Quarterly Review and Outlook Third Point funds delivered solid results in the First Quarter, driven in part by the strong performance of large cap tech companies as well as continued gains from Bath & Body Works and Vistra. Just as during COVID when everyone became an expert in virology, today you would be hard pressed to find an investor who does not have a complex view on “macro.” Last week was a good example of equity investors’ currently mercurial nature, with the market experiencing sharp swings based on interpretations of noisy and inconclusive reports of economic data. While we don’t attempt to make short-term “macro” forecasts, our current framework is to look at three important dynamics—two economic and one structural—to guide our current allocations and where to look for new opportunities.
1) Rates and inflation: We believe that we are no longer in an inflationary economy based on our analysis of labor, rents, and other key components of inflation and thus, absolute and real rates are near the top of their ranges.
2) Economic growth and earnings: We are in the soft-landing camp and see some impending weakness in labor which could impact demand in certain sectors, but expect that future Fed action will temper the extent of future economic declines.
3) Artificial Intelligence and Energy Transition: We shared our views on AI’s transformational potential in recent letters and it is a key element of the thesis for nearly half of our equity positions today. Unlike in past periods of technological paradigm shifts, this new technology favors incumbents who are deploying their financial and intellectual war chests to win the AI arms race. Right now, what we see as the best-run “legacy” companies like Microsoft and Amazon (both of which we own) have built enormous competitive advantages and seen their growth vectors accelerate. Below, we discuss London Stock Exchange Group, Alphabet and TSMC, positions whose catalysts for further value creation are primarily “AI-driven”.
We also see the energy transition and growth in data centers affecting scores of industrial, materials and energy companies as demand for infrastructure and certain commodities surges. In the category of “better to be lucky than smart,” below we also share how a deep value name we acquired for its capital allocation and attractive valuation became an AI darling.
Vistra
Vistra is one of the largest independent power producers (“IPPs”) and retail electricity providers in the country. In 2023, Vistra’s natural gas, nuclear and coal plants generated over 20% of electricity consumed in Texas.
Unlike regulated utilities, where profits are determined by capital invested, Vistra operates in deregulated markets (primarily ERCOT and PJM), where they generate and sell electricity at market prices. Historically, Vistra has been valued at a steep discount to both the regulated utility sector and the broader market in part due to the challenging fundamentals of merchant power. Stagnant domestic electricity demand combined with an oversupply of natural gas has made US electricity prices among the lowest in the world. Meanwhile, significant growth in subsidized renewable generation has created major intraday price volatility in Vistra’s core markets, with power prices sometimes going negative during periods of abundant sunshine or wind. Bankruptcies, including Vistra’s former parent company TXU in 2014, have become commonplace in the sector over the last decade.
In response to this challenging environment, we believe Vistra’s capital allocation strategy has been brilliant. The company shut down unprofitable coal plants to improve its carbon footprint and mitigate oversupply. Given the market was valuing its remaining gas assets at pennies on the dollar relative to the cost of new builds, management patiently invested in maintaining the existing fleet and deployed excess cash flow into share purchases, reducing its share count by ~33% from 2018 to 2023 at an average purchase price of about 1/3 of current trading levels.
In March of 2023, Vistra made its latest smart capital allocation move, acquiring the nuclear generation assets of Energy Harbor (yet another bankrupt IPP), which served as the catalyst for Third Point acquiring shares. The timing of this deal was prescient, as nuclear is finally being recognized for its merits as the only carbon-free source of 24/7 power generation. Unfortunately, onerous regulatory requirements make it very difficult to build nuclear plants economically outside of China. Vistra seems to have channeled Mark Twain and thought: “buy nuclear generation, they’re not making it anymore”.
After our investment, the market started to focus on two trends that we think will have profoundly positive impacts on Vistra’s future value: increased penetration of intermittent generation (renewables), and an inflection in power demand from AI/data centers and electric vehicles.
While increased renewable generation is clearly a positive for reducing CO2 emissions, many supporters of green energy are naïve to the impacts of solar and wind on the reliability of grids (lower) and volatility of electricity prices (higher). Price signals have become so distorted that rather than building new dispatchable (gas) power generation, we are shuttering capacity. However, when the wind stops blowing or the sun stops shining, we are increasingly reliant on huge amounts of dispatchable generation to turn on quickly, even though it may have been losing money an hour prior. Counterintuitively, electricity generated by gas in Texas has grown 30% since 2016 despite a twenty-fold increase in solar and a three-fold increase in wind. Texas recently created a $10 billion fund to incentivize new gas generation to strengthen the grid. We believe the only way to incentivize new capacity is to improve the long-term profit outlook for dispatchable gas assets via market reforms or other legislation and expect this to benefit those like Vistra who already own existing dispatchable generation.
