The Value of Necessity: Now it Gets Hard…But That’s Alright (Part IV)

Author’s Note: This post will serve as the fourth of several related blog posts on this topic. This series is meant to look at the current realities facing land developers, acquisition agents, and investors in the energy infrastructure sector when seeking to design, develop, and operate projects which produce or generate power within the United States. Instead of studying this from the perspective of each energy sector individually, it instead looks at the common conditions, dangers, potential headwinds, and hoped for tailwinds faced by all such endeavors. This post is meant to examine the notion that land acquisition and energy infrastructure projects must exist in a world where the human element and rapidly developing artificial intelligence-based models are effectively two sides of the same coin. Each will both augment and challenge the other as we quickly head toward a world where a land developer not properly and fully utilizing both will not be able to compete with a developer who implements a truly bilateral approach.

This discussion is meant to go beyond the utilization of A.I. models for back-office support such as human resources compliance controls or accounting checks and protocols (even though they offer massive efficiencies and associated savings there as well). This discussion will focus on how A.I. and humans can work together to revolutionize the way that energy infrastructure and development projects identify, acquire, and manage the land necessary for their construction and operational needs.

Humans and Artificial Intelligence: Together On a Parallel, Not an Intersecting Course

       Intelligence. Data gathering and evaluation. Decision making ability. These abilities and tasks were once the sole purview of humankind, but increasingly sophisticated artificial intelligence-based models and systems are beginning to challenge that monopoly. This post is in no way meant to imply that A.I. will take over every task currently handled by humans or that humanity itself is doomed. Nor does the author believe that artificial intelligence will end all forms of human employment any more than Microsoft Windows-enabled computers threatened to end large scale employment in the 1980’s. Instead, the purpose of this discussion is to look at the changes that A.I. will likely bring to the land acquisition, management, and development process, and to consider how these new capabilities might enhance the role of the actual humans in the process.

       The initial presumption is that artificial intelligence is not by itself either a force for good or bad. In truth, artificial intelligence, as it is currently constituted and enabled, is not actually “intelligence” as most would define the term. Even more advanced agentic A.I. models cannot fully emulate and reproduce the immersive and wholistic problem solving approach offered by humans of even modest creative abilities. The hallmark of an intelligent being is not only the ability to gather and process information but to understand and appreciate that information as well. Humans have emotions, a conscience (most of us), social intelligence, and the ability to approach problems and solutions creatively through non-linear thought progression. Artificial intelligence, by contrast, is currently limited to gathering, collating, and analyzing vast stores of data on a scale and at a pace so far beyond human abilities that it is beyond the capacity of most humans to truly comprehend.

       The result produced by this process is not that of a computer “thinking” its way through a problem, but rather the mathematical solution to an equation or algorithm applied to a unique set of data and variables. The larger the data set applied, the more precise the question asked, the more times that similar queries have been run in the past, the more prior results available to review, and the better the subsequent corrections the more accurate and useful the result. But as sophisticated and truly amazing as this new reality is, it still cannot fully supplant the intuitive, creative, and contemplative human element in the decision-making process. This means that for at least the foreseeable future, artificial intelligence and its advocates must accept that humans still have an equally integral place in the process as well.

       Humans, for their part, have three options when confronting this new reality. They can (1) bury their heads in the sand and hope that this passes over them without causing too much damage, (2) they can rage against the winds of change for all that will gain them, or (3) they can adapt to this new reality and find a way to make it profitably work for them. The first option is nothing but a false hope. Artificial intelligence in ever greater scope and sophistication is here to stay. Ignoring that or wishing that it would go away will get you nothing but left behind. As for the second option, go down to the beach and try holding back the tide. You will make just as much headway. This leaves the third option as the only reasonable alternative left to those who seek to have some level of control and success in their lives and careers.

       Humans must learn to cope with ever greater machine process presence and augmentation as a fact of life. There is not a job or career in the United States that cannot be impacted by the introduction of artificial intelligence. The use of the term “impacted” as opposed to “eliminated” is quite intentional. The fear is that eventually artificial intelligence can eliminate great swaths of the American employment landscape. Perhaps, but that would be the lazy approach on the part of managers, boards, companies, and investors.

