Stock Analysis · Bulls vs Bears

ERock, Inc. EROC

🐂 Bulls vs Bears 🐻Report #312 Oct 2026
★ Willem owns this stock · 3% of the portfolio
58
elevated risk▼ −4 vs last
🐂 9vs11 🐻
Bears lead
30%
▲ Chance of the biggest tailwind the forecasts miss
30%
▼ Chance of the biggest headwind the forecasts miss
The verdict

On demand the evidence leans bullish: Three rival engine makers confirm the shortage lasts until 2028. ERock holds about $610M of customer cash. A new 1,543 MW permit application shows a pipeline well beyond the $1.7B backlog. On delivery and Texas the evidence leans bearish: The orders scheduled at 31 March cover only about $296M of the $363–393M of H2 revenue that guidance needs. Three control weaknesses are still…

The Bulls vs Bears podcast is being made
Two AI hosts debate this report, NotebookLM-style. It appears here automatically within a few hours.

Written by Claude (Anthropic’s AI) from public sources · Willem owns EROC · AI can make mistakes — numbers in brackets point to the sources · Not financial advice

🐂 Bull arguments 9

  1. 1Rivals confirm nobody else delivers firm gas power before 2028Strong
  2. 2About $610M of customer cash backs the order bookStrong
  3. 3A 1,543 MW Midland permit application shows pipeline beyond backlogMedium
  4. 4No debt, $627M cash and an undrawn $250M credit lineMedium
  5. 5Engine long blocks contracted through 2028 with HD HyundaiMedium
  6. 6McCloud's cost per kW undercuts recently approved Texas gas plantsMedium
  7. 7The price has reset to about 1.4–1.9x 2027 salesMedium
  8. 8One plant, three revenue lives, with a growing service tailWeak
  9. 9Low emissions and small footprint suit dense sitesWeak

🐻 Bear arguments 11

  1. 1Booked orders don't yet cover the H2 2026 guidanceHigh
  2. 2Texas permit freeze hits pending applications like Caprock RanchHigh
  3. 3Judges found McCloud had no tender and no alternativesMedium
  4. 4One customer is 28–37% of revenue and 62% of receivablesMedium
  5. 5Three material weaknesses, no remediation date, and a late filingMedium
  6. 62027 consensus needs $100–300M of orders not yet bookedMedium
  7. 7Shorts hold ~43% of tradeable float ahead of December unlockMedium
  8. 8Small engines cost more to run than turbines per MWhMedium
  9. 9Much bigger rivals are doubling or tripling capacity for 2027–28Medium
  10. 10TRA, Class M units and 78% sponsor control leak valueMedium
  11. 11Korean long blocks carry an unpriced tariff billLow

Bull arguments show how well they survive the devil’s advocate (strong / medium / weak); bear arguments show the damage they could do after the counter-argument (high / medium / low).

Bull case

🐂 Bull #1Rivals confirm nobody else delivers firm gas power before 2028

  • The point: ERock sells time-to-power. Its competitors' own statements on lead times say that window stays open for at least two more years.
  • Evidence:
    • Caterpillar (2026-08-04 call):
      • Turbines are booked to "the back half of 2028 and into 2029".
      • It is taking orders "into 2029 and 2030".
      • "No one is slowing down at the moment" [21].
    • Generac (Q2 call):
      • Its own lead time is 40–45 weeks; others are at "70, 80 weeks", in "some cases 2 years".
      • Its $1.6B data-centre backlog has nothing in it for 2028 [22].
    • GE Vernova: 116 GW of gas backlog plus slot reservations, with orders priced more than 20% above 4Q25 per kW [26].
    • ERock: most new contracts deliver in 12–18 months (10-Q). Capacity is sold out "through 2027 and into 2028" [1,4].
  • Why it matters: while the shortage lasts, 2027 revenue depends on how many gensets ERock's own plants can turn out, not on winning orders. Scarcity also supports price: management says pricing and margin trends are "positive", with a lag [4].
  • 😈 Devil's advocate:
    • Generac's CEO calls behind-the-meter recip power a "bridge technology" and "not permanent", because on-site energy costs more than grid power [22]. The market ERock serves may therefore shrink once grid connections catch up, around 2028–2030.
    • The better pricing reaches ERock's results only "mainly 2028 and beyond" [4], which is the same window in which rivals' new capacity lands (Bear #9).
  • Verdict: Strong Three independent competitor statements confirm the gap through 2028. Beyond 2028 it is unproven.

🐂 Bull #2About $610M of customer cash backs the order book

  • The point: the order book is backed by cash. Customers have prepaid more than ERock's entire market value of equipment it has yet to deliver.
  • Evidence:
    • At 2026-06-30, contract liabilities were $528.4M, up from $170.0M.
    • Deferred income was $92.3M, up from $35.4M. The $56.9M rise is "associated with a recently executed customer agreement".
    • Net of $10.2M of contract assets, customer-funded liabilities total $610.5M [1].
    • Backlog is ~$1.7B (up ~10x year on year) and GAAP remaining performance obligations (RPO) ~$1.8B [1,3].
    • Named buyers: Anthropic (470 MW, equipment only), Meta via El Paso Electric (366 MW), Microsoft (63 MW) [2,3,4].
    • CFO Blakely: "What we put into contracted backlog is firm" [4].
  • Why it matters:
    • Deposits fund the inventory build, which rose $62.4M in H1 [1].
    • Deposits make cancellation expensive for the customer.
    • The reference names help sell the next order.
  • 😈 Devil's advocate:
    • Customers may cancel before installation is complete. The prospectus warns ERock "may not recover" its costs [2].
    • The size of the termination payments is not disclosed.
    • Anthropic's order is equipment only. Carrington confirmed Anthropic is not the "Customer A" of the IPO materials [4], so Customer A's identity is also unknown.
  • Verdict: Strong Few companies of ERock's size hold this much prepaid customer cash. Cancellation terms remain the blind spot.

