By Gerald DeVito July 2, 2026
Why the Grid Alone Cannot Keep Up
By Gerald DeVito July 2, 2026
Feeding People and Powering Machines: AI, Farmland, and Food Security
By Gerald DeVito June 20, 2026
What AI Can Do for Climate The potential of AI to address climate instability is real and already partially realized. Here is what the evidence shows: Energy systems optimization. AI systems trained on real-time weather, industrial load, and consumer behavior data can predict electricity demand 24 to 72 hours ahead with greater than 95% accuracy, allowing utilities to dispatch renewable generation at maximum efficiency and reduce reliance on fossil fuel peakers. Google's DeepMind achieved a 40% reduction in data center cooling energy using reinforcement learning. The same technique is now being applied to regional grid optimization. Accelerating renewable siting. AI-powered satellite analysis and geospatial modeling can evaluate millions of potential solar and wind sites in hours, identifying optimal land while avoiding sensitive habitat, flight paths, and transmission constraints. Tasks that previously required years of manual engineering analysis are being compressed into weeks. Weather and extreme event modeling. NVIDIA's FourCastNet and Google's GraphCast now generate 10-day global forecasts in under one second, compared to hours for traditional numerical weather prediction. They are beginning to model hurricane intensification, atmospheric river behavior, and wildfire spread with a precision that gives emergency managers days more warning than previously available. Materials discovery. Google DeepMind's AlphaFold transformed protein structure prediction. The same deep learning approach is now being applied to discover new materials for solar cells, batteries, and carbon capture. Microsoft's MatterGen AI model identified over 2 million new stable inorganic materials in 2025, including candidates for next-generation solid-state batteries and high-efficiency photovoltaics. Carbon monitoring. AI systems analyzing satellite imagery can now detect methane leaks from oil and gas infrastructure in near real-time, identify illegal deforestation within hours, and track industrial emissions with a granularity no human inspection regime can match. An Honest Assessment AI can optimize existing energy and transportation systems by 15 to 40 percent. This is proven, deployed, and scaling now. AI can accelerate materials discovery by 10 to 100 times, potentially unlocking next-generation clean energy technologies within a decade rather than three. AI cannot substitute for political will, infrastructure investment, or behavioral change. AI's own energy footprint must be powered cleanly for its climate contributions to be net positive. That is the central argument for the infrastructure Solar DC Power is planning to build. What Comes Next Post 4 in this series covers AI and food security: how AI is helping address the challenge of feeding 9.7 billion people by 2050, and where Solar DC Power's agrivoltaic model connects directly to that story. Publishing June 22. Solar DC Power is planning to develop agrivoltaic-powered rural data centers and community microgrids in Georgia, the Carolinas, and Costa Rica. Learn more at www.solardcpower.com .
By Gerald DeVito June 19, 2026
What AI Actually Is: And Why It Matters Now
By Gerald DeVito June 19, 2026
The Hidden Cost of Waiting: How Engineering Experience Saves Years on Data Center Permitting How 42 years of civil engineering experience translates into faster permitting and fewer costly surprises. The data center industry is in a race. Artificial intelligence, cloud computing, and the digitization of nearly every sector of the global economy have created a demand for computing power that existing infrastructure cannot meet. Developers are scrambling to site, permit, and build new facilities as fast as capital allows. But capital has a clock. Every month a project sits in a permitting queue is a month of interest payments on land, equipment financing, and investor capital. A project that takes three years to permit instead of one does not just cost time. It costs millions of dollars in carrying costs before a single kilowatt-hour is generated. Most data center developers are technologists or financiers. They are exceptionally good at what they do. But permitting is a different discipline, and colliding with a regulatory agency mid-project is one of the most expensive mistakes in development. What Permitting Actually Requires Permits are not just paperwork. They are the product of relationships, sequencing, and a thorough understanding of what each agency needs to say yes. Environmental reviews, utility easements, stormwater management plans, zoning variances, and local land use approvals each follow their own timeline and their own logic. Miss the sequence, and you can find yourself waiting on an approval that was contingent on something you submitted six months too late. I spent 42 years as a civil engineer, split between location design and construction management. That included transportation infrastructure with WSDOT in Portland, Oregon, and water and sewer projects with the City of Wilmington, North Carolina. In that time I learned that the agencies approving your project are not obstacles. They are stakeholders. Understanding what they need, and giving it to them correctly the first time, is the difference between a project that moves and a project that stalls. Why Agrivoltaic Data Centers Are Permittable Solar DC Power's model was designed from the ground up with permitting in mind. Agrivoltaic arrays sited on active farmland in rural Georgia and the Carolinas operate behind the meter with no grid interconnection required. That single design decision eliminates one of the most time-consuming regulatory processes in renewable energy development, the interconnection queue, which in some states runs three to five years. The land remains in agricultural production. The farmer continues farming. There is no conversion of farmland to industrial use, which means the project does not trigger the land use reviews that stop conventional utility-scale solar projects in their tracks. The data center co-located on the same property benefits from power that is generated, stored, and consumed on-site. How a Civil Engineer Assembles the Team Most people think of a civil engineer as someone who designs roads or bridges. That is part of it. But the civil engineer's most important role