       Instead, those responsible for growing America’s companies and industries should look to see how this new and developing addition to the corporate toolkit can augment their existing work force. How can employees and managers be empowered and supported by a greater and more comprehensive toolkit than at any other time in human history? How can a company or endeavor be designed and supported to advance by exponential leaps instead of timid steps?

       Employees have a role in this transition as well. If they are working at a well-managed company, then they should have the opportunity to demonstrate that the incorporation of artificial intelligence into their roles and responsibilities will only make them far more effective, efficient and therefore valuable. If an employee has cause to believe that this opportunity will not be available to them in their position or at their current employer, then they might wish to explore alternate options.

       When tying these concepts and questions to the land acquisition and development process it is therefore best to see the human factor and artificial intelligence not as separate, competing elements but rather as intrinsically interconnected elements focused on solving the same problem. Essentially, they are two sides of the same coin. There are several immediately identifiable functions in the land acquisition process where this necessary coordination between human and machine can profitably coexist.

The Negotiation Process:

       Fundamental to any land acquisition and development program is the financial offer and term determination process and the follow-on negotiations with the land or rights owner. Until and unless this process is successfully completed a land developer, no matter how well intentioned, supported or funded, has nothing but an idea on a whiteboard. Playing on the two-sided coin metaphor, this is a process that relies and even thrives on both the human and the artificial intelligence contributions being present and valued.

The Human Contribution

       Where we live matters a great deal to our interpretation of the world around us. The “where” influences who we become just as much as do our family and our educational experiences. While there are many points of similarity to those living in the same nation, basic geographic and community level differences factor into our development. The vast majority of individuals living in America interact with the world using the English language (either as their primary or secondary language), they attended a public school through their eighteenth year, watch the same movies, television shows, root for or against the same teams in the same professional sporting leagues, and we all pay taxes to the same federal government. But a child born to parents living in a small town of 1,500 inhabitants in rural Kansas will have a different lens with which to view their world then will a child who grows up in a large subdivision containing 2,000 homes and 8,000 inhabitants, and their experiences will in turn differ from a child raised in one of the nation’s large cities. These children can salute the same flag, play the same sport, study for the same grade school exams, and even worship the same god in the same denomination church, but grow up viewing life very differently.

       This inherent distinction in viewpoint which naturally and necessarily develops between individuals raised in different types of communities is an essential factor to consider and respect when seeking to acquire, use, or encumber land. It is also a conceptual distinction that is best appreciated and exploited by empathic humans and somewhat lost on an artificial intelligence constituted only by quantifiable, empirical data (regardless of how accurate they might be). The computer is forced to reduce the world to a series of algorithms and probabilities. It can only calculate the mathematically most advantageous end result and put that forward as a reasonable initial offer to a counterparty. A human sees the individual behind that data.

       A human understands that the person on the other side of the offer is only partially a creature of logic. Humans, even the most dispassionate amongst us, are only partly swayed by logic alone. A person is the sum total of their upbringing, their hopes and fears, their successes and their failures, and their simultaneous debts and responsibilities to the past, present, and future. It is very difficult to accurately reflect all of this in an algorithm.

       Artificial intelligence can reasonably and accurately tell you that the fair market value a given tract of land two states away from your office is worth $1,000 per acre and by all quantifiable logic it will be correct. Artificial intelligence cannot, however, account for that tract’s owner having to reconcile that number with the four generations of sweat his family put into it nor can it properly account for the future value of that tract to his son’s grandson. Landowners do not care what their tract of land is worth to you. They only care about the tract’s value to them. If this differential cannot be amicably or reasonably bridged, then a negotiation will fail. Humans, again, even the most logical and dispassionate amongst us, are at times irrational beings willing to act against our own seemingly best interests. This reality is sufficient to add a statistically significant likelihood of inaccurate outcomes to even the most precise and well-conceived artificial intelligence-based acquisition models.

