Description

SUMMARY:  Hidden Data Center Power Play @3x FCF

Earlier this year, Amazon (“AMZN”) agreed to purchase a data center asset from Talen Energy Corp.  While the transaction was notable in itself because of the attractive return the seller made on the transaction (>2.5x MOIC), the details and the implications for the market are very telling.  Amazon has the right to purchase approximately a gigawatt of annual capacity for 10 years at a fixed estimated price that is more than 50% above the current market (estimated >$80/MW v $55/MW on futures).  To commit to an agreement of this scale underscores Amazon’s significant concern regarding reliability of power in the coming years.  Only a few weeks later, news articles began to circulate on “Stargate”, an initiative from Microsoft to spend $100 billion on data centers encompassing 5GW of capacity.        

U.S. power prices are on the precipice of a major upward shift driven by both supply-side factors (closure, regulation, intermittence, weather) and demand-side factors (EVs, IOT, data-centers, onshoring). For example, Goldman Sachs expects a rapid acceleration in growth for power consumption driven by data centers (15% CAGR from 2023-2030), but capacity does not appear to be taking notice.  In fact, Goldman Sachs estimates 47 GWs of capacity is needed in the US through the end of the decade from gas and renewables. However, the MISO market, which encompasses much of the Midwest, is currently facing a declining supply curve right as demand is set to accelerate. While the market has been chasing the likes of Vistra, Constellation Energy, NextEra, etc., we believe Hallador Energy (“HNRG”) with a 30% free cash flow yield based on contracts signed to date, is the only play in the market to gain exposure to the regional power network that has the most upside risk. 

In fact, HNRG screens as a coal producer but has recently transformed into a vertically integrated IPP.  We think the combination of limited analyst coverage and public filings that emphasize its coal operations has masked the significant value of a MISO power plant. Based on the current forward curve, HNRG is currently trading at a 25% FCF (free-cash-flow) yield, and at a 29% FCF yield on pricing they have recently been realizing in the futures market.  Furthermore, if power pricing spikes from here (as we believe to be the case), these assets can generate more than 50% of its market cap in a single year.

Source: Internal Estimates, Company Disclosure

We also take comfort that management agrees with us as evidenced by the language from its most recent 10-K filing (see quote below from MD&A section).  If the company were to exceed $25/MWh in margins, we think HNRG would generate roughly a 30% cash flow yield.  It is also worth emphasizing that power prices have strengthened materially since the 10-K filing in March then thereby pointing to much higher levels of profitability. 

“Based on the currently available forward power price curves, we believe over time, the margins earned on energy and capacity sales will be more than double our historical margins of approximately eight dollars per ton on coal production. Furthering this belief, in Q3 we reported contracted sales of 3.4 million MWh to be delivered in 2026-2028 at MWh margins that we believe could exceed twenty-five dollars per MWh.”  

MISO OVERVIEW:

Broadly speaking, the power grid and the electrical system as we know it is facing a wide set of complex challenges, including, but not limited to changes in the fleet of power generation (i.e. wind/solar not providing similar reliability/base power as gas/coal), regulations (forced retirement of reliable generation such as coal in lieu of environmental friendly generation such as solar), extreme weather conditions (severe weather conditions seem to be increasing as evident by issues in Texas and 2024 expected to be the largest hurricane season on record) and load growth (driven by not only by a pickup in economic activity/onshoring, but the growth of incremental power demand that is only starting to proliferate across the nation; EVs/charging, Internet-Of-Things/broader connectivity, data-centers/AI, etc..). This has further been exacerbated by delays in new resources coming online – i.e. in the MISO, there are ~25GW (gigawatts) of new capacity that are averaging delays of 650 days to commercial viability ⁽¹⁾.

While the statement above could be a blanket statement that speaks to all regional areas of electricity distribution, the area that could see the most strain is the Midcontinent Independent System Operator (MISO). In fact, the North American Electric Reliability Corporation (“NERC”) in its 2023 Long-Term Reliability Assessment report ⁽²⁾ determined that there are two “high risk” areas in the US as defined as areas where “the supply of electricity for these areas is more likely to be insufficient in the forecast period and that more firm resources are needed.”