on a complex project is as the integrator, the person who assembles the right team, sequences their work correctly, and keeps every discipline coordinated from site selection through construction completion. On a data center project, that team typically includes a geotechnical engineer to assess soil bearing capacity and foundation requirements, a structural engineer to design the building and equipment foundations, an electrical engineer to design the power systems, a mechanical engineer for HVAC and cooling, and an environmental consultant to navigate wetlands, stormwater, and any site-specific regulatory requirements. The civil engineer does not replace any of them. He coordinates all of them. One coordination requirement that is easy to overlook is the server cooling system. Data centers generate substantial heat, and the mechanical engineer selected for the project must have specific experience with high-density cooling systems, not just conventional commercial HVAC. The civil engineer works with the architect early in the team selection process to ensure the right mechanical engineer is brought on board before design begins, not after a general contractor has already made that choice based on lowest bid. The architect handles building design and local building code compliance. The civil engineer works alongside the architect from the beginning, ensuring that site grading, utility connections, stormwater management, and access roads are designed in parallel with the building, not as an afterthought. When those two disciplines are aligned early, the permit package that goes to the local jurisdiction is complete and internally consistent. Reviewers can approve it. When they are not aligned, the permit package comes back with comments, and the clock starts over. Water: The Hidden Infrastructure Challenge Data centers have two areas of public contention in rural communities: water and electricity. Both must be addressed early in the planning process, before site selection is finalized and before the first community meeting is held. A data center outside Atlanta recently consumed 30 million gallons of water, effectively creating a localized desert around the facility. Thirty million gallons is roughly six acre-feet, enough to fill an acre-sized pond six feet deep. A pond of that scale on a rural site would likely require fencing and its own permit approval process, adding time and cost to a project already navigating multiple regulatory workstreams. Rural areas present a unique set of sensitivities around both resources. Farmhouses typically rely on private wells for drinking water. Crops depend on ponds, irrigation systems, and pumps that have served families for generations. A data center that strains those resources will face community opposition that no permit application can overcome. Addressing water and power demand transparently, and designing systems that coexist with rather than compete against existing rural infrastructure, is not just good engineering. It is the prerequisite for earning the community's trust. Solar DC Power's cooling approach addresses the water challenge directly. Our system is designed to cool and recirculate water rather than consume it, substantially reducing total water demand. Two options are under evaluation. A closed-loop pond system at reduced depth, approximately three feet, minimizes the permitting footprint while providing thermal mass for cooling. Alternatively, a field of a deep well or wells equipped with filters and a backwashing system could prove cost-competitive and would have the added benefit of providing the surrounding community with a source of clean water, a genuine contribution to the rural areas where we plan to develop. Cost modeling for both options will need to be generated early in the planning process. The civil and mechanical engineers, working with the architect, will be responsible for that analysis. It is one more reason why assembling the right team at the start is not a formality. It is a financial decision. Keeping the Project on Schedule Permitting delay is almost always a sequencing problem. An agency cannot approve a grading permit until the stormwater plan is complete. The stormwater plan cannot be finalized until the grading plan is set. The grading plan cannot be set until the geotechnical report is in hand. Each dependency has a lead time, and an experienced civil engineer knows how to run those workstreams in parallel rather than in series. On Solar DC Power projects, the civil engineer also manages the critical relationship between the energy infrastructure and the construction timeline. The agrivoltaic array, battery storage system, and Bloom Energy solid oxide fuel cells must be designed, permitted, and on track for delivery before the data center building reaches the point where it needs power. That coordination happens between the civil engineer, Bloom Energy, and our EPC partners. It is not something that can be improvised late in a project. It has to be built into the schedule from the first day of planning. The difference between a project that permits in eight months and one that permits in two and a half years is rarely the complexity of the project. It is usually whether someone with engineering experience was managing the sequence from the start. Solar DC Power brings that discipline to every project we plan to develop. It is not a feature. It is the foundation. The Value of Experience There is no substitute for having stood on a job site while a regulatory hold was issued, worked through the resolution, and kept the project moving. That experience lives in the details, knowing which agency has jurisdiction over a drainage easement, how to sequence a stormwater permit with a building permit, when to request a pre-application meeting and what to bring to it. Solar DC Power brings that experience to every project we plan to develop. For investors and data center operators evaluating sites, it means a realistic permitting timeline built on engineering judgment rather than optimism. Time is money. In data center development, it is a great deal of money. Getting permitting right from the start is not a formality. It is a competitive advantage .
By Gerald DeVito May 29, 2026
What a Fish Farm in Spain Taught Me About Energy
By Gerald DeVito April 17, 2026
The Grid Is the Vulnerability: Architecture Eliminates the Risk
By Gerald DeVito April 16, 2026
Reversing Desertification - and What It Means for Agrivoltaics