       This means, at least for the foreseeable future, there is a need for the human element in at least the negotiation and consideration phase of the land acquisition process. There is a need in the process for a human to add that last degree of finesse reasonably calculated to craft an offer or terms that will be well received and acceptable to the land or rights owner. The final push that every deal needs to get across the finish line is for each counterparty to blink and exhibit at least a modicum of trust in their opposite party. No computer program can provide that minimal but required level of trust that convinces a landowner to trust. No computer understands or can offer the incredible power of a handshake. Without this human touch the odds of a successful transaction will not be as great as it otherwise could or should be.

The Artificial Intelligence Contribution

       As essential to the offer and negotiation process as the just discussed human element necessarily is, the dispassionate logic and raw computational power offered by a well-designed artificial intelligence-based model is equally foundational to a successful land acquisition and development program. Good decisions are predicated on good data and good data assumes a thorough intake and review process. Assuming all data is equally accessible and accurate, a well-conceived and programmed artificial intelligence-based model will produce vastly superior results compared to a strictly human powered review process in every instance and it will do so in a fraction of the time, and, eventually, at a fraction of the cost. Data, as it relates to land, can cover many facets of the acquisition program.

As to Land & Tract Valuation and Comparative Sales Data

       Land values are not per se secret, but neither are they necessarily widely publicized. There are public databases that offer recent transaction data, but they are often incomplete and not always timely updated. Fee-based, private databases normally have somewhat better and more timely data, but they too are subject to sales and pricing data being made available by the parties to a transaction. Alternatively, local realtors will normally have a good understanding of general land values and availability within their locality. Perhaps the best source of data is the local landowner. There isn’t a farmer or rancher in this country who cannot accurately tell you the value of every tract within walking distance of their front gate.

       But these sources are inherently and intrinsically limited in their worth. All real estate, famously, is local. The best realtor in Hallettsville, Texas, will be of no help to a buyer in Fort Dodge, Iowa. An understanding as to the value of pastureland in the Oklahoma panhandle does not confer similar insight to an almond grove in California’s Central Valley. Even nationwide realty companies are just a collection of separate, hyper-regional offices with very little internal connection or data sharing ability (or need) beyond required corporate metrics. It is very difficult for a human to possess the requisite knowledge or to be able to readily access the resources necessary to make land valuations in locations and regions beyond their immediate vicinity. This, in turn, leads to a great deal of errant land value data becoming the basis of project economic planning with obvious and sometimes disastrous results.

       Sourcing, classifying, and evaluating land sales and valuation data therefore becomes a task essential to any land acquisition or development project. A human can be taught to complete such a task and over time they can find efficiencies that will allow them to increase their review speed. But after the initial period of education and experimentation, during which a human can rapidly expand their capabilities, their continued progress will be linear in nature. Once peak efficiencies are approached by that human actor, a period of diminishing returns is usually reached as either boredom and complacency manifest in the form an unmotivated employee or the high-performing employee is identified, rewarded, and promoted to new duties thus necessarily forcing the learning curve to begin anew with another individual. Programs managed by artificial intelligence do not suffer the same limitations. These programs, once properly dialed in, can search a vast number of individual transaction records spread across multiple databases in a fraction of the time it would take a human to accomplish the same task. As more records are digitalized and made available for public access this capability only grows. Once a model has been refined and to the extent possible perfected, these sorts of data studies can be run repeatedly with very little if any loss of fidelity.

As to Landowner Specific Data

       Land data also relates to the identification and understanding of the landowner themself. Just as no two tracts of land are truly alike, no two landowners are either. The ability to search out, identify, and develop an accurate profile on target landowners is an area in which artificial intelligence models can bring great value to a land acquisition project.

       When working in the oil & gas or hard mineral mining sectors, landowner identification is, in one sense, a less onerous task than when working in the renewable energy, geothermal, or carbon sequestration sectors. In the former, the deposits of economically recoverable resources will be found under certain tracts of land but not under others. Of course, identifying exactly which tracts fall into either category is part science and part art, but the basic fact is that these projects are tied to a specific geographic location. In those instances, the developing company does not have a choice as to which landowners it must negotiate with. If the landowner is fortunate enough to possess a mineral rich tract of land, then they must be negotiated with, and their interests must be acquired in some fashion. Conversely, renewable energy, geothermal, and carbon sequestration projects, while certainly having identifiable and important land metrics themselves, are far more forgiving in terms of specific geographic location. If enough suitable acreage can be identified and acquired for these projects, there is less concern as to precisely where they are located within a town, county, or region. While this reality does confer certain advantages on such project types, such as the ability to walk away from a problematic landowners when necessary, it also introduces a level of complexity to site prioritization and selection not always found on more rigid site-specific land acquisition programs.