🐂 Bull #3A 1,543 MW Midland permit application shows pipeline beyond backlog(new)

  • The point: ERock's next order may already be visible in a public permit file.
  • Evidence:
    • Enchanted Rock LLC applied to TCEQ (the Texas environmental regulator) for an air construction permit for the "Caprock Ranch Energy Facility", south of Midland.
    • The application covers 3,428 engines and up to 1,543 MW. That works out to ~450 kW per engine, ERock's genset size (my arithmetic).
    • The plant would provide "bridge power" to an unnamed "third party", then serve as backup and as a grid "peaking" plant.
    • The public notice is dated ~2026-07-14 and the application is pending [19].
    • The size is more than four times McCloud's 366 MW [11].
  • Why it matters:
    • Interpretation: at McCloud's ~$0.79M/MW, 1,543 MW would be roughly $1.2B of equipment revenue (estimate). That is about 70% of today's entire backlog.
    • It is not disclosed as an order, so it is not in backlog or, in firm form, in consensus.
  • 😈 Devil's advocate:
    • It is a pending application filed by ERock, not a signed order. The customer is unnamed.
    • A large NSR-type air permit sits squarely inside Abbott's 2026-09-21 freeze. TCEQ "has paused the issuance of all air and water permits and authorizations" for data-centre infrastructure [13,15].
    • Interpretation: the same filing that shows the pipeline is also the clearest example of the freeze biting.
  • Verdict: Medium The demand signal is real and very large, but it cannot become revenue until Texas lifts the freeze.

🐂 Bull #4No debt, $627M cash and an undrawn $250M credit line

  • The point: ERock does not need to raise equity to fund its growth.
  • Evidence:
    • Debt was $0 at 2026-06-30: the 2025 term loan was repaid on 06-11 and the convertibles settled on 05-13 [1].
    • Unrestricted cash was $626.6M, plus $34.2M restricted [1].
    • The $250M JPMorgan ABL (an asset-based credit line) is undrawn. It costs SOFR + 2.00%, and its only covenant is a $85M minimum liquidity until fixed-charge coverage exceeds 1.0x [1].
    • H1 operating cash flow was +$268.9M and capex only $8.8M [1].
    • A new 1.2 GW production line costs about $15–20M [5].
    • Management says cash is sufficient for 12 months "and beyond" [1].
  • Why it matters: rental-fleet rivals have to finance their assets. ERock's customers finance ERock.
  • 😈 Devil's advocate:
    • Strip out the customer deposits and cash is only ~$98M (my arithmetic: $626.6M − $528.4M).
    • H1 cash flow excluding the deposit build was about −$98M after capex, or ~$16M a month (estimate from [1]).
    • The 10-Q itself says "additional equity or debt financing may be required" [1].
    • The ABL's borrowing base is not quantified, so its real availability is unknown.
  • Verdict: Medium Strong. The balance sheet is clean, but the cash belongs to deliveries that haven't happened yet.

🐂 Bull #5Engine long blocks contracted through 2028 with HD Hyundai

  • The point: the scarcest input is locked in for the backlog period.
  • Evidence:
    • HD Construction Equipment will supply 22-litre gas-engine long blocks (the bare engine) for 2027–28.
    • The deal is worth 390bn won ($290M), signed 2026-10-07 [10].
    • The Korean press calls it HD's first large continuous-duty contract in North America [10].
  • Why it matters:
    • It removes the biggest supply question for 2027–28.
    • Interpretation: if long blocks are a minority of a ~$357K genset's revenue, $290M of blocks implies output far above the ~$1B/yr run rate. Unit prices are not disclosed, so this is directional only.
  • 😈 Devil's advocate:
    • HD has no North American track record in this duty.
    • ERock filed no 8-K for the deal [7].
    • CARB-DG certification (California's emissions approval) for the new engine was not found.
    • Korea's 15% US tariff rate may apply (Bear #11).
    • The 10-Q showed "no outstanding purchase commitments" at 06-30 [1]. Interpretation: the supply chain behind 2027 was only contracted three months before 2027 begins.
  • Verdict: Medium The volume is secured. Qualification and tariff costs are not yet known.

🐂 Bull #6McCloud's cost per kW undercuts recently approved Texas gas plants(new)

  • The point: on the regulator's own record, an ERock plant is not expensive to build.
  • Evidence: the PUCT Proposal for Decision (PFD), Docket 59076 [11]:
    • McCloud's capital cost is $1,609/kW, on a $499.8M total.
    • Entergy Texas plants approved on 2025-12-09 cost $1,764/kW and $2,122/kW.
    • El Paso Electric's generic combustion-turbine reference is $1,264/kW, but EPE says turbine prices have risen since.
    • SemiAnalysis puts all-in recip costs at $1,700–2,000/kW, against $3,000–4,000/kW for fuel cells [44, snippet only].
    • The units provide N-1 redundancy: 225 MW stays available even with units offline [11].
  • Why it matters: speed doesn't come at a capital-cost premium, so utilities can justify an ERock plant on capex alone.
  • 😈 Devil's advocate:
    • The same record shows ERock plants cost more to run (Bear #8).
    • The judges found El Paso Electric did not show McCloud is cost-effective [11].
  • Verdict: Medium ERock competes on capex, but not on lifetime cost.

