AI workloads are forcing a reckoning. The mega data centers going up today will reshape communities, strain power grids, and redraw the boundaries between technology companies and energy companies. Most organizations still cannot answer a basic question: what is the carbon cost of their compute? Dan Versace from IDC joined the Supply Chain Revolution podcast to explain why energy intelligence is now a strategic operating discipline—not a reporting exercise.
From the Cherokee National Forest to Enterprise Sustainability Research
Dan Versace is a Senior Research Analyst for ESG Business Services at IDC, and his path to enterprise sustainability research is unlike almost anyone in the field. He started with a four-year degree in environmental science focused on aquatic ecology and the effects of large operations on watersheds in New Hampshire. He then spent a year in the Cherokee National Forest in Tennessee conducting invasive species and climate change research, studying how giant carbon sinks across the United States are responding to a warming planet.
His team worked inside a single four-square-mile research plot. When Dan asked his colleagues when they would move to the next area, the answer was ten years. That was the moment he decided to scale his impact. He enrolled in the University of Vermont's sustainable innovation MBA program and joined IDC, where he now leads sustainability strategy and services research across both providers and users of professional services.
I met Dan through the lens of sustainability research, and his scientific foundation shows in every analysis. When someone has studied carbon sinks at the species level and then translates that rigor into enterprise sustainability strategy, the perspective is qualitatively different from professionals who learned sustainability through a business lens first. That combination of environmental science training and market strategy is rare, and it shapes everything about how Dan sees the AI energy conversation.
How Organizations Connect Sustainability to Business Value
Dan identifies two primary axes through which organizations are connecting sustainability to measurable business value. The first is the brand and market capitalization argument. Younger generations, particularly Gen Z and millennials, are voting with their wallets. Transparent, truthful sustainability initiatives attract customers, improve brand perception, and increase market capitalization. This works across sectors where buyers increasingly prefer sustainable products and services, not just consumer packaged goods.
The second axis is operational efficiency. Organizations create sustainability programs—sometimes initially to satisfy reporting requirements—but in the process they drive efficiency gains, push their supply chain partners to be more sustainable (compounding those gains), and improve talent retention and attraction.
The talent retention argument for sustainability is one of the most underappreciated drivers in enterprise strategy. The companies with the strongest sustainability commitments enjoy some of the highest employee retention rates—not because of marketing campaigns, but because they stick to their values publicly and consistently.
This dual-axis framework dissolves the false binary between doing the right thing and doing the profitable thing. Both axes point in the same direction. The organizations that understand this are building compounding advantages over those that treat sustainability as a compliance exercise. For supply chain leaders, this is the business case you bring to the CFO: sustainability is not a cost center. It is a value multiplier across brand, operations, and workforce.
The Cynical and the Hopeful Answer for Why AI Changes Everything
When I asked Dan why AI workloads are forcing enterprise leaders to rethink energy, infrastructure, and sustainability simultaneously, he asked if he could give his cynical answer first.
The cynical answer is that it is all about cost. Organizations do not care about the reasoning behind saving energy. They care about the number on their balance sheet. Dan argues that the push to do it for the right reasons may actually slow progress, because it adds moral friction to decisions that would otherwise move faster on economic logic alone.
The hopeful answer is that the sheer scale of physical infrastructure being built for AI workloads makes sustainability impossible to ignore. Data centers are not abstract cloud services. They are massive physical installations that consume grid capacity, affect local communities, and create measurable environmental footprints. Internal sustainability champions who previously struggled for a seat at the table now have one, because the impact is material, daily, and visible to the C-suite.
Dan's critical observation is that both answers lead to the same outcome. Whether the motivation is cost or conscience, the result is that energy intelligence—the ability to measure, manage, and optimize the energy footprint of enterprise compute—is becoming a required operating discipline for any organization running significant AI workloads. That convergence of economic and environmental logic is what makes this moment different from previous sustainability conversations.
Data Center Sustainability and Community Impact: Why Technology Companies Are Now Energy Companies
One of the most provocative claims in our conversation came when Dan assessed the community impact of mega data centers. His position is that the impact on communities will be tenfold what it was during the Industrial Revolution when industries placed coal mines in towns.
Yes, data centers bring investment and employment. But they also consume grid capacity that communities need for residential and commercial use. In some cases, there will not be enough energy in the grid to power both the data center and the existing homes and businesses in that state. The issues of grid resiliency, energy security, and energy equity are deeply interconnected, and communities that are already hit first and worst by environmental burden are disproportionately affected.
Technology companies are now energy companies. That identity carries a social license to operate that most have not yet fully reckoned with. Dan is advocating for national-level guidelines that determine not just where data centers are wanted, but where they are actually needed. He also argues that sustainability subject matter experts should be physically embedded in data centers to monitor, tune, and optimize energy performance at the most granular level—not managing sustainability from a spreadsheet in a headquarters office, but standing next to the racks.
We need sustainability SMEs inside the data center. Not managing energy from a dashboard in a headquarters office. Physically present, tuning performance at the rack level. That is where energy intelligence becomes operational, not theoretical.