AI and the Climate

AI and the Climate: The Question Is Not Whether, but How

Post 3 in our series on AI, energy, and the future of rural America.

The last post explained what a data center is: a building full of computers that can draw as much power as a small city. The obvious next question is what all that electricity means for the climate. The headlines tend to pick a side. Either AI is an environmental catastrophe in the making, or it is the technology that will solve climate change. The truth, as usual in engineering, lives in the details.


The honest accounting

Start with the demand side, because the numbers are real. AI data centers are among the fastest-growing sources of new electricity demand in the United States. Utilities across the Southeast are fielding interconnection requests that would have seemed absurd five years ago. Some of that demand will be met by building new gas generation. Some of it is keeping older fossil plants running past their planned retirement dates.

If that is the whole story, AI is a climate problem, full stop.

But it is not the whole story, because nothing about a data center requires it to run on fossil power. The computation does not care where the electrons come from. A processor fed by solar panels does exactly the same arithmetic as one fed by a coal plant. The climate impact of AI is not a property of the technology. It is a consequence of infrastructure decisions, and those decisions are being made county by county, right now.


The grid is the bottleneck

Here is the part that rarely makes the coverage. Even when a data center operator wants clean power, the traditional path runs through the public grid, and the grid is jammed. New solar and wind projects wait years in interconnection queues. Transmission lines take a decade or more to permit and build. Meanwhile the demand keeps arriving.

This is why the behind-the-meter model matters so much for the climate question. When a facility generates its own solar power on site and consumes it on site, the grid bottleneck disappears from the equation. No queue. No new transmission. No pressure to keep an old fossil plant alive to cover the load. The clean generation and the demand are built together, matched to each other, on the same land.

This is the approach Solar DC Power is planning to develop: agrivoltaic solar arrays sized to the data centers they serve, with battery storage carrying the load through the night. The panels go up on working farmland. The crops keep growing beneath and between them. The facility never asks the grid for anything.


The other side of the ledger

There is also a credit column, and it deserves honest mention without overselling it. AI systems are already being used to improve weather forecasting, optimize irrigation, reduce fertilizer waste, design better batteries, and squeeze more capacity out of existing power lines. Agriculture in particular stands to gain, which matters to every farming community weighing whether this technology is friend or foe.

None of that cancels out a data center running on fossil power. Efficiency gains elsewhere do not excuse dirty generation here. But it does mean the technology itself is not the enemy. The build-out is where the climate outcome gets decided.


What communities should demand

If a data center wants to locate in your county, the climate question comes down to a short list anyone can ask at a public meeting. Where does the power come from? Is new clean generation being built for this facility, or is it leaning on the existing grid? Who pays for the infrastructure? What happens to the land?

A project with good answers to those questions is an asset. A project that dodges them is a liability wearing a ribbon. The difference is not the technology. It is the engineering and the intent behind it.

The next post in this series turns to a subject close to home for every farming family: what AI and the infrastructure behind it mean for food security and the future of agricultural land.

Solar DC Power is planning to develop agrivoltaic-powered rural data centers across Georgia, the Carolinas, and beyond. Behind the meter, no grid interconnection required. Learn more at www.solardcpower.com or write to gerald@solardcpower.com.