       In the modern age all but the most reclusive of us have some social media presence. At the very least we have some identifiable online presence that can be researched and harvested for information. Past purchases, real property transaction records, interview transcripts, social media posts, inferred biases, and countless other breadcrumbs exist in perpetuity to connect an individual’s past behavior, likely current attitudes, and probable future decisions. We are effectively the sum of our past decisions, and those decisions are quite often available to view if a researcher or model knows where to look. Artificial intelligence models, with their ability to rapidly search through vast troves of public data, are superior tools with which to gain the measure of your counterparty. Knowledge is power; certainly, at least the power to conduct negotiations at a distinct advantage relative to a landowner who likely knows very little about you and doesn’t have the same resources to conduct a similar in-depth search.

       When project siting requirements are so malleable and forgivable that in effect a project could theoretically be developed anywhere, it means by extrapolation everywhere becomes a theoretical development location. This theoretical reality can add hundreds or even thousands of potentially suitable landowners to a land acquisition program. It is exceedingly difficult for a developer and their associated vendors to reduce this overall number of potential targets to a more realistic target list. When such a reduction is not handled in a competent manner a great amount of waste and delay will be introduced into the development process as too many landowners are identified, approached, and brought into negotiations. This can greatly increase project costs in terms of both funding and scheduling. Further, and likely a greater threat to overall project success, is the worry that by means of this inefficient process the best landowners and tracts will be lost amidst the noise of the other lesser potential options.

As to Tract Specific Data

       As discussed above, the value of a tract of land and any available insight into the landowner matter a great deal. Artificial intelligence-based models do a great job of capturing and organizing this type of information. Of equal importance is any available data regarding the tract of land itself. Artificial intelligence-based models also provide a superior product here as well. Sometimes the land acquisition is necessary only to serve as the physical site of the intended project. In such cases there will be required parameters which must be met in order for the tract to be of use to the developer but beyond those requirements there will be no specific significance given to anything else about the tract. In such instances, slope and gradient, soil compaction, foliage and coverage, distance from roads, railways, and population centers, previous land use, presence in or near a flood plain, and other definable physical characteristics will control the acquisition process. Sometimes, however, the value in a tract of land comes not from its location and surface characteristics, but instead from the presence (or absence) of valuable minerals and other resources which can be extracted from beneath the subject tract. In either type of land acquisition A.I.-based modelling will bring a significant benefit to the project.

       For any development projects in which the developer is concerned with the physical characteristics of either the surface of the subject tract or the minerals contained on and under that tract, the benefit of artificial intelligence will come from the ability to rapidly sort through vast volumes of land data, isolate pertinent facts and data, and efficiently and quickly position the tract into an established prioritization scheme.

       There are vast troves of free public land data that contain reams of useful information describing all of the factors mentioned above. Federal government entities such as the USDA, the EIA, NASA, the EPA, the BLM (to name but a few) and the Departments of Agriculture, Commerce, Energy, Interior, and Transportation all provide pertinent information that can be utilized in land acquisition development programs. All of this information is made available to the public free of charge. Every state collects, maintains, and publishes their own data as well, but the quality varies greatly from state to state. Oil, gas, and mineral production history and forward-looking permitting activity can be researched to understand what competition (or potential partners) exist in the vicinity. County level ownership title and probate information can be researched quickly and efficiently to create ownership reports (the bedrock of all acquisition and development efforts). Tax rolls denote both ownership and activity. There is no shortage of raw data available. Knowing how to utilize this data in imaginative ways is the key.

       Because government data can sometimes be grouped, stored, and managed in a cumbersome manner, many private companies have entered the space to facilitate data identification, acquisition, and usage for a fee. Many of these companies are backed with tremendous financial resources and big-name investors. Jeff Bezos and Bill Gates recently led a consortium of investors to raise $537 million to spur the development of KoBold Minerals (the private company now has a total valuation of over $3 billion); a Berkeley-based company which utilizes artificial intelligence-backed search programs to look for rare and critical minerals all over the globe purely via A.I.-based record reviews. While this is an example of an extremely well-funded private A.I. driven enterprise, there are many such entities seeking to bring similar capabilities to the market, albeit on a smaller scale.