🐂 Bull #7The price has reset to about 1.4–1.9x 2027 sales(new)

  • The point: the stock already prices in a lot of disappointment.
  • Evidence:
    • The close on 2026-10-09 was $11.95, against a $21.50 IPO price and a post-IPO closing range of $8.99–18.75 [32].
    • Market cap is ~$2.6B on 219.4M Class A + B shares, or ~$3.3B including the 52.4M Class M-equivalent shares (my arithmetic, [1]).
    • Enterprise value net of $627M cash is ~$2.0–2.6B.
    • 2027 consensus revenue is ~$1.42B (StockAnalysis; WallStreetZen ~$1.4B from 10 analysts) [35,36].
    • Wolfe's 2028 EBITDA estimate is $400M (target 14x) [40].
    • That gives EV/2027 sales of ~1.4–1.9x and EV/2028 EBITDA of ~5–6.5x (my arithmetic).
    • All named brokers rate it Buy. The mean target is ~$22.50, against Guggenheim $20 and JPMorgan $21 after cuts [37,38].
  • Why it matters: if the H2 ramp and 2027 deliveries happen, the multiple has room to rise.
  • 😈 Devil's advocate:
    • The EV understates claims on the company: the TRA's estimated $658M early-termination value [2] and deposits that are owed back as product.
    • The 2027 multiple only holds if consensus holds, and consensus needs orders that are not yet booked (Bear #6).
  • Verdict: Medium The stock is cheap on forecasts that are themselves the open question.

🐂 Bull #8One plant, three revenue lives, with a growing service tail

  • The point: a RockBlock earns first as bridge power, then as backup, then as grid peaking capacity. Each sale also comes with a service contract.
  • Evidence:
    • Services attach to more than 95% of deliveries, with a ~10-year average term and a 100% renewal rate (company claims) [5].
    • Annualised recurring service revenue is $23.6M, up 18% [3].
    • H1 services revenue was $29.2M, +38% year on year [1].
    • The installed base is 1,104 MW [3].
    • Caprock Ranch was filed on exactly this bridge → backup → peaker design [19].
  • Why it matters: recurring revenue grows with the installed base and smooths out the lumpy equipment sales.
  • 😈 Devil's advocate:
    • Q2 services revenue fell to $13.4M from $15.8M [3].
    • The judges challenged McCloud's "second life" for ratepayers [11].
    • An intervenor argued the 50,000-hour engine life means only ~5.7 years at continuous running, against the 20 years assumed [11].
  • Verdict: Weak Medium. The model works on paper, but the regulator questioned its flagship example.

🐂 Bull #9Low emissions and small footprint suit dense sites

  • The point: ERock's units are compact, air-cooled and run on three-way catalysts, which helps where land, water or NOx limits bind.
  • Evidence:
    • McCloud places 366 MW on 31 acres, with no water use. It is registered under TCEQ's Standard Permit with dry three-way catalysts, below Title V thresholds [11].
    • The units hold CARB-DG certification [2].
  • Why it matters: where siting is the bottleneck, a cleaner and smaller plant gets permitted first.
  • 😈 Devil's advocate:
    • Abbott's freeze ignores how clean a project is [13].
    • Caprock's ~7M t/yr of greenhouse gas is drawing activist attention [19].
    • An intervenor argued McCloud could exceed its 100 tpy CO limit [11].
  • Verdict: Weak in Texas today, Medium elsewhere.

Dropped since report #2: "Texas rules require backup generation." The final SB 6 rule (16 TAC §25.194, adopted 2026-09-18) explicitly left behind-the-meter generation "outside the scope" [18]. The tailwind did not materialise in the rule.

Bear case

🐻 Bear #1Booked orders don't yet cover the H2 2026 guidance

  • The point: FY26 guidance needs H2 revenue roughly five times H1, and the delivery schedule disclosed in March falls short of that.
  • Evidence:
    • FY26 guidance is revenue of $435–465M and adjusted EBITDA of +$3M to +$9M [3].
    • H1 revenue was $71.6M and H1 adjusted EBITDA −$26.4M. H2 therefore needs $363–393M of revenue and about +$30–36M of EBITDA [1,3].
    • The RPO schedule at 2026-03-31 showed $335.8M for the rest of 2026 [2]. Q2 recognised $39.9M, which leaves ~$296M for H2 (interpretation). About $67–97M must come from bookings made after March, or from earlier timing.
    • Q2 revenue fell 42% year on year ($39.9M vs $68.5M), and H1 fell 23% [1].
    • Three material weaknesses remain open (Bear #5).
    • The one public Q3 consensus figure ($20.4M revenue, TipRanks) looks stale [37].
  • Why it matters:
    • A miss on its first-ever guidance would hit a newly listed company at the moment its lock-up ends.
    • The Hagens Berman "investigation" page already cites the revenue decline [42].
  • 🛡️ Counter-argument:
    • Revenue is booked at factory acceptance, and bill-and-hold allows booking even when a customer's site is late [1].
    • Inventory has already doubled [1].
    • The Q2 year-on-year drop reflects a one-off prior-year "large delivery of generators out of inventory" [1].
    • Management named $360–390M of 2026 conversion, which matches the guidance [4].
    • Anthropic, the 06-30 RPO rise and the "recently executed" agreement were all booked after March [1].
  • Severity: High It is the next hard test: Q3 needs roughly $140M+ to keep the year plausible (estimate, assuming a 40/60 Q3/Q4 split).

