Grid Alpha: Daily Power Market Briefing (ERCOT, PJM, CAISO)

Grid Alpha
  • 5.0 (2)
  • Investing
  • Daily

Grid Alpha turns real-time data from all nine North American power markets (ERCOT, PJM, CAISO, ISO-NE, NYISO, MISO, SPP, AESO, IESO) into short, trader-ready signal. Each daily episode is a two-host conversation: Ever asks the questions a trader would ask, and our lead analyst reads the tape: fuel mix, LMP/DART spreads, congestion, storage response, and LinkedIn commentary from analysts, developers, and policy watchers who actually move size. No background music, no fluff, just the setups that matter this week for U.S. power and gas traders. Live dashboards at gridalpha.us.

  1. 6h ago

    A $1607 PJM Spread That Is All Congestion and No Confirmed Driver

    PJM's real-time nodal spread hit $1607.18/MWh in the 11:50 UTC interval, 06:50 ET on a Sunday morning, with the RTO hub sitting at $244.29. Yorkana 13 kV Load topped the stack at $923.28; Safe Harbor 13 kV Units 8 through 12 anchored the bottom, all five printing -$683.90. Nothing in the constraint feed or the news tape explains it yet. The decomposition leaves no ambiguity about what kind of spread this is. Every node at both ends carries the same $241.48 system energy component; losses are $3.28 at Yorkana and $1.70 at Safe Harbor. The entire gap is congestion: $678.52 pulling Yorkana up, $927.08 pushing Safe Harbor down. Yorkana is not alone on the high side. Redfront 115 kV, Prospeme 115 kV, Whellam 115 kV and Yoe 115 kV printed $914.93 to $922.59 with $668.90 to $677.81 of congestion each, which reads as a coherent load pocket on one side of a binding element, with the Safe Harbor generator bus on the other side being told its output is worth less than nothing. Five identical unit-level prices at -$683.90 say the constraint sits upstream of the plant, not in a per-unit dispatch quirk. What should be visible and is not: the snapshot's three listed constraints, Manor 230 kV, ALLLUD4J 138 kV and Mardela 69 kV, all show null shadow prices, and none is an obvious south-central Pennsylvania path. The 48-hour RSS lookback returned two items, both on data center policy, neither on a PJM transmission event. The outage feed logs 144 PJM events over seven days and shows planned outages near 25.3 to 25.6 GW, forced 3.1 to 4.4 GW and maintenance 1.9 to 2.5 GW, but all five sampled rows carry the same 2026-11-09 05:00 timestamp, so treat that as a revision series rather than a fresh read on today's corridor. System conditions do nothing to explain the pocket either: load 87,736 MW, net load 85,748 MW, gas on the margin at 07:00 ET. This is a local topology event sitting on top of an ordinary Sunday, and whether the Safe Harbor units are even running into that price is unknown; the feed carries no unit output or status. The angle frames this as a basis-widening watch, and the test is whether the pocket survives the morning ramp or dissolves as a single-interval artifact. If the Yorkana 13 kV congestion component is still at or above $678.52/MWh in the 14:00 UTC (09:00 ET) interval, the constraint is persistent and the day-ahead congestion at these five nodes becomes the thing to size against; if it has collapsed to within a few dollars of the $241.48 energy component by then, the 11:50 print was transient and the setup is dead. If a shadow price appears on any of the three named constraints before 13:00 UTC while Safe Harbor still shows -$927.08 of congestion, the driver is identified and the trade stops being a guess about topology. If PJM-RTO moves off $244.29 but the Safe Harbor component does not budge, the move is energy, not basis, and the pocket is still intact. > A spread of $1607 built entirely of congestion on a Sunday morning, with no named constraint behind it, is a corridor telling you something the outage feed has not caught up to. Not investment advice. For informational purposes only. Yesterday's tape: SPP SPPNORTHHUB lmppeak crosses 26.72 — verified (observed 119.43). Grid Alpha — daily briefs · live dashboards, all nine markets · public call record

