NAVITAS Energy Whitepaper

Texas Market Overview- the Value of Flexible Load 

1. Forward power prices in Texas have increased significantly over the past 4 months

Because of a combination of factors including the Ukraine conflict spiking almost commodities, abnormally hot early summer weather causing energy price spikes, and adjusted expectation of renewable generation additions forward wholesale power prices in ERCOT have increased 45% or $17/mwh over the last 4 months. 

2.  Demand response value for Flexible Load comprises an increasing larger component of power costs with potential to reduce power prices by $15-30/mwh

Energy curtailment value (explained #2) and RRS value and significantly reduce power costs.  These opportunities to decrease net power costs should increase over the next 5 years given the changes in the ERCOT market.

Note that there is no “execution risk” to this strategy since BTC mining can turn off instantaneously so every 5 min as soon as the RT price is printed.  This reduction in power also does not include the value of selling the ancillaries or ERS in all the other hours (as detailed further in this document).  The optimization opportunity between Flexible Load Curtailment and the electricity should increase over the next 5 years. 

3.  High likelihood of continued supply shortage and resulting energy price spikes

ERCOT relies on a scarcity pricing mechanism, which manifests as energy price spikes when operating reserves are low, to signal the market’s need for capacity. Over the past three years, reserve margins have been around 8-12%, by far the lowest of any major market in North America, contributing to high price spike frequency. Additionally, changes in grid capacity away from thermal and towards intermittent renewables has increased the volatility of supply. As a result of these factors, over the past several years, volatility in energy prices has grown exponentially.

In 2019, day-ahead prices at North Hub exceeded $1,000/MWh for a total of 35 hours, and real-time market prices hit the $9,000/MWh price cap 1 for around four hours in August. In February 2021, a winter storm combined with capacity-constrained system conditions led to an upsurge in energy prices lasting an entire week. Tight operating margins and further price spikes have continued to occur throughout 2021-2022. The high susceptibility to price spikes due to scarcity events is expected to continue going forward.

Flexible Load can capture scarcity pricing either through curtailing load and selling energy or through ancillary service (AS) provision: AS prices correlate strongly with energy prices. AS prices increase not only when energy prices are high but also when energy prices are depressed due to low load and high wind conditions. In general, Flexible Load capture greater value through AS in the near-term when Responsive Reserve Service (RRS) prices are projected to be highest. Second, simply being available to bid in real-time, whether dispatched or not, makes a unit eligible to receive various real-time energy price adders designed to reflect scarcity pricing.

The chart below shows the BTC equivalent power price per hour for the latest gen ASIC miner at the time since 2016. A miner would curtail mining operations, if it had purchased fixed price power and sell power back to the market during hours it would be more profitable than mining bitcoin or simply stop mining if you were buying index power.   The “BTC strike” ranged from below ~$100/mwh in 2016-2017 to $600/mwh in 2021 and is currently ~$150/mwh.  BTC operations should have curtailed on average 100-300 hours per year to capture on average $5/mwh every hour assuming no Uri, $30/mwh with Uri.  2022 through July has seen the highest amount of hours where curtail BTC mining and selling power would have resulted in higher profit compared to the previous 6 hours with 330 hours on a day ahead resulting in a net reduction of $18/mwh in a Flexible Loads power cost for the year

4. ERCOT official forecast significantly overestimates future reserve margin

ERCOT’s May 2022 Capacity Demand Reserve (CDR) report significantly over-states market reserve margin levels over the next 5 years (same in subsequent reports). Solar is the most significant source of new builds in ERCOT. However, the CDR assumes that the contribution of solar resources to reserve margins is 81% of nameplate capacity, whereas, in reality, the net peak load will shift towards evening hours as more solar comes online, reducing its effectiveness to the Reserve Margin. After factoring this in for 10 GW of solar resource additions, its net reserve margin contribution is approximately 40%. Further, the CDR assumes all generators with a signed Interconnection Agreement (IA) are anticipated to come online on time, which is more optimistic than historical data suggests.

Whereas the most recent CDR shows reserve margins in the 30%+ range over the next several years, the true reserve margin is probably closer to 12% for near term and largely flat thereafter, even with an influx of solar. While all markets go through cycles of being long or short on capacity, ERCOT’s energy-only market makes it more susceptible to wide swings in supply and an increased likelihood of short conditions that can be capitalized on.

5. Significant price premium in RRS ancillary service product, capturable around-the-clock without excessive cycling of Flexible Load operations

ERCOT’s unique RRS ancillary product averaged $29/MWh in 2019, making it among the highest priced ancillary service in any major organized market. This was driven in part by the scarcity price spikes. During extreme years such as 2021, as discussed above, the upside potential was huge: RRS prices averaged at $312/MWh in 2021, while excluding Winter Storm Uri, the average remained at $22/MWh.  Forward RRS can be hedged for $10-25/mwh.

 

On an hourly level, RRS prices can spike up for two main reasons. One is when energy prices are very high, and many units are running to serve load, leading to high opportunity costs and fewer MW available to serve RRS. Another is when energy prices are low or negative, and few thermal units are running to serve load. In this case, thermal units must stay online while running at minimum load to provide RRS even though they are out of the money for energy, leading to high cost of service provision and high RRS prices (i.e., the cost to keep the unit running, unprofitably, when it would otherwise be offline if not for market RRS requirements). Flexible Load can take advantage of high RRS prices in both cases. Unlike thermal units, Flexible Load have no minimum load restrictions and minimal variable cost burdens to serve RRS.