🐻 Bear #2Texas permit freeze hits pending applications like Caprock Ranch

  • The point: more than 80% of ERock's revenue is in Texas, and Texas has stopped issuing data-centre permits with no end date.
  • Evidence:
    • Abbott's 2026-09-21 letter: "pause the issuance of all permits related to data center projects until ERCOT completes its review". It has no MW threshold and no carve-out for self-powered sites [13].
    • TCEQ confirmed it paused "all air and water permits and authorizations" for data-centre infrastructure [15].
    • Fermi America's 5 GW air permit meeting was cancelled [15].
    • Law firms say it is unclear whether standard permits and permits-by-rule are covered [14,16].
    • ERCOT targets 2026-12-10 for its audit filing [17].
    • TCEQ reports to the Governor on 2026-10-19 [13].
    • Abbott wants the January 2027 legislature to "eliminate any financial incentives for data centers" [13]. The sales-tax exemption is at "serious" risk [43].
    • Caprock Ranch (1,543 MW) is pending [19].
    • ERock's 10-Q reported "no material changes" in risk factors and does not mention TCEQ [1].
  • Why it matters: existing deliveries can be booked as bill-and-hold, but new Texas orders need permitted sites. That makes the 2027–28 order flow, not 2026 revenue, the part at risk.
  • 🛡️ Counter-argument:
    • Already-issued permits appear unaffected. McCloud is already registered [11,14].
    • The pause is tied to an audit with a December target.
    • Grid connection is getting slower: Batch Zero (ERCOT's first study batch for big new loads) is now expected to miss its 2027-04-09 deadline [17]. That raises the value of on-site power once permits flow again.
    • Anthropic's Rockdale capacity uses Riot's existing grid connection [45], so not all hyperscaler Texas demand depends on new air permits.
  • Severity: High The pause is real and broad, and the next three dates (10-19, 12-10, January) decide whether it is a hiccup or a regime.

🐻 Bear #3Judges found McCloud had no tender and no alternatives

  • The point: ERock's largest-ever project has a hostile record at the regulator.
  • Evidence: the PFD, issued 2026-09-23 [11]:
    • El Paso Electric ran no tender (RFP), considered no alternative to ERock, and "has never worked with" ERock.
    • Meta's preference was "a contributing factor".
    • The judges recommend approval only if customers are held harmless from capital and operating costs; otherwise denial.
    • Every intervenor urged denial or conditions. Staff made no recommendation. More than 1,000 public comments were filed [20].
    • Exceptions are due 2026-10-13 and replies 10-20. SOAH (the administrative-hearings office) should return its letter by 11-06, and the open-meeting date is not set [12].
    • The schedule calls for construction complete in May 2027 and power in June 2027 [11].
  • Why it matters:
    • McCloud is probably ERock's top customer: ~$290M, inferred from the prospectus's 813-generator/366 MW utility order [2].
    • A denial would put deliveries and the reference case at risk.
  • 🛡️ Counter-argument:
    • Meta already bears costs during the bridge period, so a hold-harmless condition largely describes the existing deal. Approval with conditions is the judges' own first branch [11].
    • The judges rejected the City's harsher cost cap ($565.6M) and its availability guarantee [11].
  • Severity: Medium High. A denial is unlikely, but ERock's contractual protection if it happens is undisclosed.

🐻 Bear #4One customer is 28–37% of revenue and 62% of receivables

  • The point: a handful of buyers decide each quarter.

  • Evidence:

    Period Top customer share of revenue Other large customers
    Q1 2026 37% 13%, 12%
    Q2 2026 28% 13%, 13%, 11%
    FY2025 18% 16%, 14%
    • One customer was 62% of receivables (~$63M) at 06-30 [1,2].
    • Texas is more than 80% of revenue [2].

  • Why it matters: a single delay, dispute or cancellation moves the quarter and the cash position.

  • 🛡️ Counter-argument:

    • Concentration is an artefact of large hyperscale orders, and those buyers are among the most creditworthy anywhere.
    • Deposits cover much of the exposure [1].

  • Severity: Medium–High.

🐻 Bear #5Three material weaknesses, no remediation date, and a late filing

  • The point: the control environment is weak exactly when volumes multiply.
  • Evidence:
    • The auditor found three material weaknesses during the FY2025 audit [1]:
      • segregation of duties;
      • too few accountants with public-company experience;
      • IT general controls.
    • All are unremediated, disclosure controls are "not effective", and no remediation date is given [1].
    • Founder-director McAndrew filed his Form 4 for 06-11 transactions only on 09-04 [8].
    • CEO Carrington previously ran Stem (FY2024 net loss $854M, 1-for-20 reverse split) [46].
  • Why it matters: a 5x revenue ramp through a new plant raises the risk of cut-off errors, such as booking bill-and-hold revenue in the wrong period.
  • 🛡️ Counter-argument:
    • Deloitte is the auditor [2].
    • These weaknesses are common at newly listed companies, and the Section 404 test starts only with the FY2026 10-K.
    • No restatement has occurred.
  • Severity: Medium High.

🐻 Bear #62027 consensus needs $100–300M of orders not yet booked(new)

  • The point: analysts model more 2027 revenue than the disclosed schedule supports.
  • Evidence:
    • 2027 consensus revenue is ~$1.42B; EPS is ~$0.49–0.51 [35,36,38].
    • The March RPO schedule showed $864.5M for 2027 and only $14.8M for 2028 [2].
    • Anthropic's 470 MW (booked in Q2) runs "into 2028" [3].
    • Interpretation: $864.5M plus a share of Anthropic (~$420M at ~$0.9M/MW, estimate, split over 2027–28) plus ~$50–60M of services gives roughly $1.1–1.3B.
    • That leaves a gap of about $100–300M that new orders or Anthropic installation work must fill.
    • FY2028 consensus of ~$2.2B [36] needs much more.
  • Why it matters: if the Texas freeze slows new orders, 2027 estimates come down even if ERock delivers everything it has booked.
  • 🛡️ Counter-argument:
    • RPO rose to ~$1.8B by 06-30 [1], and the year-by-year split after Q2 is not disclosed. Part of the gap may already be booked.
    • Capacity is described as sold out "into 2028" [4].
  • Severity: Medium

🐻 Bear #7Shorts hold ~43% of tradeable float ahead of December unlock

  • The point: a crowded short meets a large supply of unlocking shares.
  • Evidence:
    • Short interest rose from 6.25M (08-14) to 12.0M (09-30), with 4.3 days to cover [33].
    • Ortex estimated 12.75M on 09-28, with cost to borrow up to 4.78% [34].
    • The tradeable float is about 28.8M Class A shares, since EIP's 19.35M are locked up [9]. On that basis, short interest is ~43% of float (my arithmetic).
    • The lock-up ends on day 180, 2026-12-06 (a Sunday), so the first trading day free is 12-07. At that point 20.27M Class A shares become sellable, plus exchanges from 171.2M Class B units [2].
    • Every insider sale to date was an IPO-related repurchase by the company at $19.85. There have been no open-market buys [8].
  • Why it matters: a weak Q3 a few weeks before the unlock could compound price pressure.
  • 🛡️ Counter-argument:
    • The shorts have to buy back: 12M shares is about 4 days of volume.
    • EIP holds 78% and has every reason to sell in an orderly, marketed way [9].
  • Severity: Medium (price, not business).