       Humans do not have the ability to even approach the capacity of artificial intelligence-based models to work through and efficiently and accurately utilize the immense troves of raw land data available in the world today. This is one instance where people should not try to compete. Instead, admitting the vast superiority of employing A.I. models in this manner, humans should instead harness this newfound power to bring a level of cost discipline, speed, and accuracy to land acquisition projects such as has never before been possible.

A Combined Approach Wins Through in the End

       As asserted at the beginning of this section, the data gathering and follow-on negotiation process is best managed when both human and artificial intelligence driven aspects are present. The raw computing power coupled with the ability to rapidly search every nook of the internet sets artificial intelligence powered data gathering into a realm never before possible. Decision makers now have the ability, with only reasonable inputs as to cost and process evolution, to have reams of accurate data available to them during the planning and conceptual stages of a development project. When done correctly, this allows project developers access to a toolset superior to anything available in the past. When this advantage is coupled with creative, qualified, and experienced individuals, land acquisition campaigns are advanced to a previously unattainable level of accuracy. When properly managed, this should lead to cost efficiencies and savings that can be either redirected to other project phases or returned to the project’s bottom line.

       Going forward, companies will either adapt to this new bifurcated approach to project responsibilities and become a faster, leaner, and more efficient competitor or they will ignore it at their peril and suffer the consequences for having done so. In a few years’ time it will be very difficult for a development firm not utilizing this approach to realistically compete with one that is. Project design, research, and development support from artificial intelligence models is not going to end humanity or the need for human interaction, or our relationship building abilities, or our curiosity, but it will create winners and losers in the process. Those firms and individuals who choose to adapt and master this new world of data gathering and offer creation will quite simply outcompete those that do not.

Building a Smart Energy Program For an America of the 21st Century:

       This year the nation celebrates its 250th birthday and the changes witnessed by humanity between 1776 and 2026 are starker than during another other period in human history. At the nation’s birth illumination at night was accomplished with a combination of hearths, candles, and mirrors, but for the most part, people just accepted that darkness was a part of their daily life. Horsepower was indeed a measure of work accomplished over a given period of time, but it was actual horses being described as the nation was created just prior to the dawn of the Industrial Age. This nation cleared its first fields, cut its first roads through the forests, and built its first cities using human and animal muscle (and a surprising understanding of physics). Cooking, cleaning, and house chores were laborious and done with strong forearms. The unloading of ships at America’s port cities and the planting and harvesting of her crops were likewise laborious and done with strong backs (sometimes willing and sometimes forced). America at its birth was built on energy; human, horse, oxen, and wind energy.

       As the nation progressed into the 19th Century, mechanically derived forms of energy entered the mix. Steam engines allowed us to cross the breadth of the nation, elevators allowed us to push buildings into the sky, and steam powered factories began mass manufacturing at a scale soon to shock the world. By the end of the 19th Century this nation would lead the world in technical innovation fostering a concentrated Age of Invention never before seen in human history. While the century came in powered by muscle and wind and transitioned to being powered by coal and wood fired steam, it ended with spinning turbines generating electricity.

       The 20th Century witnessed humanity change in ways barely imaginable during any prior period. In a mere 70-year period (1903 to 1969) this nation bridged the gap between humans never having left the surface of the earth to one small step for a man that was truly a giant leap for mankind. Prior to the dawn of the 20th Century no human in had exceeded a speed of 82 miles per hour (two speed record attempts conducted by the New York Central Railroad in 1891 and 1893) with most “high-speed” express trains only averaging 60 miles per hour. Within three generations the Apollo 11 astronauts achieved a peak speed of 25,000 miles per hour during their Earth departure burn. This exponential leap occurred as scientists learned to unlock and then harness the raw power contained in oil, gas, and other natural elements (liquid oxygen and kerosene in the case of the Apollo Program’s Saturn V rockets), and by splitting atoms themselves.