Unlike regulation up/down, RRS is called to dispatch very infrequently. During normal conditions, RRS is typically called only a few times per month, mostly for very short duration deployments (<10 minutes). During emergency conditions, RRS capacity is “released” to the real-time market and dispatched according to its energy price bids. Bids up to $5,000/MWh are allowed. Critically, units are dispatched according to the real- time locational marginal price (LMP), not the settlement energy price (SPP) which is the sum of LMP plus various scarcity price adders. Even during system emergencies, most of the time, LMP is lower than $5,000/MWh and the scarcity price adders bring the SPP up to the $5,000/MWh cap. This means that Flexible Load will not necessarily have to dispatch continuously when the energy price is $5,000/MWh – thereby reducing the risk of serving RRS even during an extended system emergency. Because of this, RRS can be provided continuously by Flexible Load with little cycling and careful management. Flexible Load, particularly crypto mining, can provide RRS through primary frequency response: the most privileged category of RRS, and one not capped out by MW procurement limits.

 6. Robust demand growth across ERCOT supports higher price levels and absorbs new entrants

While other markets have experienced tepid growth or even shrinking demand, demand in ERCOT has grown rapidly, benefiting from Texas’ economic development and population expansion. Over 2015-2021, ERCOT- wide demand growth averaged about 2.7%, and demand growth for the first six months of 2022 has averaged 12% above 2021 levels. The drivers of growth are varied and include:

  • Oil and gas drilling in the Permian: As of June 13, 2022, oil prices at WTI are trading around $121/bbl, a highly profitable range for most Permian producers that points to continued strong electric demand growth from drillers in upcoming years.
  • Large LNG export terminals are to be located along the Gulf Coast, mainly in the ERCOT South load zone. The export terminals, and the broader gas sector they support, have increased ERCOT’s total demand. Feed gas deliveries for LNG exports are increasing along with growing exports to Mexico forecasted at 6.5 bcf/d and 2.2 bcf/d, respectively, by 2045.
  • Higher EV penetration than assumed in ERCOT’s long-term load forecasts (and adopted by ICF for the first 5 years of the forecast) could provide an upside to demand growth. ERCOT’s forecast is based largely on econometric regression and does not assume explicit additional loads from EVs or other electrification.
  • Commercial expansion in the ERCOT region as Silicon Valley tech firms are moving to leverage Texas’s offerings. For example, in 2020 and 2021, several large technology firms such as Hewlett Packard, Oracle, and Tesla announced that they are moving operations and/or headquarters from California to Texas.
  • Crypto mining – ERCOT continues to be a prime location for US-based cryptocurrency mining. Some studies estimate ERCOT could add 5 GW of crypto mining load by 2023, or an estimated 7% of the 2021 peak load served. 

Load growth supports power and ancillary prices by pressuring reserve margins and absorbing new generations.  Higher average power prices will increase overall power costs for Flexible Load.  Conversely, participating Flexible Load can benefit from higher ancillary services price to help offset base power costs.

  • Population growth is observed at the main load centers, especially in the ERCOT North and Houston regions. Based on the latest census bureau projections, the Texas population is growing at 1.24% on an annual basis.

7. Fast-ramp capability of Flexible Load increasingly aligned with ERCOT’s operational needs

ERCOT already has over 30 GW of wind and nearly 10 GW of utility-scale solar online, and ICF projects another 11 GW of wind and 12 GW of solar online by 2025. ICF does not assume further extension of renewable tax credits; any renewal or new incentive structures would likely bring further renewables online particularly over 2025-2030. This increase in renewables will result in greater need for ancillary services and will increase volatility in real-time markets. Additionally, projects with low start costs and fast ramping capabilities will increasingly be dispatched over slower plants to meet net load ramps and forecast error in renewable output. Operational stresses on the grid translate to price spikes capturable by Flexible Load and increased costs for units not able to start and ramp quickly, which could lead to more retirements.


8. ERCOT market reforms in response to Winter Storm Uri in February 2021 must be monitored and considered when managing Flexible Load in the power market

Winter Storm Uri highlighted the vulnerability of the market to extreme weather as well as the capacity shortage in case of contingency conditions. As of Jan 2022, the Public Utility Commission of Texas (PUCT) and ERCOT are undertaking a comprehensive review of the market structure in response to the February 2021 blackout. The changes that the PUCT are focused on include the following:

Phase 1 reforms, largely approved during fall 2021:

  • Updated ORDC curve parameters and systemwide offer cap. The PUCT mandated that the high systemwide offer cap (HCAP) drop from $9,000/MWh to $5,000/MWh, and that the minimum contingency level (MCL, which sets the level of online reserves below which the price is automatically set at the HCAP) be increased from 2,000 MW to 3,000 MW.
  • Updates to demand response and Emergency Response Services (ERS) interruptible load program rules. ERS also provide another program that Flexible Load can participate if Flexible Load does not want its operation controlled automatically by is Qualified Scheduling Entity when participating the RRS markets. 
  • Acceleration of previously approved improvements to ancillary services such as ECRS and FFR.  Flexible Load will be able to participate in these new ancillary services.

Phase 2 reforms currently being considered include one or more of the following:

  • LSE Reliability Obligation, a pseudo-resource adequacy (RA) like construct that would require load serving entities to contract for a minimum amount of capacity
  • Dispatchable Energy Credits, a program requiring a certain amount of new 2-hr battery and/or extremely efficient, quick-start gas capacity
  • Backstop reliability service, in which guaranteed contracts would be given to a subset of plants purely for emergency needs and extracting them from the energy market