  2. 1d ago

    SPP Nodal Spread Hits $438.59 as SPS Blackhawk Sinks to -$218.19

    SPP's real-time nodal spread reached $438.59/MWh in the 11:50 UTC interval on Saturday, 05:50 CST. OPPD.PLVIEW cleared at $220.40/MWh while SPS.BLACKHAWK1 and SPS.BLACKHAWK2 printed -$218.19/MWh. SPPNORTH_HUB sat at $26.72/MWh at 12:00 UTC, so the entire gap lives in the congestion component rather than the energy price. The decomposition tells the story. Every node shared a $23.08 energy component; OPPD.PLVIEW carried $196.25 of congestion and the Blackhawk pair carried -$238.73. Losses were a rounding error at $1.07 for PLVIEW and -$2.54 for Blackhawk. The angle frames this as a north-south split, and the tape is messier than that. SPS.RICHRDSN1 and SPS.ENGCARBON1, both inside the SPS footprint, were among the five highest nodes at $202.90 and $202.83/MWh on $182.12 of congestion each, so SPS occupied both tails of the distribution at once. In Nebraska, OPPD.TRTL1 and OPPD.TRTL2 cleared at $131.85/MWh on $107.81 of congestion, while three LES nodes, LESTBGSCC1, LESLFG1 and LESTBGSCT4, printed -$139.27/MWh on -$163.17 of congestion. Positive and negative pockets sit side by side in both Nebraska and the Southwest, which reads like a constraint or set of constraints pricing specific injection points against specific sinks, not a single regional seam. The system backdrop is consistent with that read. SPP load was 31,996.761 MW at 11:55 UTC with net load of only 11,814 MW and wind as the main supply source; a low net load, high wind morning is the textbook condition for generator-side nodes to go deeply negative behind a binding element. The snapshot lists TP104733629, TP105933633 and TEMP09_26226 at the top of the congestion feed, but each shows a shadow price of n/a, so the specific element producing the Blackhawk and PLVIEW components is unconfirmed. The seven-day outage log counts 1,172 generation outage events, with sample rows from 2026-11-07 showing coal outages between 5,927 and 7,538.6 MW, but that count may be pagination-truncated and nothing in the data ties it to this path. This is one five-minute interval, with no prior or subsequent prints in the pack, so persistence is the whole question. If the SPS.BLACKHAWK1 congestion component still prints at -$238.73/MWh or lower at 12:00 CST, the Southwest pocket has survived the midday ramp and the constraint is structural for the day; if the component has narrowed materially by then, the morning print was a wind-ramp artifact. If the OPPD.PLVIEW congestion component falls below $196.25/MWh by 09:00 CST while the SPS.BLACKHAWK1 component stays below -$238.73/MWh, the two ends are decoupling and a single-constraint explanation is wrong. Check SPPNORTH_HUB against $26.72/MWh at 12:00 CST as the control: if the hub moves while nodal congestion components hold, the story has shifted from transmission to energy. > A $438.59 spread on a $26.72 hub is a congestion story with no named constraint, and until a shadow price shows up, the only trade is watching whether Blackhawk's -$238.73 survives noon. Not investment advice. For informational purposes only. Grid Alpha — daily briefs · live dashboards, all nine markets · public call record