       These new power sources allowed humans to achieve once unimaginable goals while also transforming the everyday lives of people around the world. Average citizens now traveled well beyond their homesteads and towns for employment allowing median family incomes to soar compared to previous generations. Companies could decentralize and spread their operations out as needed to grow into multi-state and eventually multi-national monoliths drawing together the world’s disparate workforces and manufacturing capacity. The world leaving the 20th Century was smaller and more interconnected than at any other time in human history.

       But as technologically powerful and economically productive as the nation and world were leaving the 20th Century, they were really just maximizing the tasks, goals, and dreams of the previous millennia. The dawn of the 21st Century would introduce foundational changes never seen in human history that would have the effect of, in many ways, not only impacting their human creators but of changing humans at a fundamental level. The first quarter century of the new millennia will be remembered as the dawn of the Information Age. Yes, armies, bullets, and missiles all still represent power in its most raw and brutal form but increasingly access to and control of information represents true power. Whomsoever possesses the most useful information typically comes out ahead economically, socially, politically, and militarily. Information requires not only technological superiority but vast amounts of energy as well.

       It has therefore become axiomatic that continued economic growth and progress relies on concurrent growth in energy availability and affordability (so much so that energy price spikes have served as precursors to most economic recessions since World War Two). The raw data is out there for the taking and waiting to be analyzed, but exactly who will make the necessary commitment to build the data centers, and to create the necessary regulatory framework in which A.I. can grow with only the minimum required fetters, and who has access to sufficient energy are the questions facing this generation.

Artificial Intelligence and a Smart Grid: A Symbiotic Relationship

       Artificial intelligence models, in sum total, require vast amounts of energy constantly available to the computational process while the smart grids responsible for providing that power can utilize artificial intelligence to better understand the load and supply fluctuations affecting their own cost and growth prospects. Roman logicians would have referred to this as “circulus in probando” – a circle in proving.

       The holy grail in domestic and global energy development is understanding what sources are available to provide power while simultaneously understanding what and who will use that power once available, when and for how long it will be needed, and in what quantity it will be needed. Complicate this with some energy sources falling off (coal), some new ones being brought online (solar and wind), some decreasing in costs (natural gas) while others increase in costs (nuclear), entirely new demand sources constantly appearing and then striving for new efficiencies, and the intermittency issues inherent with renewable energy sources and you quickly see why there is no reliably accurate way for humans alone to forecast actual market needs.

       Domestic interconnection queues and global energy markets are constantly shifting targets for players seeking to enter or take from those markets. While this inefficiency creates potentially lucrative arbitrage opportunities, it makes the job of sourcing and providing adequate power in modern, complicated, and growing economies difficult at best. When this process fails, the inevitable brownouts, blackouts, and uncertain utility infrastructure support results in the inability for individuals, investors, and companies to reliably make plans for future development and expansion and the associated loss of productivity and revenue growth. Artificial intelligence models, with their speed, reach, and unrivaled ability to process massive amounts of data, can bring more certainty to this process than humans ever could thereby potentially providing the power and infrastructure certainty that is required for continued economic growth.

       The primary responsibility, and most difficult task, for any electric system operator is to balance available supply with current demand at all times. The more information and advance warning of non-routine events that an operator or utility can have, the greater the likelihood that they will be able to fulfill their obligations. Americans have grown accustomed to being able to turn on a light, or the heat, or air conditioners on a whim without themselves giving thought to what it takes to allow those actions to take place. To accomplish this, in part, accurate weather forecasting is imperative. It is the grid operator that needs to reconcile an incoming hurricane or blizzard and the expected consequences with what variable energy generators (e.g.: wind, solar) and which dispatchable or firm energy generators (e.g.: nuclear, natural gas, and coal fired plants) are online, are able to be brought online, are to be held in reserve, or are down for maintenance or intermittency. This is difficult enough, but when the normal and frequent track, speed and severity of forecasts constantly change it can be very difficult for the humans running utilities with existing algorithms to keep up.

       Beyond unscheduled intrusions of weather, human activities such as large sporting events, concerts, or rallies can temporarily displace a large percentage of a city’s population even if just for a few hours. For a period of time, these shift where on a grid supply is needed.