🐻 Bear #8Small engines cost more to run than turbines per MWh(new)

  • The point: ERock wins on speed and capex, but loses on running cost. That limits the plant's life once the grid arrives.

  • Evidence: from the PFD [11]:

    Measure McCloud (ERock) El Paso Electric generic turbine
    Heat rate 12,000 Btu/kWh 10,564 Btu/kWh
    Fixed O&M $59/kW-yr $24/kW-yr
    Fuel cost $41.75/MWh $26.75/MWh (existing units)
    Service life 20 years 30 years
    • Generac's CEO said on-site energy costs more than grid power [22].

  • Why it matters:

    • Interpretation: a $1/MMBtu rise in gas adds ~$12/MWh at a 12,000 heat rate (my arithmetic).
    • Once the grid connects, the economic case for continuous running weakens. That caps the "three lives" story (Bull #8) and invites utilities to tender against turbines.

  • 🛡️ Counter-argument: turbines aren't available until 2028–30 [21,26], and for a hyperscaler, compute revenue dwarfs the fuel cost.

  • Severity: Medium. It matters more in 2028–30 than today.

🐻 Bear #9Much bigger rivals are doubling or tripling capacity for 2027–28

  • The point: the scarcity premium shrinks as incumbents add supply.
  • Evidence:
    • Generac: on a path to triple large-megawatt capacity within 12 months [22].
    • Caterpillar: restarting a 10 MW gas recip with ~1.5 GW of capacity, shipping from Q4 2026 [21].
    • Wärtsilä: 1.2 GW of firm data-centre orders and capacity rising to 2.2x by Q1 2029 [24].
    • Cummins: a 130-litre gas prime-power platform, and a 1 GW deal with EquipmentShare (09-29) [23,27].
    • INNIO/VoltaGrid: 1.5 GW of packaged 25 MW units by 2028 [25].
  • Why it matters: ERock's better pricing is due "mainly 2028 and beyond" [4], which is when this capacity arrives.
  • 🛡️ Counter-argument: rivals are scaling but still quote 40–80+ weeks. ERock's 12–18 months for a complete plant, with permitting and installation, remains differentiated [1,22].
  • Severity: Medium

🐻 Bear #10TRA, Class M units and 78% sponsor control leak value(new)

  • The point: public Class A holders own a thin slice, with several claims ahead of them.
  • Evidence:
    • The TRA (tax receivable agreement) pays 85% of tax savings to pre-IPO owners. Early termination is estimated at $658.1M [2].
    • 57.07M Class M profit units can be exchanged into ~52.4M Class A shares, 23.9% on top of 219.4M [1].
    • The plan reserve has ~18.7M shares left (9%) [1].
    • EIP holds 78.1% as-exchanged under its 13D [9].
    • $156.9M of IPO proceeds bought units from pre-IPO holders [1].
  • Why it matters: per-share value is diluted and the TRA takes cash from future tax savings.
  • 🛡️ Counter-argument:
    • Class M units only convert above threshold values, and the 52.4M is an estimate at the 06-30 price [1].
    • The TRA pays only when tax savings are real.
  • Severity: Medium

🐻 Bear #11Korean long blocks carry an unpriced tariff bill(new)

  • The point: the HD supply deal may come with a tariff bill.
  • Evidence:
    • Korea's rate under the US–Korea deal is 15% [31].
    • Section 232 adds 50% on steel content where it applies [31].
    • The rate on engine long blocks (HS 8408/8409) was not found.
    • The prospectus says ERock cannot always pass costs through [2].
  • Why it matters: 15% × $290M ≈ $44M over 2027–28 (my arithmetic).
  • 🛡️ Counter-argument:
    • The IEEPA tariffs were struck down. Caterpillar booked $392M of recoveries [21].
    • Contracts may include pass-through clauses (not disclosed).
  • Severity: Low Medium.

Tailwinds NOT in the forecasts

Per-share numbers use ~273M fully exchanged shares (219.4M Class A + B plus ~52.4M Class M equivalents) and a 25% tax rate (estimate). The share basis behind consensus EPS is not disclosed.

  • Caprock Ranch (1,543 MW) turns into a firm order — 30% chance by 2027-12-31

    • Impact on revenue: +$300M, +14% vs FY2028 consensus ~$2.2B.
    • Impact on EPS: +$0.12 per share, +11% vs FY2028 consensus ~$1.09.
    • How we got the number (estimate): 1,543 MW × $0.79M/MW ≈ $1.2B order. About half of the first-year deliveries (~$600M) is assumed to be above what FY28 consensus already assumes for new orders, so +$300M. Then $300M × 15% incremental pre-tax margin × 0.75 ÷ 273M ≈ +$0.12.
    • Expected value: 0.30 × $0.12 ≈ +$0.037 per share.
    • What it is: the pending 3,428-engine TCEQ application south of Midland becomes a disclosed order [19].
    • Why it's not priced in: it is not in backlog, and no analyst coverage of it was found.
    • Trigger to watch: TCEQ lifts the pause after the 12-10 audit, a permit is issued, or Q3/Q4 backlog jumps above ~$2.5B.
    • Why this probability: the filing proves serious intent, but it needs a frozen permit and an unnamed customer to sign. That is two hurdles, each roughly coin-flip.