Small Local Models vs. Cloud AI: The Distributed Compute Pivot
Dan made a compelling case for small local models (SLMs) as a counterweight to cloud-dependent AI. The technology exists today for AI-enabled PCs to run local models that handle a significant portion of the analytical work that organizations currently route through cloud data centers. The energy footprint is a fraction of what a data center query requires.
His framing was direct: organizations do not need maximum compute for every task. A significant share of enterprise AI use cases—analyzing a graph, summarizing a document, running a classification model—can be handled locally without sending a query to a data center on the other side of the country.
The challenge is market dynamics. The largest AI providers have already cornered enterprise procurement. Organizations have invested in cloud AI licenses and infrastructure. Pivoting to local models requires both technical willingness and a shift in organizational thinking. Unless the major AI providers themselves adopt the local model narrative, the transition will be slow.
For supply chain and operations leaders, this mirrors the distributed energy generation conversation. Just as rooftop solar provides a lower-footprint alternative to centralized power generation, local AI models provide a lower-footprint alternative to centralized compute. The strategic answer is not one or the other. It is a diversified AI architecture that matches compute requirements to the appropriate scale of infrastructure—cloud for training and heavy inference, local for the everyday analytical workload.
The AI Energy Measurement Gap: Why Granularity Matters for Credible Sustainability Commitments
Dan pushed back on the assumption that energy data is invisible. His position is that it is visible at the 50,000-foot level, but the granularity is insufficient for credible sustainability commitments. What is one data center pulling from the grid? What is the carbon intensity of that specific grid at that specific time? What is the energy consumption per rack?
If a data center could run entirely on renewable energy, it would offset approximately 70% of its total societal impact. But making that claim with confidence—in an auditable and assured way—requires measurement protocols that do not yet exist in standardized form. The regulatory bodies pushing progress, including ISSB and EU supply chain regulations, are driving the right direction, but progress is slower than needed, partly due to bureaucratic friction and partly due to the lobbying power of the largest AI companies.
The cross-functional accountability question is equally important. Energy intelligence cannot live in one department. IT owns the infrastructure. Finance owns the cost. Procurement owns the vendor relationships. Sustainability owns the commitments. Operations owns the efficiency. Without a cross-functional governance model, each function optimizes for its own KPIs and the measurement gap persists. This is the same silo problem that supply chain leaders have been solving for decades, applied to a new domain: enterprise energy architecture.
Lightning Round: Six Answers Every Enterprise Leader Needs
The biggest misconception about AI and sustainability
That AI has few practical uses for sustainability and is a net negative for environmental progress. In reality, AI has massive potential to accelerate sustainability outcomes when deployed with the right governance, the right data, and the right human oversight. The technology is not the problem. The deployment model is the variable.
The one energy metric more leaders should understand
Energy consumption per rack. This is the most granular, actionable metric for understanding data center energy performance, and most enterprise leaders have never heard of it. If you are making decisions about cloud providers, AI infrastructure, or data center partnerships, this is the number that tells you what your compute actually costs the grid.
The one question every company should ask its cloud provider
How can you help me be more sustainable? Not: what is your sustainability report? Not: what are your carbon offsets? Rather: what can you actively do to reduce the energy and carbon footprint of the services I am buying from you? The distinction is between a reporting relationship and an operational partnership.
The hidden carbon impact most companies underestimate
Loss in the grid. Energy transmitted across the grid loses a measurable percentage before it reaches the point of use. Most organizations do not account for transmission loss in their carbon calculations, which means their reported footprint is systematically lower than their actual footprint.
The skill young professionals should build
The ability to diagnose sustainability issues at the root rather than from a million-foot view. The unicorn is not the AI. The unicorn is the person who understands both the science and the systems—who can move between environmental data, enterprise technology, supply chain operations, and business strategy without losing precision.
What gives the most optimism
Sustainability is currently a CFO and CEO level issue. When sustainability moves from the CSO's office to the CFO and CEO agenda, it stops being a reporting exercise and becomes a strategic operating discipline. That shift is happening now, and it is structural, not cyclical.
Listen to the full conversation with Dan Versace on the Supply Chain Revolution podcast.
Energy is not a line item. It is the operating architecture beneath every AI decision, every supply chain optimization, and every sustainability commitment an enterprise makes. The organizations that treat energy intelligence as a strategic discipline—not a reporting obligation—will be the ones that scale AI responsibly, retain talent, protect communities, and build durable competitive advantage. The reckoning is here. The question is whether your organization is measuring what matters.
Sheri Hinish is the Founder and CEO of Supply Chain Revolution Global LLC (d/b/a Supply Chain Queen®), a supply chain thought leadership, advisory, media, and community platform. A former Senior Partner at both EY and IBM Consulting, where she led global practices in sustainable supply chain and sustainability services respectively, she advises Fortune 500 companies and governments on supply chain transformation, AI-enabled operations, sustainability, circular economy, and just-transition strategy. She is recognized as a Top 250 Global Leader in Sustainability, Top 100 Women in Supply Chain and Technology, and Top 100 B2B Influencer in North America. She hosts the Supply Chain Revolution podcast, now in its third season, and speaks globally at COP, NY Climate Week, CES, and major industry events.