Energy Finance (Course Review & ML Application)
Reflections on Energy Finance and its Business Implications
As I wrap up my full-time MBA journey at Carnegie Mellon University's Tepper School of Business, I'm taking time to document and share the knowledge I've acquired and the projects that have honed my skills in preparation for future business ventures. I'm starting with the most recent courses, as these experiences are still fresh and this approach offers the perfect opportunity to reflect on my two-year journey, highlighting my personal growth and advancements.
Energy Finance: A Surprising First
Among all the classes I took during my MBA, the first I wish to discuss is Energy Finance. It might surprise some, as this course wasn't my initial top choice. When I first looked at the course list, I admittedly ranked Energy Finance at the bottom. As an individual who thrives on challenges, I believed this course would merely echo my experience as a Commodity Specialist at my previous company, with concepts like hedging schemes and procurement pricing all too familiar to me. My plan for the final semester was to absorb as much as I could from Machine Learning and Computer Science courses. Hence, I started Energy Finance with slightly less enthusiasm compared to my other choices.
Despite the manageable workload, I deeply appreciate the professor for the well-selected course materials and the comprehensive project. As a commodity trading advisor at Samsung Futures, I had the opportunity to deal with a vast array of energy products traded as financial assets. However, while the global market began shifting towards green energy, Korea and its primary industries remained somewhat obsessed with traditional energy sources like oil and gas. Consequently, renewable energy was somewhat of a blind spot for me (except for the carbon emissions rights traded actively in EUREX). This course served as a bridge, broadening my understanding from conventional energy to renewable sources.
The Business Potential of Standalone Batteries
Our group project focused on evaluating the financial feasibility of standalone batteries connectable to the grid. Batteries are a critical component of the green revolution. The reason renewable energy has been able to replace a significant portion of traditional power sources is due to improvements in battery technology. I firmly believe this trend will persist into the future, which is why I steered our team towards this area of exploration.
Our project revealed the standalone battery business's considerable potential. While we didn't delve into detailed recommendations (being out of our project's scope), I recognized this area as a promising intersection for AI/ML applications.
The standalone battery's primary revenue source is energy arbitrage - charging when prices are low and discharging when they're high. The challenge lies in predicting future peak prices. However, if we could enhance demand forecasting accuracy, the arbitrage space is substantial enough to offset substantial upfront costs.
So, how can we predict more precise time windows with variable daily discharge times? Transformer models could be the solution. These models overcome the vanishing gradient problem and utilize parallel modeling to handle sequential prediction. The energy sector has an abundance of long-accumulated data, which, with proper model design and training, could generate highly accurate predictions. Fortunately, I had the opportunity to study machine learning with large datasets during the S23 semester, a challenging yet rewarding course, especially regarding AI/ML model scaling, like the GPT model. The course taught us how to scale models to handle large amounts of data. Given the success of transformer-based models like ChatGPT, I'm confident that a well-trained model could make highly accurate short-term power demand predictions. Theoretically, the daily price gap could be as wide as $80 per MW, which could push the IRR of a simple battery business over 20% - highly lucrative indeed!
Navigating the Business Landscape: Funding and Risks
Starting a business in this domain has another advantage: funding is relatively accessible despite
the ongoing banking system issues and rising interest rates. Thanks to tax equity financing, this business can attract investors looking for an initial tax credit as a return on investment, significantly lowering the burden of initial investment.
Lastly, it's crucial to touch on the risks associated with this business. Both upside and downside risks exist, with government policy changes being the most significant factor. This is an ongoing issue, as depending on where the battery is located, not only the charging price but also the subsidies can vary significantly.
One compelling example is Texas. Texas's less regulated energy market makes it an attractive location for standalone battery operations. The state has a unique energy-only market design, which allows for high price spikes, making it an ideal environment for energy arbitrage. The high potential for profitability coupled with the state's commitment to renewable energy makes Texas an exemplary region for individual battery businesses.
Powering the Future: The Untapped Potential of Battery Storage and Distribution”
In the push for a green revolution in the energy sector, a substantial amount of funding is dedicated to power generation. However, I am convinced that the efficient storage and timely distribution of power are as critical as harnessing energy from renewable sources. Given the high potential for improved efficiency and capacity in the battery sector, it promises to be a compelling business model.
Looking forward, I am eager to revisit and update this post, whether it's because I have embarked on my own business venture or helped others to start theirs. Regardless of the outcome, I want to conclude this reflection by expressing my profound gratitude to my professor and my teammates. Their invaluable contributions have greatly enriched my learning experience and have laid a solid foundation for my future endeavors in the fascinating intersection of energy finance and technology.