We also believe US electricity demand is poised for significant growth for the first time in decades, and Vistra’s fleet of baseload power (both nuclear and natural gas) is uniquely positioned to benefit. As the AI arms race commences, McKinsey estimates new data center build could drive an incremental 800 TWh of global electricity demand by 2030, with 40% of this driven by Generative AI1. The US is expected to capture roughly half of this demand, as bargain-basement power prices make it the world’s leading destination for new data centers despite higher labor and real estate costs. We expect data centers to drive domestic demand growth to accelerate 1-2% per annum over the next five years against a baseload supply backdrop that continues to decline as coal is phased out. Rising electric vehicle penetration is subsequently expected to add an additional 1% to annual growth (Pacific Gas & Electric, who serves a region with over 20% EV penetration, highlights that every two new EVs is equivalent to a new household in terms of demand)
Because many data centers require 24/7 power (which cannot easily be provided by renewables) and connection timelines for utility grids exceed three years in some cases, hyperscalers have shown growing interest in contracting directly with nuclear plants behind the meter to ensure speedy, consistent access to electricity. For example, Amazon recently signed a 20-year agreement with nuclear operator Talen to buy power at a ~60% premium to market prices. Despite this premium, nuclear remains by far the cheapest direct source of clean energy for a hyperscaler; the estimated levelized cost of electricity for a renewables system fit to provide 24/7 power is an eyewatering $200 per MWh, roughly triple what Amazon will pay Talen2.
Vistra is in the pole position to capitalize on these trends, and we expect the discount applied to their assets to continue to narrow as their business becomes increasingly essential to serving domestic power demand.
Alphabet
During Q1, the funds made a substantial investment in Alphabet as the market worried about the impact of LLMs, personal assistants, and answer engines such as Perplexity AI on Google Search. We have owned Alphabet in the past and have long admired its exceptional business model and its proven ability to maintain a leading position across an array of preeminent products such as Search, Gmail, Android, GCP, and YouTube.
The concern that in an AI world, changes in the way consumers will eventually interact with their personal devices and with the internet can result in risks to Alphabet’s core business Search is not entirely unfounded. Alphabet, however, has both a substantial distribution and technology advantage over competitors and is positioned to use its AI capabilities to unify, enhance, and better monetize the entire suite of its products.
In early March, Gemini’s initial blunders further contributed to the narrative that Alphabet will end up an AI loser. Assigning primacy to a small operational misstep while demoting the fact that the company has been building world-class capabilities in AI for over a decade, created an attractive entry point for a long-term investor. While it is easy to forget that the original paper on Transformers which paved the way for the rise of LLMs was published by Google engineers, and that it was Alphabet that built two of the leading AI research organizations (Google Brain and DeepMind), when news reports previewed Apple’s intentions to embed Gemini into iOS, investors paid attention. We believe the moment when Gemini takes a seat at the economic table is approaching. Page 7
As to the question of disruption of Google Search, it is important to imagine that generative AI is likely to take content creation costs to zero, not unlike when the internet took distribution costs to zero. The post-AI internet will have lower barriers to entry in a lot of content- and information-dependent industries while at the same time increasing customer acquisition costs. Google stands to benefit from this megatrend. We envision the eventuality of a world of greatly reduced content creation costs and a proliferation of information, a lot of it in the form of ad-supported content full of AI fakes. Against that backdrop, Google Search will be increasingly important as a source of truth.
Despite recent efforts to contain expense growth, Alphabet full-time employees have grown at a 14% CAGR since 2017 and the median total compensation per employee, at $280,000, is the second highest among major technology companies. This excessive increase in the company’s employee base, combined with a slowdown in revenue growth, has resulted in Alphabet achieving significantly lower revenue per employee than companies like Netflix, Apple, Meta and Nvidia. However, we are encouraged by improvements evident in the most recent quarter as well as management’s decisive handling of employee unrest, in sharp contrast to its prior capitulation to a mob of angry employees in its cancellation of Project Maven in 2018.
Sincerely, Daniel S. Loeb CEO
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.