companies don’t lack algorithm talent—they lack professionals who can translate business problems into data problems and data insights back into business actions. These hybrid talents typically require 5–8 years of development; starting now isn’t early.
5. Measure AI investments with economic metrics, not buzzwords. Every AI investment should have clear expected economic returns: How much inventory capital does it free up? How much does it improve perfect order rates? How much does it reduce transportation costs? If you can’t answer these questions, the project may not be ready.
6. Maintain “Human-in-the-Loop” decision architecture. In supply chain, a single AI algorithm error can cause millions in losses. Before fully trusting AI, ensure critical decision points have human review. Gradually expand automation scope as trust and validation systems mature.
Conclusion: The Opening Act of a New Era
Whether or not you agree with every assertion in Supply Chain Management in the AI Era, it marks a pivotal moment: the world’s top supply chain academic minds collectively acknowledge that AI’s impact on supply chain management is no longer a question of “whether” but “how fast and in what way.”
As David Simchi-Levi has said on multiple occasions: “The future of supply chain management is not AI replacing humans, but supply chain teams that master AI replacing those that don’t.” That, perhaps, is the single most important message from this 40-scholar joint declaration.
Source: Cohen, M.C., Dai, T., Perakis, G., et al. (2025). “Supply Chain Management in the AI Era: A Vision Statement from the Operations Management Community.” SSRN Working Paper | INFORMS Open Forum | Lokad Critique by Joannes Vermorel
