Essay · Capital & Regulation
Introduction: The Expensive Lesson I Had to Learn
When I started Vena Resources in 2003, I thought I understood how to operate in remote Peru. Get the permits from the government. Hire local workers. Pay taxes. Create jobs. The communities would follow the money.
I was wrong. And it cost me millions.
I walked away from high-potential geological sites because the communities that owned the land said no. Not "maybe later" or "negotiate harder." No. Our geologists had identified what they believed was significant mineral potential—millions in market valuation. We had government exploration rights. And none of it mattered because the real estate belonged to the communities, and they were not interested in our pitch.
I learned a hard truth in those early years: government permits are theater. Community control is real. And once you lose the trust of a community, you lose access forever. They talk. The next company hears the same "no."
I did not want to move on to the next opportunity. But I had to. And that failure reshaped everything about how Vena operated.
Now, AI data centers are arriving in remote towns making the exact same assumptions I made in 2003. They will learn the same lesson. The question is whether they will learn it proactively or after destroying millions in projected returns.
Part One: Who Actually Owns the Land
In Peru, the government grants exploration rights and mining concessions. On paper, this looks like ownership. You can develop the site, extract resources, build infrastructure.
On the ground, it means something else entirely. The real estate belongs to the communities that live there. You are essentially negotiating to lease their land. If they say no, the government's permission means nothing.
This is not unique to Peru. It is how remote communities everywhere approach extractive operations. The land has been theirs for generations. They understand the costs of sharing it.
The mistake I made, and the one data center operators are making now, is assuming that regulatory approval is the gate. It is not. The community is the gate.
When we first approached promising exploration sites in the Andes, we went through the government process. Permits came through. Our geologists were excited. We had the legal right to operate.
Then the communities said no.
Their reasoning was not emotional or irrational. It was economic. "If you extract resources from here, what do we lose? What do we gain? Are you leaving or staying? What happens to our water, our agricultural land, our future?"
We had answers, but not good ones. We had not asked. We had assumed money and jobs would be enough.
Part Two: The Cost of Getting It Wrong
Losing access to a high-potential exploration site is not just a delay or a setback. It is the loss of millions in potential market valuation. Our geological team had identified deposits that could have defined the company's trajectory. We had the skills, capital, and timeline to develop them.
We had one thing we could not get: community permission.
You cannot renegotiate after you have failed. Local leaders remember. They talk to other communities. The next company that shows up in that region starts from zero trust, not a fresh slate.
For an exploration company, this is catastrophic. For a data center operator, the dynamics are different but the principle is identical: lose community trust and you lose the site. You cannot build a billion-dollar facility if the community decides not to allow it.
The financial impact is not theoretical. Consider a mid-size data center: construction timeline, 3-4 years. Capital requirements, $500 million to $1 billion. Delayed permitting adds years and millions in carrying costs. Blocked access makes the entire investment worthless.
Mining and extraction industries have learned this lesson at enormous cost. Data centers have not.
Part Three: Learning to Listen
After losing those early sites, we changed our entire approach. Instead of assuming, we asked.
We sat down with communities and asked: "What do you need? What are you losing if we operate here? What would make you comfortable supporting this project?"
The answers were not complicated. They needed specific things:
Education. Their children had limited access to skilled training. If we were going to operate a mine, we needed technicians, equipment operators, geological support staff. Why not train local workers? We put our engineers into the schools. They taught math and science tied to real operations. Dozens of students came through that program over the years. Everyone who passed the tests got hired. By the time Azulcocha was running full operations with 400+ workers, a significant portion came from that school.
Water. The Azulcocha region drew water from rivers used for agriculture. We could not simply extract water and pretend there were no consequences. So we funded wells. But this was not top-down charity. The communities managed the wells. We paid for them. Joint venture. They had skin in the game.
Healthcare. There were no doctors or dentists for hundreds of kilometers. We arranged for physicians and dentists from Lima to visit monthly. Not permanent clinics. Just reliable, scheduled access to medical care they would never otherwise have.
The negotiation was clear: "Support our operation, and we will commit to these specific things." No ambiguity. No vague promises about "community investment." This is what you get. This is how we deliver it. This is how you hold us accountable.
It worked. Not perfectly. But it worked.
Part Four: The Zero-Sum Calculation on Water
Water is where the math becomes unmistakable. At Azulcocha, we were diverting water from rivers that fed agricultural areas. This was not an environmental abstraction. This was farmers' livelihoods.
A farmer relies on river water for irrigation. If a mining operation draws down that water, the farmer loses yields, loses income, loses stability. The community sees this as a trade: we lose agricultural productivity so you can mine. What do we get in return?
The answer has to be: jobs and economic opportunity that exceed the losses. Not match them. Exceed them.
If mining creates 400 jobs, those jobs need to generate enough income and economic activity to compensate the entire community for the lost agricultural productivity. This is not charity math. This is zero-sum economics.
Data centers face the same calculation. A large facility consumes 1-2 billion gallons of water annually. In water-stressed regions, this is not abstract—it means less water for agriculture, municipal supply, and future growth. The community's question is identical: "What do we gain that exceeds what we lose?"