  3. 2d ago

    $331.94 PJM Nodal Gap: ALLDAM6 Congestion, Brunner Island Negative

    PJM's real-time nodal spread reached $331.94/MWh in the 11:50 UTC interval, 06:50 ET on a Friday morning. ALLDAM6 138 KV GEN 6 and GEN 5 printed $324.12/MWh at the top of the stack while BRUNNERI18 KV UNIT01 and UNIT02 cleared at -$7.82/MWh at the bottom. The PJM-RTO hub was $42.08/MWh ten minutes later, so this is a wires story, not a system one. The decomposition makes that explicit. Every node in the top five and bottom five carried the identical $43.25/MWh energy component, which means the whole gap sits in congestion and losses. ALLDAM6 carried $280.50/MWh of congestion against $0.37 in losses; the KITTANNI138 KV T62, T61 and T2 buses sat one step down at $291.88/MWh with $247.47 of congestion and $1.16 in losses. On the other side, the Brunner Island 18 kV unit buses and the BRUNNERI230 KV AUX10, AUX30 and T20 points all showed -$51.95/MWh of congestion and $0.88 in losses. A cluster of 138 kV generator buses priced far above energy and a cluster of 230 kV plant buses priced below zero is the signature of a constraint or set of constraints separating the two pockets, with the Brunner Island side effectively bottled. What the data does not do is name the constraint. GridAlpha's snapshot lists BRUNNERI230KVBRU-WHE1LN, MANOR230KVMAN-OTTC1LN and BUTLAPS138KVBUT-KAR1LN as the top congestion facilities at 12:05 UTC, but all three show shadow prices as not available, so it is unconfirmed whether any of them was actually binding at 11:50 UTC. The Brunner Island 230 kV line appearing on that list alongside negative congestion at the plant's own buses is consistent with a local export limit, but that is inference, not a sourced driver. System conditions give no help either: load was 96,127.77 MW at 12:05 UTC, net load 90,094 MW, gas the main source. GridAlpha logged 144 PJM daily outage events over the trailing seven days, but the sampled rows carry no facility detail and the aggregation level is unstated, so nothing in the outage feed can yet be tied to these nodes. That leaves the nodal basis itself as the thing to test, and the test is persistence. If the ALLDAM6 congestion component still exceeds $280.50/MWh in the 09:00 ET (14:00 UTC) real-time interval while the RTO hub holds below $42.08/MWh, the 06:50 ET print was not a single-interval artifact and the pocket is carrying a sustained constraint into the morning ramp. If BRUNNERI18 congestion remains below zero at 09:00 ET and PJM posts a shadow price above zero on BRUNNERI230KVBRU-WHE1_LN, the export-limit read gains confirmation; if the shadow price stays absent and the Brunner Island buses rise back above the $43.25 energy component by that interval, the discount was transient and the setup is gone. The third check is ordering: if KITTANNI138 congestion is still within the $247.47 to $280.50 band set against ALLDAM6 at 09:00 ET, both pockets are being priced off the same facility rather than two separate ones. > A $331.94 spread with a flat $43.25 energy component across every node is pure congestion looking for a named constraint; until a shadow price appears, nodal basis at ALLDAM6 and Brunner Island is the only tradable fact. Not investment advice. For informational purposes only. Grid Alpha — daily briefs · live dashboards, all nine markets · public call record

  4. 3d ago

    SPP Prints $742 Spread: OKGE $278 vs KCPL $-464 at 05:50 CST

    SPP's real-time nodal spread reached $742.01/MWh in the 11:50 UTC interval, 05:50 CST on a Thursday morning. OKGE.VOLT.0125 cleared at $277.93/MWh while KCPL.VOLT.0271 printed $-464.08/MWh. The system energy component at both ends was an identical $42.00/MWh; none of the gap is energy. Congestion did all the work. OKGE.VOLT.0125 carried a $233.34/MWh congestion component and KCPL.VOLT.0271 a $-504.20/MWh one, with losses adding $2.59 and $-1.88 respectively. The high side is a cluster rather than one outlier: OKGE.VOLT.0035 at $274.62, OKGE.RIVERVALLEY at $270.29, OKGE.NUEN.LAM1 at $267.62 and AECCHYDRO13 at $266.80 fill out the top five, each with congestion between $222.39 and $233.34. The low side runs deeper and wider. AECCFULTON sat at $-379.60 on $-423.42 of congestion, and three WAUE.MRES.EXR units printed an identical $-155.86 on $-196.30 of congestion apiece. SPPNORTHHUB traded $40.56/MWh at 12:00 UTC, roughly the $42.00 energy marginal, so the hub saw none of this. SPP listed FULPATLONSAR, TP105933633 and TP1352_33578 as its top constraints at 12:01 UTC but published no shadow prices, so which flowgate pins KCPL.VOLT.0271 at $-504 cannot be read off the feed. The "Kansas City pocket" framing is a reasonable inference from the node name; the data do not state the node's physical location or the constraint it sits behind. The backdrop supports a trapped-generation reading without proving it. System load was 29,974.718 MW against net load of 20,764 MW at 11:50 UTC, with natural gas the lead fuel, implying roughly 9,200 MW of renewable output netted out. A $-504 congestion component at a generation-side node under that renewable share is what curtailment pricing looks like, and SPP logged at least 1,943 VER curtailment events over the last seven days. The caveats matter: the only returned curtailment samples are solar readings from 2026-11-05 that all show zero MW, the 1,172 outage-capacity events came back as five readings from 2026-11-05 13:00 showing 8,019.7 to 8,969.7 MW of coal offline, and wind output in the KCPL and WAUE areas is not in the feed at all. Day-ahead LMPs for either node are also absent, so whether this congestion was priced in overnight is an open question. Only the 11:50 UTC interval is shown, so persistence is the first thing to check. If KCPL.VOLT.0271's congestion component still prints below $-200/MWh at 08:00 CST, treat the spread as structural for the session rather than a single five-minute artifact, and the OKGE-to-KCPL gap becomes the day's intra-SPP trade. If OKGE.VOLT.0125 congestion drops under $100/MWh by 09:00 CST while KCPL.VOLT.0271 stays negative, the binding element is on the export side of the low node and curtailment, not import scarcity, is the mechanism. Watch whether SPP posts a shadow price on FULPATLONSAR, TP105933633 or TP135233578 by 12:00 CST; any value above $200/MWh names the wire and lets the DART comparison actually be made. > A $742 spread on a $42 energy price is a transmission story, and until SPP posts a shadow price nobody can say which wire wrote it. Not investment advice. For informational purposes only. Grid Alpha — daily briefs · live dashboards, all nine markets · public call record

