Thoughts on Energy

The Final Bottleneck

Sir Frank Bowling, Shells Beyond the Seahorse, 2021

Just a few weeks ago, on account of the emergency heatwave that had enveloped most of Los Angeles, the local government’s policies kicked in overnight, shutting off most of the power grid for the day. With temperatures scheduled to range as high as 105 degrees that night, the unexpected outage was met with dismay and outrage, with many citing potential health concerns for the many stuck at home. In a world as modern and developed as found in one of the most prominent cities here in the United States, to experience first-hand such primitive policies was both humorous and shocking. Humorous to see my schools’ libraries packed to the brim with overheated 18-year old’s with no intent to study and shocking in the sense that for many of the developing world, power-rationing remains a viable tool in these countries’ efforts to stabilize power usage.

Despite the shackles that the developing world still finds itself bound to, it is for a far different reason that energy has become the flavor of the month here in the United States. As artificial intelligence continues to extend its tendrils across much of the digital infrastructure, hyperscalers (companies that provide cloud-based infrastructure to exponentially growing user sets/services, thereby developing the necessary framework for AI applications) have begun to utilize gaping amounts of energy in harnessing such forces. Despite most of the world still largely dependent on traditional fossil fuels such as petroleum, there have been increasing calls for the utilization of “bridge fuels” such as LNG (liquified natural gas) as nuclear energy infrastructure remains woefully inadequate for the demands of the present.

While short term and long-term alternatives do exist in the context of the clean energy transition because of the exponential growth of energy consumption, this article seeks to understand as to what result would the development and application of these nascent solutions imply in the context of both the energy industry itself and humanity’s process of development. After all, we are in the business of both predicting and enabling the future as investors. It's clear that as much as we seek (and want!) to enable the growth of the bleeding edge, the current state infrastructure and resources serve as the main constraint for the unleashing of such growth. This brings me to the notion of Jevon’s paradox, an increasingly cited economic observation that serves as the main mechanism of model for the catalysts driving most of the markets/industries in the world today.

Jevon’s paradox essentially argues that because of the pent-up tension between constrained resources and excessive expectations (stemming from a seemingly idealized/fantasized view on what would occur if such resources would become “cheaper”), once resource usage becomes more efficient, consumption of said resource exponentially increases. Furthermore, when this demand is satiated, the dynamics of the industry are permanently altered, resulting in new inherent characteristics.

The logical underpinnings of Jevon’s paradox are sound. After all, if you let your kids into the pantry, who knows how many chocolate chip cookies would remain when you got after work. With energy serving as the last bottleneck for an “intelligent world” capable of unbridled prosperity and innovation, the notion of Jevon’s Paradox increasingly hold salience in today’s world. Alongside labor and intelligence, energy has existed as one of the main factors for understanding and driving growth in economic productivity. As the traditionally defined economic factors of capital taking on a lesser role than “intelligence”, Jevon’s paradox exists has an appropriate framework for what is to come as this “tension” continues to define the short term prospects of the global economy.

The advent of artificial intelligence (just 2.5 years into the release of the first GPT model) represents a landmark moment in the continuum of mankind. While some of the “ai-enabled” wrapper-esque products and offerings remain that of a dubious value proposition (I’ve written about this in my AI:DOTCOM bubble piece), AI has become an accelerant for the role that human capital and labor has had in the modern economy. With gpt4 and o1 performing up to par on conventional metrics (standardized tests, logic puzzles), current solutions boast the frightening capacity to conduct complex analysis and have even created mobile applications; showing the capacity to deconstruct our notions of labor productivity. Unbounded by restraints of the physical and coupled with the modern library of Alexandria that is the internet, artificial intelligence has already “unlocked” two of the 3 fundamental variables underpinning economic growth: intelligence and labor.

This leaves us with the last remaining constraint, energy. For the past few decades, we’ve taken energy as a constant, unaware of the frighteningly inefficient infrastructure/pricing mechanisms underpinning our continuous use electricity and other energy forms. It has only been within the past few years, with heightened geopolitical tension and the advent of artificial intelligence, that the current state of our apparatuses has been re-assessed. As much as the G7 nations have discussed the pressing need to introduce carbon-neutral policies and clean energy alternatives, existing solutions and infrastructure remain both resource-inefficient and blisteringly CAPEX heavy.

However, it is AI that will emerge as the key to the decades long stultification of our infrastructure and energy usage, developing viable infrastructure and alternatives to the current miasma that are our current systems of consumption. Given the achievements that the likes of Mistral and OpenAI have had despite their relatively nascent beginnings, there is no doubt that AI will help to raise or understanding of novel energy sources and minimize the endemic inefficiencies of current systems. In short, AI will help to unlock the third and last bottleneck for humanity to embark upon new horizon, doing so as a function of newfound intelligence and labor rather than the continuous capital misallocation and ill-advised usage of our existing infrastructure.

This is not a question of “if” but of “when”. The prospect of expanding our sources of energy to near unlimited boundaries is revolutionary. With all three of the economic productivity inputs soon to be at scale, it is only a matter of choice as to whether mankind accelerates into the future.

We’ve already discussed Jevon’s Paradox in how improvements of resource utilization result in almost overwhelming consumption, ultimately resulting in a completely new paradigm. More than of the previously outlined resources or factors, energy has historically existed as the chief constraint in technological/economic innovation. Much like the power rationing that is now of common practice in places such as Europe, it has been the result of debilitating energy prices (which in turn, stems from the inefficient infrastructure and solutions) that has often restrained humanity’s exploratory efforts into new frontiers. Ramped up consumption of energy has the capacity to re-shape industries that have historically been stagnant on account of significant CAPEX and input costs.

Radical transformations across infrastructural capacities will enable breakthroughs in sectors such as telecom and agriculture, reflected in significantly different pricing/incentive mechanisms and end-customer bases. With inputs to be held low costs, a radical transformation would occur in industry wide dynamics, traditionally oligopolist/monopolistic dynamics transitioning into one denotated by significant intra-industry competition that is driven by transient competitive edge. One can even posit that rather than existing as standalone entities with overlapping factors, industries could coalesce into holons, or small subsystems within the entire system itself. While it was the rise of software that initiated this transition, man’s soon-to-be domain over labor, intelligence, and energy through artificial intelligence will solidify this secular trend. The result of this would be the development of smaller sub-ecosystems demarcated by solely function that adhere to the same pricing and operations model and remain affected by industry/systemwide factors.  

We have arrived at the point of maximal tension illustrated within Jevon’s paradox. Spurred on by the successes of artificial intelligence, mankind’s expectations of the fantasies and objectives to be accomplished with the advent of clean, cheap energy have exploded in a matter of months.

The new world is here.