Most data center operators have not even calculated the answer. They assume water is a utility they can buy. It is not. It is a finite shared resource that communities depend on.
At Azulcocha, we did not solve this perfectly. There were problems. There were always problems. But we acknowledged it directly: your water matters to us, we are affecting it, here is what we are doing about it. Transparency and commitment built trust. Pretending the problem does not exist destroys it.
Part Five: Why Data Centers Are Repeating This Pattern
Data center operators are not miners. They are not extracting mineral resources. But they are extracting something equally critical: water, electricity, land, and fiber capacity.
A modern data center facility is essentially infrastructure—it sits on a site, consumes massive amounts of power and water, and creates a limited number of permanent jobs. The pitch to communities is predictable: "We will bring jobs, tax revenue, and economic growth."
All true. All incomplete.
The jobs are often 50-200 permanent positions requiring specialized skills. Most locals do not have these skills. The operator says "we will train you" but often the training is outsourced, and the jobs go to people from other regions anyway.
The tax revenue flows to government, not directly to the community. When the operator relocates (and they will), the jobs disappear. But the community keeps the costs: depleted water tables, stressed electrical grid, land occupied by industrial infrastructure.
This is the mining pattern from twenty years ago. Communities recognize it. They oppose it.
The difference between a blocked project and an approved one is not the permit. It is whether the community sees the operator as a partner or an extractive force. Mining learned this painfully. Data centers are about to learn it the same way.
Part Six: What Responsible Operations Actually Look Like
There is a pattern to getting this right. It is not perfect. But it works.
Listen First, Pitch Later
Sit down with community leaders and ask: What do you need? What are you losing? What would make you confident we are a good partner? Do not lead with your business plan. Lead with their concerns.
Workforce Development Tied to Real Opportunity
If your facility needs 100 permanent employees, commit to training and hiring 60-70 locally. Embed your engineers into local schools to teach skills tied to actual jobs. This is not charity education. It is the cost of operating. When your employees are from the community, they have roots. They stay. Retention improves. Culture improves.
Water Management as a Negotiated Commitment
If your operation affects water availability, acknowledge it directly. Fund wells if they are needed. Set up monitoring programs the community runs. Commit to specific consumption targets and publish results quarterly. Do not pretend the problem does not exist. Face it. Solve it together.
Healthcare and Basic Services
In remote areas, basic healthcare is often unavailable. A data center can fund monthly clinics from regional specialists. This is not charity. It is removing friction from community life. When families have access to medical care, they are more stable, more invested, more likely to support the operation.
Local Sourcing and Economic Multiplier
Commit to sourcing 40-50% of non-core supplies locally. Maintenance, catering, landscaping, security, logistics. One direct facility job becomes 3-5 indirect jobs in the community. Local businesses grow. The community becomes invested in your success.
Written Commitments and Dispute Resolution
Put it in writing. What the operator commits to, how disputes are resolved, what happens if commitments are not met. Communities know when operators hide agreements. Make them public. Make them specific. Make them enforceable.
Part Seven: This Is Risk Management, Not Altruism
Some will say this approach is expensive. It is not. It is economically rational.
The cost of community opposition is orders of magnitude higher than the cost of proactive partnership. Legal delays. Construction stoppages. Reputational damage. An expansion that never happens. For mining companies in the 1990s and 2000s, the cost of community opposition ranged from months of delay to complete project cancellation.
Companies investing 2-3% of revenue in community relations and genuine partnership typically avoided 20-30% project delays. The math works.
Second, a community that benefits directly from an operation and has a voice in it is invested in its success. A community that sees itself as exploited will find ways to oppose it. Stability has measurable value.
Third, this approach scales. A data center operator that becomes known for honest partnership can build multiple facilities in a region because communities trust the operator. First mover advantage compounds if the first mover behaves well.
Finally, long-term workforce and supply-chain resilience matter. Employees with roots in a community have lower turnover. Local suppliers become reliable partners. Infrastructure improves. These advantages compound over 20-30 years of operation.
This is not theoretical. It is how the mining industry eventually learned to operate. It is the only way that works.
Conclusion: The Choice Is Not Whether to Engage, But When
Data center operators are about to face a choice mining companies faced decades ago: engage communities as partners, or treat them as obstacles to be managed.
Mining chose poorly at first. It cost the industry years of opposition and millions in write-downs. Companies that eventually learned to partner became the ones that expanded, thrived, and built lasting operations.
Data centers do not have to follow that path. The playbook is documented. The economics are proven. Communities are ready to negotiate if someone shows up prepared to listen.
The question is not "Should we engage with communities?" It is "Will we engage proactively, or will we engage after we have lost the site?"
I learned the answer the hard way. It cost me millions and some high-potential geological sites. The next operator does not have to make the same mistake.
But most will. Because they have not yet lost what losing feels like.
Juan Vegarra is the author of An Outsider’s Playbook (2026). He founded Vena Resources and ran it from 2003 to 2016, including the Azulcocha operation in the Peruvian Andes. More essays · Free toolkits · Advisory · Write me