  5. 4d ago

    CAISO Bishop Node Prints Negative $1,048, Spreading $1,109 Against Northern Sink

    A maximum real-time LMP spread of $1,108.68/MWh defined the CAISO interval on November 11, driven by a stark divergence between two distinct geographic clusters. PODBISHOP1UNITS-APND printed an LMP of -$1,048.26/MWh, while WDT2737_N008 settled at $60.42/MWh. The gap crystallizes a severe local oversupply condition in the Bishop area against a stable, positively priced northern node. The mechanism is entirely congestion-driven. At PODBISHOP1UNITS-APND, the energy component remained a modest $28.52/MWh, but the congestion component plummeted to -$1,090.62/MWh. This negative congestion value indicates that export capacity from the Bishop area is binding, forcing generators to pay to inject power. Conversely, WDT2737N008 displayed a positive congestion component of +$19.50/MWh on an energy price of $25.58/MWh, resulting in the $60.42/MWh LMP. The pattern is not isolated to a single node; PODCONTRL1POOLE-APND and PODCONTRL1QF-APND matched the Bishop print exactly at -$1,048.26/MWh, while PODBISHOP1ALAMO-APND tracked closely at -$1,047.15/MWh. On the high side, five WDT273_7 nodes (N004 through N008) all printed identical $60.42/MWh values, suggesting a uniform constraint or pricing block in that northern zone. Traders should monitor the persistence of the negative congestion component at the Bishop cluster versus the stability of the WDT273 block. If the congestion component at PODBISHOP1UNITS-APND remains below -$1,000/MWh in the next real-time interval, the export constraint is structural rather than transient. Watch the spread between the WDT2737_N008 node and the Bishop cluster; if the spread narrows below $500/MWh, it signals relief in the export constraint or a shift in local load. The identical pricing across multiple WDT273 nodes suggests a single binding constraint, so relief would likely manifest simultaneously across all five. > The $1,109 spread is a pure congestion arbitrage: pay to inject in Bishop, collect at the northern sink, and wait for the export constraint to break. Not investment advice. For informational purposes only. Yesterday's tape: MISO ILLINOIS.HUB lmp_peak Grid Alpha — daily briefs · live dashboards, all nine markets · public call record