  • Margin surprise before 2028 — 17% chance within FY2027

    • Impact on revenue: none.
    • Impact on EPS: +$0.12 per share, +24% vs FY2027 consensus ~$0.50.
    • How we got the number (estimate): +3 percentage points of gross margin × $1.42B FY27 consensus = $43M × 0.75 ÷ 273M ≈ +$0.12.
    • Expected value: 0.17 × $0.12 ≈ +$0.020 per share.
    • What it is: newer orders (Anthropic, reportedly ~$0.9–1.0M/MW) ship earlier than management's "2028 and beyond" [4].
    • Why it's not priced in: the CFO pointed to 2028+, and Wolfe models a mid-20s% gross margin only in 2028 [40].
    • Trigger to watch: adjusted gross margin above ~25% in any 2027 quarter.
    • Why this probability: older contracts dominate 2027 deliveries, and the HD engines may carry tariffs.

  • Anthropic adds installation and O&M to its 470 MW order — 35% chance by 2027-06-30

    • Impact on revenue: +$150M, +7% vs FY2028 consensus ~$2.2B.
    • Impact on EPS: +$0.04 per share, +4% vs FY2028 consensus.
    • How we got the number (estimate): FY25 installation was 53% of product revenue ($47.8M / $90.1M) [2]. Applied to ~$420M of Anthropic product, that is ~$220M, of which ~$150M falls in the first full year. Then $150M × 10% margin × 0.75 ÷ 273M ≈ +$0.04.
    • Expected value: 0.35 × $0.04 ≈ +$0.014 per share.
    • What it is: the equipment-only purchase order is widened to full scope, as the Q2 call flagged was still to be negotiated [4].
    • Why it's not priced in: BofA reportedly models product sales only [41, secondary].
    • Trigger to watch: a backlog step-up without a new customer, or an Anthropic site named.
    • Why this probability: Anthropic may self-install through its data-centre partners. The Rockdale site, for one, uses existing grid power [45].

  • A long ERCOT audit and missed Batch Zero deadline push more projects to self-power — chance not quantifiable

    • Impact on revenue and EPS: not quantifiable. It depends on how many of ~200 GW of queued loads fail verification [17], and on whether the TCEQ freeze lifts at the same time.
    • What it is: the slower grid makes bridge power the default.
    • Trigger to watch: the ERCOT 12-10 filing and Batch Zero delays.

Headwinds NOT in the forecasts

  • The Texas freeze or the 2027 legislature delays new Texas sites in the 2027–28 order book — 30% chance of a visible delay by 2027-06-30

    • Impact on revenue: −$250M, −11% vs FY2028 consensus ~$2.2B.
    • Impact on EPS: −$0.10 per share, −9% vs FY2028 consensus ~$1.09.
    • How we got the number (estimate): ~300 MW of new orders slip × $0.83M/MW ≈ $250M × 15% incremental margin × 0.75 ÷ 273M ≈ −$0.10.
    • Expected value: 0.30 × −$0.10 ≈ −$0.031 per share.
    • What it is: the pause extends beyond the 12-10 audit, or bills restrict on-site gas or data-centre incentives [13,43].
    • Why it's not priced in: guidance (08-11) predates the freeze, and no broker has cut estimates for it.
    • Trigger to watch: TCEQ's 10-19 report, the ERCOT 12-10 filing, and bill filing from November.
    • Why this probability: the freeze already covers pending air permits such as Caprock. Abbott has tied its end to an audit, not to a date. Most backlog sites are already permitted.

  • ERock absorbs a 15% tariff on HD long blocks — 35% chance in FY2027

    • Impact on revenue: none.
    • Impact on EPS: −$0.06 per share, −12% vs FY2027 consensus ~$0.50.
    • How we got the number (estimate): 15% × ~$145M of blocks a year ≈ $22M × 0.75 ÷ 273M ≈ −$0.06.
    • Expected value: 0.35 × −$0.06 ≈ −$0.021 per share.
    • What it is: engine blocks from Korea face the 15% deal rate, and ERock's fixed-price contracts don't pass it through [2,31].
    • Why it's not priced in: the deal is only five days old, and no analyst comment on tariffs was found.
    • Trigger to watch: Q3 call commentary on landed cost, or 2027 gross-margin guidance.
    • Why this probability: the tariff likely applies, but HD may have priced it in, and parts of newer contracts may allow pass-through.

  • The H2 2026 ramp misses guidance — 38% chance by the FY2026 results (~March 2027)

    • Impact on revenue: −$60M, −13% vs FY2026 consensus ~$452M.
    • Impact on adjusted EBITDA per share (EPS is meaningless in FY26): −$0.04.
    • How we got the number (estimate): $60M × ~20% gross margin ≈ −$12M EBITDA ÷ 273M ≈ −$0.044.
    • Expected value: 0.38 × −$0.044 ≈ −$0.017 EBITDA per share.
    • What it is: revenue slips into 2027, and EBITDA guidance of +$3–9M turns negative.
    • Why it's not priced in: FY26 consensus of ~$452M sits at the guidance midpoint [35].
    • Trigger to watch: Q3 revenue below ~$140M, or reduced guidance.
    • Why this probability: the March RPO covered only ~$296M of H2, and the ramp runs through a new plant under weak controls. Bill-and-hold and Q2 bookings narrow the gap. Up from 34% in report #2, given the RPO arithmetic.

  • Anthropic or another large customer cuts or delays an order — 12% chance within 36 months

    • Impact on revenue: −$210M, −10% vs FY2028 consensus.
    • Impact on EPS: −$0.12 per share, −11% vs FY2028 consensus.
    • How we got the number (estimate): half of a ~$420M order × 20% gross margin × 0.75 ÷ 273M ≈ −$0.115.
    • Expected value: 0.12 × −$0.115 ≈ −$0.014 per share.
    • What it is: a hyperscaler re-plans its sites and uses its termination rights [2].
    • Why it's not priced in: consensus treats backlog as firm.
    • Trigger to watch: backlog falling without matching revenue, or termination-fee income.
    • Why this probability: deposits deter cancellation, but the Texas freeze increases the risk that sites move.