Source: 2023 Long-Term Reliability Assessment

While SERC-Central is the other high-risk area, its shortfall is a timing issue. As you can see below from the same report, the anticipated reserve margin for MISO is stepping down every year and will be below the reference margin level required for the area while SERC-C steps down for a year and then steps back up. Ultimately, it means if the planned shutdowns of coal continue to take place, MISO will be short 4.7GW. Note that the reserve margin is the amount of unused availability capacity of a system as a % of total capacity (i.e. think of this as a cushion for extreme weather events) – Anticipated Reserve Margin is what matters here. This is all ultimately driven by the fact that MISO has one of the highest % of coal generation as a % of total electricity being produced (see charts below). Keep in mind that this report is somewhat dated (~6 months)  and based on recent data from primarily sources such as ERCOT (May 24^th Report  the situation has worsened across the nation for all regions.

MISO                                                                                 SERC-C

Source: 2023 Long-Term Reliability Assessment

Another way of seeing this is from BNEF data that we have spread below. This is a simple version of what capacity is being added regardless of what type (i.e. coal vs solar).

Source: BNEF, Bloomberg.

While this underscores the situation in MISO is more dire than other regions, this is only half of the issue. MISO is replacing heavy baseload capacity (i.e. coal) with intermittent (i.e. solar) capacity, so it is not a 1 for 1 trade off. As you can see below, while installed capacity (“maximum amount of energy that resources could theoretically produce if they ran at their highest output levels all the time and never shut down for planned or unplanned reasons”) is growing (the below assumes no delays), the accredited capacity (“how much energy resources are realistically expected to produce during times when they are needed the most by accounting for their performance, which includes limiting factors such as their forced outage rates during adverse weather conditions”) is declining precipitously. The accredited capacity determines the likelihood of insufficient capacity. F2A as can be evident in the report is the middle case using various assumptions for new capacity ramp and decarbonization targets. The takeaway here is that even in a flat demand environment, we think there will be a significant supply/demand imbalance.

Source: Miso Futures Report – Series IA – November 2023

While we addressed some of the mega data center projects above, we cannot underscore enough the significance of the demand acceleration to come.  According to the Daily Journal, Indiana introduced tax incentives to attract data center investors across the state and since introduction of these incentives, there have been 8 data center projects announced including 4 since October ⁽⁴⁾. When you see headlines like the ones below from the likes of Google and Amazon and overlay this information with industry projections it is reasonable to conclude pricing must dramatically increase to incentive new capacity.

While investors have myopically focused on the ERCOT and PJM power producers, it does appear that MISO is actually a larger beneficiary of data center builds than the other regions.  In fact, the rumored Stargate project from Microsoft was referenced to involve areas within the MISO market.   

Source: Grid Strategies, December 2023

While demand has been flat historically, market pundits last November were already starting to call for ~1% growth over the several years. When factoring in all these data centers hitting over the next 3-4 years, we think the market could be grow much more rapidly than expected.

Source: Bloomberg, Internal Estimates, Miso Futures Report – Series IA – November 2023

HNRG OVERVIEW:

Throughout the life of its existence, HNRG has been known as a coal producer in the Midwest. They have a few different mines in the Illinois Basin in Indiana. They would mine and sell the coal to the likes of Duke Energy, Hoosier Energy, and Indianapolis Power & Light. In 2021, according to management, HNRG signed an interconnect agreement with Hoosier Energy, who owned the Merom power plant. This plant was expected to be retired in 2023, but the regulatory body reversed course on it in 2022. Ultimately, Hoosier ended up selling the plant to Merom in 2023 for pennies on the dollar for a few reasons. First, the interconnect agreement meant that HNRG was the only one able to sell power to the grid from that location past mid-2023 (this was again signed at a time when Hoosier was expecting this plant to close). Second, even if there was a workaround, the plant was money losing given the spike in coal pricing. Given HNRG had the interconnect agreement and was vertically integrated (could procure at a cost lower than market rates), they were the only logical buyer and were able to buy the plant for ~$5mm mainly made of WC and reclamation. 