  6. 5d ago

    MISO's $1,606 Nodal Split Was All Congestion, No Shortage

    MISO's maximum real-time nodal LMP spread printed $1,606.45/MWh in the five-minute interval ending 11:50 UTC Tuesday, 05:50 CST: EMBA.WDFSOL at $1,020.45/MWh against EAI.WMULD_1.AZ at -$586.00/MWh. Ten minutes later ILLINOIS.HUB cleared $71.11/MWh. The system was not short; two pockets of it simply could not reach each other. The decomposition is unusually clean. Every one of the ten extreme nodes carried the identical $258.51/MWh energy component, so the entire gap is congestion and loss. WDFSOL's congestion term was +$756.10/MWh and WMULD1.AZ's was -$851.33/MWh, a range above $1,600/MWh on that component alone, with losses contributing single digits at both ends. Neither tail is a one-node artifact: EAI.WSTMMP.ARR printed $977.65/MWh, TVA.BRMNH1.ARR $952.86/MWh, and EAI.GACP.GMPWR and EAI.DRVRSOL both $825.99/MWh on the high side, while EAI.WMULDLD printed -$585.88/MWh, EAI.WMPSOL -$521.34/MWh, and EAI.BIGCYP120 and BIGCYP20 both -$408.84/MWh on the low side. System load was 73,599 MW with net load 67,827 MW at 11:55 UTC, natural gas the main marginal source. MISO's top congestion list at that snapshot named WestMemphisAT1161500kVfloSansSouciDriver alongside EA50024CLDWTRCLDWTRSENATO3A and ITC34066MONROE4MONRONSTAR121_1, but all three carried shadow prices reported as not available, so neither the marginal value of the bind nor which element is splitting the two Arkansas-area pockets is established by the data. The geography of the nodes and the geography of one named element are suggestive and nothing more. For a desk carrying nodal exposure in MISO South, the checkable questions are narrow. If the congestion component at EAI.WMULD_1.AZ is still below -$100/MWh at 09:00 CST, the bind outlived the morning ramp and intraday basis risk is a position problem rather than a five-minute print. If ILLINOIS.HUB holds under $100/MWh over the same window while the EAI cluster stays below -$100/MWh, the event remains internal basis and the hub is the wrong instrument to hedge it. And if shadow prices on those three elements populate at any value above $0/MWh in the next publication cycle, the bind acquires a marginal cost that can be compared against the $756.10 and -$851.33 congestion terms; until then the constraint attribution stays open. The most recent MISO generation outage estimate in hand is dated 2026-11-04 05:00, at 17,394 MW planned, 3,977 MW forced, and 4,823 to 5,038 MW derated, which is context for the week and not a read on Tuesday's supply stack. Background matters here too. The anomaly feed logs at least 15,829 real-time binding-constraint events across MISO in the trailing seven days, though that count is flagged as a truncated sample. And the transmission conversation around the RTO is live on two fronts: Invenergy filed a complaint at FERC on November 9 contesting MISO's position that merchant HVDC resources are not capacity resources, and EPRI's CEO argued the same day that transmission utilization should be judged by how the system actually operates after something fails, pointing to a specific constraint and a specific action. > A $1,606 spread with a $71 hub is not a shortage signal; it is a transmission bill, and right now MISO has not told anyone what it costs. Not investment advice. For informational purposes only. Yesterday's tape: ERCOT ESCONDIDESR negativelmp ESRN lmp_peak >= 58.96 — verified (observed 760.2). Grid Alpha — daily briefs · live dashboards, all nine markets · public call record

  7. 6d ago

    $1440 Node Split Puts ERCOT Battery Revenue on Geography

    At 05:50 CST Monday, ERCOT's real-time market priced the same five-minute interval at $1110.60/MWh and -$329.10/MWh simultaneously: a $1439.70/MWh gap between TSECAESRN and ESCONDIDESR. HBNORTH settled at $58.96/MWh ten minutes later. Nothing about systemwide conditions justified either tail. Load was 45,765 MW with net load at 40,753 MW, natural gas carrying the margin: a mild Monday morning, nowhere near scarcity. TSECAESRN did not clear high alone. That cliff between fourth and fifth place is the signature of a binding constraint isolating a small pocket rather than a tight system. ERCOT's top constraint list for the 11:55 UTC snapshot named ESCONDGANSO11 alongside IDUPPIDUPS11 and NLARSWPILONC11, though no shadow price was published for any of the three. The underlying density is not in dispute: 21,875 shadow price events in SCED over the trailing seven days, against 2,271 unplanned outage events in the same window. The component split of that $1440 is unconfirmed, and so is the specific constraint that drove it; both matter for anyone sizing a hedge, because a congestion-dominated spread is CRR-expressible and a loss-dominated one is not. What I would watch first is persistence at the node level: whether ESCONDIDESR prints negative LMP again during the forecast elevated-price windows on 2026-11-11 and 2026-11-12, HE18 through HE21, while TSECAESRN, HAMIBESSRN and AMISTADALL clear above HBNORTH in the same hours. If that asymmetry repeats across both evenings, the pocket is structural and the next CRR auction between these nodes should reprice. If ESCONDIDESR instead settles within $20/MWh of HB_NORTH by HE21 on 11-11, Monday morning was a one-interval dislocation and not a regime. Two supply details are worth tracking against those hours. PLATINUM, a 152 MW battery, has been on Maintenance Level 1 with 0 MW available since 23:00 CST on 2026-11-04, a full 152 MW reduction. PINN_SLR has logged forced solar outages cutting 76 MW, 79 MW and 168 MW off a 218 MW maximum. On the fuel side, Enbridge's 48-hour notices include an MNUS customer-specific Operational Flow Order effective 11/7 and planned maintenance at the Garden Banks South Marsh Island 76 platform. Any of those tightens the local picture into the 11-11 evening block, and none of them shows up in a hub print. > When one battery node clears at $1110.60 and another at -$329.10 in the same interval on a 45,765 MW morning, the trade is no longer in the hub: it is in the interconnection paperwork. Not investment advice. For informational purposes only. Yesterday's tape: PJM WESTERN HUB lmp_peak >= 46.86 — verified (observed 381.51). Grid Alpha — daily briefs · live dashboards, all nine markets · public call record