  • The PUCT denies McCloud, or El Paso Electric rejects the conditions — 14% chance by 2027-03-31

    • Impact on revenue: −$150M, −11% vs FY2027 consensus ~$1.42B.
    • Impact on EPS: −$0.08 per share, −16% vs FY2027 consensus ~$0.50.
    • How we got the number (estimate): ~$150M of undelivered McCloud gensets is deferred × 20% gross margin × 0.75 ÷ 273M ≈ −$0.08.
    • Expected value: 0.14 × −$0.08 ≈ −$0.011 per share.
    • What it is: commissioners follow the "deny" branch, or El Paso Electric won't accept the hold-harmless condition [11].
    • Why it's not priced in: ERock's filings treat McCloud as under construction [3].
    • Trigger to watch: El Paso Electric and Meta exceptions (10-13), the SOAH letter (11-06), and the open-meeting memo [12].
    • Why this probability: Meta already pays during the bridge period, so a conditioned approval is the base case. The judges' language is unusually sharp.

  • Heavy selling around the 12-07 lock-up, amplified by the short base — 45% chance of a noticeable price drop in December

    • Impact on revenue and EPS: not quantifiable. It affects the price, not the business.
    • What it is: 20.27M Class A shares unlock, and Class B exchanges follow [2].
    • Trigger to watch: an S-3, Form 144 filings, or a marketed secondary.
    • Why this probability: short interest has doubled into the event, and the sponsor holds 78%.

Investment Risk Score: 58/100 (elevated)

# Factor Weight Measured value Score Why
1 Liquidity & funding runway 15% $626.6M cash, no debt, undrawn $250M ABL; but only ~$98M excluding $528M deposits; H1 burn ex-deposits ~$16M/month 42 FCF is positive only because of deposits. Excluding them, runway is roughly 6 months on cash, or ~20+ months with the ABL. No going-concern language.
2 Leverage 10% Net cash, $0 debt, $29.8M leases; TRA early-termination estimate $658M (contingent) 18 Net cash. The TRA is a contingent claim, not debt.
3 Dilution (last 12 months + overhang) 10% 219.4M A+B shares flat since IPO; IPO added 27.9M new shares (+12.7%); Class M ~52.4M (+23.9%) + ~18.7M plan reserve 60 12-month change ≈ +13%, with an overhang of ~33% that converts only above thresholds.
4 Profitability 10% H1 adj. EBITDA −$26.4M (−37%); FY25 −12.4%; FY26 guide +$3–9M 66 Clearly loss-making, guided to improve in H2.
5 Revenue quality & concentration 10% Top customer 28–37% of quarterly revenue, 62% of receivables; services 25% (FY25) to 34% (Q2) of revenue 72 Top customer above 35% in Q1, and mostly one-off equipment revenue.
6 Execution & track record 10% 3 unremediated material weaknesses; no guidance track record; first guide needs a 5x H2 ramp; late Form 4 82 The anchor puts a material weakness in the 81–100 band.
7 Competitive position & disruption 10% 12–18-month delivery vs 40–80+ weeks for rivals; but higher heat rate and O&M than turbines; Cat, Generac, Wärtsilä and Cummins scaling 52 A strong niche today that becomes contested in 2028.
8 Cyclicality & commodity/macro sensitivity 10% Tied to the AI capex cycle; peers fell 14–24% in 2020 [28,29,30]; Q2 revenue −42% y/y on lumpiness 62 Highly cyclical demand driver, with lumpy deliveries.
9 Regulatory, legal & geopolitical 5% TCEQ freeze (no end date), McCloud PFD, 2027 legislature, Korea tariff, Hagens Berman investigation 72 Large pending rulings in its home state.
10 Valuation & expectations 10% EV ~1.4–1.9x FY27 sales; all-Buy coverage; short interest ~43% of tradeable float; 12-07 unlock 70 Not expensive on forecasts, but heavy short interest with a fragile near-term story.

Overall: weighted average = 6.3 + 1.8 + 6.0 + 6.6 + 7.2 + 8.2 + 5.2 + 6.2 + 3.6 + 7.0 = 58.1 → 58.

Change vs last: −4 versus 62 under the old six-area method. This is mostly a method effect. The v3 factors give separate weight to the net-cash, no-debt balance sheet (Leverage 18), which the old "Financial health & dilution" area blended with dilution and the TRA. Underlying risk is unchanged to slightly higher:

  • Execution is up: the RPO gap and the late filing.
  • Regulation is unchanged: the freeze, now confirmed as covering pending air permits.

Ranking check: EROC (58) sits between ALM (55) and DUOT (66). It is riskier than ALM because its revenue still rests on an unproven 5x ramp and a single Texas customer under regulatory review, against ALM's producing mine. It is safer than DUOT because it holds net cash and contracts with investment-grade hyperscalers rather than one cash-short counterparty.