       New demand sources are also constantly putting new strains on the power grid. In 2018, data centers consumed approximately 1.8% of the electricity generated in the United States (approximately 85 terawatt hours of electricity). By 2024 this single demand source had increased its consumption to 4% of the electricity generated in the United States (approximately 195 terawatt hours of electricity), and this is expected to grow to over 425 terawatt-hours by 2030. Highly sophisticated new manufacturing plants being on-shored unfortunately employ far fewer workers than in the past, but as a resulting offset, consume far greater quantities of electricity to power their computers, automated assembly lines, and robots. By 2023, electric vehicles made up 9% of all light duty vehicle sales and for the first time in history their cumulative charging needs consumed more electricity that all light rail systems in the United States combined. The U.S. power grid is not capable of handling the current rate of demand growth, let alone the project exponential increases projected to occur in the next decades.

       We do not as a nation want to impede this growth curve and there are hoped for and expected efficiencies which will likely offset some of the projected demand growth, but the story of the next several decades will necessarily be one of bringing online new sources of power and electricity generation and enlarging and modernizing the nation’s existing transmission and pipeline networks. While some of this will come from enlarging or recommissioning recently retired nuclear, natural gas, and coal power plants, much of this growth must inevitably come in the form of entirely new or expanded oil & gas production fields, utility scale wind and solar generation facilities, new nuclear power facilities, and an expansion of the nation’s hydro and geothermal capacity.

       Humans have managed to adequately respond to the evolving supply and demand situation well enough throughout the years, but the complexity of the coming decades as the nation, its companies, and its people race into the new age of information will require sophisticated planning and oversight that only artificial intelligence-based models and computing can afford. Much of the siting and planning work of the past decades have focused rather haphazardly on where it was easiest or most cost effective to bring new production online. Utilities struggled mightily, and often blindly, to select the best projects entered into their interconnection queues frequently selecting projects based either on purely financial metrics or on those most likely to be built as opposed to selecting those of the right size and location to bring the most benefit to the grid and its customers.

       There are approximately 1.9 billion acres of land in the United States. Only about five percent (5%) of that total has been “permanently developed” into the nation’s urban and suburban areas. This means that the vast majority of acreage in the nation is at least theoretically open to further energy and infrastructure development. The key hurdle that project developers and land acquisition agents face when trying to site projects measured in the hundreds or thousands of acres in a country measured in billions of acres is sorting through and prioritizing the overwhelming choice and selection of tracts they inevitably face. Any advantage that project developers can access when trying to site energy and infrastructure projects is well worth the cost and effort.

       A.I. based programs can better sort through and analyze the queued projects to select those that will bring the most benefit and efficiency to the system. This will lead to a revolution in the way that new production and generation projects are conceived, planned, and sited. Target locations can be reduced in number and refined to only those locations critical to greater development plans. This now reduced number of sites can be further analyzed and prioritized by A.I. based modelling to account for preexisting or planned competing energy, infrastructure, manufacturing, or agricultural activity, the physical characteristics of the various tracts, any known seismic, geologic, or historical mineral data, and, potentially, even landowner and community sentiment. Development companies and those individuals tasked with acquiring these prioritized sites can focus in and fight for these sites knowing their importance as opposed to constantly being willing to walk away once acquisition and development become challenging or difficult.

       With a firmer understanding as to the greater picture, as opposed to being myopically focused on a single development project or site, A.I. based programs can more readily and accurately determine what pipeline and transmission lines need to be enlarged, rerouted, or started new to better connect the production and generations sites with the refineries and load centers in need of their products.

Concluding Thoughts

       As stated at the outset of this discussion, humans and artificial intelligence-based models will be intrinsically linked as though two sides of the same coin going forward. It is quite likely that there will never again be a time in human history where humans act alone without the support of A.I. And it is equally likely that it will be an incredibly long time, if ever, before humans can ever truly be removed from the equation.

       Companies and developers seeking to plan the next generation of production and generation sites, along with the infrastructure necessary to support them, need to take into account the strengths and weaknesses of both the human and the synthetic elements of their projects. Only when they do this will this industry and this nation be able to meet head on the challenges they face in the coming decades.