  8. Nov 8

    Western Hub Prints $411.62 Overnight, $23.85 by 07:00 EST

    PJM Western Hub cleared $411.62/MWh in a single five-minute interval at 01:50 EST Sunday, then printed $23.85/MWh at the 07:00 EST read. The trailing 23.8-hour mean was $46.86/MWh, which puts that one bar near nine times the window average. Three hours before the peak the hub had set its window low at $11.79/MWh. The window runs 286 five-minute bars from 07:10 EST Saturday to 07:00 EST Sunday, so the peak is resolvable to one interval. What the series does not resolve is persistence: only the peak bar is reported, not how many consecutive intervals cleared above the baseline, and no energy, congestion and loss decomposition of the hub LMP is in hand. The 07:08 EST binding-constraint list is entirely low-voltage: TODDSTVN115KV198C1LN, SPOTSLV115KVGOL-SPOT1LN and LOCKWOOD138KVLOC-RIC11_LN, all 115 kV and 138 kV facilities rather than bulk transfer paths, and every one of them reports its shadow price as unavailable. The system backdrop at 07:05 EST was soft: 82,456 MW load, 74,462 MW net load, nuclear reported as the main source, with PJM-RTO at $21.09/MWh against the hub's $23.85 for a $2.76 positive basis. The most recent outage reads, dated 2026-11-02, show planned maintenance dominating at 23,575 to 23,940 MW against forced outages of 2,767 to 4,681 MW; the feed logged at least 144 daily outage events over the trailing seven days, and nothing newer is available. No PJM item appears in the 48-hour public coverage, so there is no outside confirmation of a grid event at 01:50. Three checks settle whether the overnight print carried information. First, level: if Western Hub real-time is still above the $46.86/MWh window mean at 17:30 EST, the overnight bar was the leading edge of something rather than an artifact. Second, basis: if Western Hub minus PJM-RTO widens from the current $2.76/MWh to $10/MWh or more by 18:00 EST, the congestion component is doing the work, which is the localized read the 115 kV and 138 kV constraint list implies. Third, attribution: if any of those three facilities reappears in the binding list with a published shadow price above $50/MWh by 19:00 EST, the event becomes nameable; absent that, the $411.62 bar stays unexplained rather than explained as scarcity. One framing note: the peak landed at 01:50 EST on a Sunday, inside the overnight hours of the same calendar day as the $23.85 read, so the transient-congestion thesis is being tested against a weekend overnight interval and not a weekday open. > A single five-minute bar at $411.62 against a $46.86 window mean is a congestion artifact until a shadow price names the constraint. Not investment advice. For informational purposes only. Grid Alpha — daily briefs · live dashboards, all nine markets · public call record

See All (136)

Ratings & Reviews

5
out of 5
2 Ratings

About

Grid Alpha turns real-time data from all nine North American power markets (ERCOT, PJM, CAISO, ISO-NE, NYISO, MISO, SPP, AESO, IESO) into short, trader-ready signal. Each daily episode is a two-host conversation: Ever asks the questions a trader would ask, and our lead analyst reads the tape: fuel mix, LMP/DART spreads, congestion, storage response, and LinkedIn commentary from analysts, developers, and policy watchers who actually move size. No background music, no fluff, just the setups that matter this week for U.S. power and gas traders. Live dashboards at gridalpha.us.

Information

You Might Also Like