Power Switch?
DeepSeek, a Chinese AI upstart rocked the tech world last weekend with the introduction of an open-source large language model (LLM) exhibiting less compute intensity than the leading and legacy US LLM efforts such as OpenAI. The stocks of firms enabling current, and most importantly future data center development and LLM processing such as the graphics processing unit (GPU) companies suffered ten percent plus losses on Monday, January 27 under the premise US AI efforts would be usurped by Chinese technology requiring less infrastructure. The Essex Global Environmental Opportunities Strategy (GEOS) has significant exposure to the power technology sector, and a few holdings had losses given the market reaction, which we believe is an overreaction as the trends supporting our power technology theme are well intact.
Power enables our economy to operate and grow — domestic electricity usage has grown 100-fold over the past 120 years, driving the productivity of capital, labor, and raw materials for domestic economic growth, primarily in the manufacturing sector (source: EIA, Boston University Visualizing Energy). While power demand has grown less than GDP over the past few decades due to various energy efficiency efforts, our current assumptions are for a meaningful upward inflection in annualized power demand for the next several decades. While DeepSeek could present a reality check for data center development, the trajectory for power demand in the US and other markets is still strongly positive. Power demand will grow based on plans for increased domestic production – onshoring, and electricity demand will be enhanced by the intact move to electric vehicles – a topic for another post. One of the most important drivers for power technology is the need to invest in a resilient and reliable electric grid which includes the construction of distributed technologies such as solar and storage. To optimize grid efficiency and effectiveness, flexibility and interconnection are of paramount importance, requiring both hardware and software solutions. Examples across our power technology theme range from electron measurement to grid weather hardening. While utilities have taken the path of least resistance, avoiding grid investments over the past few decades to optimize their returns on capital, they are now sounding shrill alarms that power prices will continue to increase as they deploy capital to grid distribution amidst increased demand.
Commercially viable technology transitions occur because they enhance the quality of life and economic growth. A more robust and enhanced power technology platform is the very definition of enabling economic growth with fewer resources – our definition of clean technology. The DeepSeek release will continue to drive our domestic AI competitive posture, enabling technology development and adoption that will enhance productivity for more applications. In the history of technology development, efficiency gains lead to increased power consumption – the Jevons Paradox. Machine learning is already in use by most of our GEOS portfolio companies, from lidar applications that automate warehouses to leak detection solutions for water utilities. Our outlook for AI is constructive, not draconian, as it enhances productivity. This all rests firmly on power technology. Without power, we have no economy – as we have written, GDP = BTU (British thermal units of energy). But let’s use modern BTUs, and optimize them – cleaner and safer solutions, rather than dirty, obsolete methods.
Disclosures:
This commentary is for informational purposes only. It does not constitute investment advice and is not intended as an endorsement of any specific investment. The opinions and analyses expressed in this commentary are based on Essex Investment Management LLC’s (“Essex”) research and professional experience and are expressed as of the date of its release. Certain information expressed represents an assessment at a specific point in time and is not intended to be a forecast or guarantee of future results, nor is intended to speak to any future periods. Accordingly, such statements are inherently speculative as they are based on assumptions that may involve known and unknown risks and uncertainties.
This does not constitute an offer to sell or the solicitation of an offer to purchase any security or investment product, nor does it constitute a recommendation to invest in any particular security. An investment in securities is speculative and involves a high degree of risk and could result in the loss of all or a substantial portion of the amount invested. There can be no assurance that the strategy described herein will meet its objectives generally or avoid losses. Essex makes no warranty or representation, expressed or implied; nor does Essex accept any liability, with respect to the information and data set forth herein, and Essex specifically disclaims any duty to update any of the information and data contained in the commentary. This information and data does not constitute legal, tax, account, investment or other professional advice. Essex being registered by the SEC does not imply a certain level of skill or training.
Please find important disclosures here.