On February 23, 2024, HNRG announced that they would be idling production at their highest cost mines to focus just on producing power at their lowest cost mines. Ultimately, they ended up committing to reduce their coal production from ~6.5mm tons to ~4.5mm tons (the likelihood is for ~3.5mm of production that can be flexed to 4.5mm in the right environment). Of that ~4.5mm in maximum annual tons, about ~2.3mm is committed to Merom, but that number could end up closer to ~3mm if the plant operates at their target rate, ~0.5mm is priced/committed to a third party through 2028 and ~1mm is exposed to the spot market or has commitments in place without a price. Ultimately, this has positioned HNRG to be a vertically independent power producer (IPP) with full exposure to the MISO market.

To provide an overview of the financials and operational leverage, we believe power price of $35/MWH+ is needed for the plant to cover variable costs. As of right now, the company has forward sold ~31% of its target power production at an average price of $55.37 in 2026 and $54.65 in 2027. To provide a sense of the sensitivity of free cash flow, the forward curve of ~$51 implies that 2026 FCF should be ~$72mm or ~25% FCF of today’s market cap. At $61, where power prices are currently standing entering 2028, the FCF yield jumps to 34%. And at $71, or still $10 below what Amazon contracted with Talen for the next decade, the FCF yield jumps into the mid 40%.  Point being, we see a scenario where the stock could be up 4x+ if power prices further inflect upwards.

Source: Bloomberg, BNEF

Finally, while HNRG may not have the allure of a Talen Corp. or a Constellation Energy to sign an exclusive off-take agreement with a data center, we don’t believe it is out of the realm of possibility. In fact, last quarter the CEO noted that “Since January, we have evaluated and continue to evaluate several major power and capacity sales opportunities, including one proposal made to us that, if contracted, will result in a more than a billion dollars' worth of potential forward power sales.” It is now our belief that that reverse inquiry was in fact from a data center.

Conclusion:

We like that our analysis is supported by prices that are lower than what HNRG is currently contracting at for 2026, but there is considerable upside to the extent that the MISO market begins to look like other US markets such as ERCOT.  For example, futures summer around-the-clock pricing for 2026 in ERCOT have appreciated by 30% year-to-date.  While ERCOT is a more liquid market with greater attention given the scope of independent power producers, the same supply/demand dynamics are taking place in MISO. While we acknowledge that MISO is a market dominated by regulated utilities, it is also important to remember that it takes 4-5yrs for new gas generation and ~2yrs for new solar/wind generation to come online, assuming no delays such as turbine deliveries or construction delays (Note: our conversations with the gas turbine manufacturers suggest a four year waitlist for deliver if one were to place an order today) So while the upside in power prices is probably capped relative to other markets, it also would not be surprising to see power pricing of $60-$80 in the 2026-2028 time period as supply simply cannot come online fast enough. Therefore, we believe it is only a matter of time before the MISO futures curve moves higher and HNRG reprices as an AI power producer.   

1. Source: “MISO’S RESPONSE TO THE RELIABILITY IMPERATIVE – February 2024”
2. Source: “2023 Long-Term Reliability Assessment – December 2023”
3. Source: “Miso Futures Report – Series IA – November 2023”
4. Source: https://dailyjournal.net/2024/04/26/googles-investment-in-indiana-data-center-project-grows-to-2b/

Disclosure: At the time of publication, the author of this article holds a position in HNRG.  This article expresses the opinions of the author. The author has no business relationship with any company whose stock is mentioned in this article.

The author of this article has a long position in the company covered herein and stands to realize gains in the event that the price of the stock increases. Following publication, the author may transact in the securities of the company, and may be long, short or neutral at any time.  The author of this report has obtained all information contained herein from sources believed to be accurate and reliable.  The author of this report makes no representation, express or implied, as to the accuracy, timeliness or completeness of any such information or with regard to the results to be obtained from its use.  Any projections, forecasts and estimates contained in this report are necessarily speculative in nature and are based upon certain assumptions. Accordingly, any projections are only estimates and actual results will differ and may vary substantially from the projections presented. All expressions of opinion are subject to change without notice, and the author does not undertake to update or supplement this article or any of the information contained herein.  This is not an offer to sell or a solicitation of an offer to buy any security.

Catalyst

Market recognition of enormous earnings power from current contracted market pricing

Substantial exposure to MISO market electricity price move