Conclusion

  • What you are really betting on: that ERock turns ~$1.8B of prepaid, contracted orders into delivered gensets in 2026–28 through a new Houston plant, and wins Caprock-scale follow-on orders before Texas politics and bigger rivals close the window.
  • The 2–3 things that will decide the outcome:
    1. Q3/Q4 2026 deliveries against the $363–393M H2 requirement.
    2. Texas: TCEQ (10-19), the ERCOT audit (12-10), the McCloud vote, and the January legislature.
    3. New orders (Caprock, Anthropic full scope) that fill the $100–300M 2027 gap.
  • Signposts:
    • Bull confirmed if:
      • Q3 revenue is ≥ ~$140M and guidance is held.
      • Backlog is above $2B, or Caprock or another ≥300 MW order is announced.
      • TCEQ resumes permits after the audit.
      • McCloud is approved with hold-harmless conditions and El Paso Electric accepts.
    • Bull broken if:
      • Guidance is cut, or revenue is shifted to 2027.
      • The freeze extends into 2027, or bills restrict on-site gas.
      • Backlog or contract liabilities fall without matching revenue.
      • A new material weakness or a restatement appears.
  • New since the last report:
    • Caprock Ranch found: a 1,543 MW, 3,428-engine ERock air-permit application (pending, unnamed customer). It is both the biggest pipeline signal and the clearest freeze exposure.
    • The delivery arithmetic got tighter: the March RPO schedule ($335.8M for the rest of 2026, $864.5M for 2027, $14.8M for 2028) covers neither H2 guidance nor 2027 consensus without new bookings. Q2 revenue was −42% year on year.
    • McCloud record read in full:
      • Docket 59076 confirmed.
      • Capital cost is $1,609/kW, below the Entergy Texas plants, but the heat rate and O&M are worse than a turbine's.
      • Exceptions are due 10-13.
      • The stock is now $11.95 (not mid-teens).
      • The TRA early-termination estimate is $658M, and Class M adds ~52M shares.

Open questions & data gaps

  • McCloud contract protection (deposits and termination payments if the CCN is denied): not found. High importance.
  • Caprock Ranch customer, permit type and status under the freeze; whether it is in backlog: not found. High importance.
  • Whether TCEQ's pause covers Standard Permit registrations, ERock's usual route: no guidance yet. The 10-19 report should tell. High importance.
  • Customer A's identity and the "recently executed customer agreement" behind the +$56.9M in deferred income: not disclosed. Medium importance.
  • Verbatim Q2 call transcript: still only condensed extracts (Investing.com, Quartr). The pricing per MW and third-line quotes are unverified. Medium importance.
  • Tariff rate on Korean engine long blocks (HS 8408/8409) and CARB-DG coverage of the HD engine: not found. Medium importance.
  • Consensus basis: EBITDA consensus not found; whether consensus EPS uses Class A only or all shares is unclear; the reasons for the JPMorgan and Guggenheim target cuts were not found. Medium importance.
  • Q3 date: unconfirmed. Ortex says 11-12 and YCharts 11-09. Low importance.
Sources
  1. ERock, Inc. Form 10-Q, quarter ended 2026-06-30 — SEC EDGAR — 2026-08-12 — https://www.sec.gov/Archives/edgar/data/2110029/000119312526346922/eroc-20260630.htm
  2. ERock, Inc. Form 424B4 prospectus — SEC EDGAR — 2026-06-10 — https://www.sec.gov/Archives/edgar/data/2110029/000119312526265898/d12401d424b4.htm
  3. ERock Reports Second Quarter 2026 Results (8-K Ex. 99.1) — SEC EDGAR — 2026-08-11 — https://www.sec.gov/Archives/edgar/data/2110029/000119312526344714/eroc-ex99_1.htm
  4. Earnings call transcript: ERock jumps 23% after Q2 2026 revenue surge (condensed extract, not verbatim) — Investing.com — 2026-08-12 — https://www.investing.com/news/transcripts/earnings-call-transcript-erock-jumps-23-after-q2-2026-revenue-surge-93CH-4855247
  5. ERock Q2 2026 slides: AI power demand drives record backlog — Investing.com — 2026-08-12 — https://www.investing.com/news/company-news/erock-q2-2026-slides-ai-power-demand-drives-record-backlog-93CH-4855296
  6. ERock Q2 2026 earnings summary (AI summary) — Quartr — 2026-08 — https://quartr.com/events/erock-inc-eroc-q2-2026_FT8SuO7H
  7. EDGAR submissions index, CIK 2110029 — SEC — retrieved 2026-10-11 — https://data.sec.gov/submissions/CIK0002110029.json
  8. Form 4 / Form 3 / Schedule 13G/A, W. T. McAndrew Jr. — SEC EDGAR — 2026-09-04 — https://www.sec.gov/Archives/edgar/data/0002110029/000149315226041632/ownership.xml
  9. Schedule 13D, Energy Impact Partners group — SEC EDGAR — 2026-06-18 — https://www.sec.gov/Archives/edgar/data/0002110029/000119312526275971/xslSCHEDULE_13D_X02/primary_doc.xml
  10. HD Construction Equipment signs ~390bn won engine long-block supply deal with ERock — Herald Corp / Korea Herald — 2026-10-07 — https://biz.heraldcorp.com/article/10895760
  11. Proposal for Decision, PUCT Docket 59076 / SOAH 473-26-14758 (El Paso Electric McCloud CCN) — SOAH / PUCT — 2026-09-23 — https://interchange.puc.texas.gov/Documents/59076_176_1686197.PDF
  12. Memo on exceptions and replies schedule, Docket 59076 — PUCT Commission Counsel — 2026-09-29 — https://interchange.puc.texas.gov/Documents/59076_177_1688242.PDF
  13. Letter to TCEQ Executive Director on data-centre permits — Office of the Texas Governor — 2026-09-21 — https://gov.texas.gov/uploads/files/press/TCEQ_Data_Center.pdf
  14. Governor Abbott Orders TCEQ to Pause All Permit Issuance for Data Center Projects — Bracewell — 2026-09-23 — https://www.bracewell.com/resources/governor-abbott-orders-tceq-to-pause-all-permit-issuance-for-data-center-projects/
  15. Data Center Permit Freeze Reaches Self-Powered Projects — Texas Dispatch — 2026-09-23 — https://thetexasdispatch.com/abbott-data-center-permit-freeze-self-powered-fermi-panhandle/
  16. Texas Pauses All Environmental Permitting for Data Centers — Beveridge & Diamond — 2026-09 — https://www.bdlaw.com/publications/texas-pauses-all-environmental-permitting-for-data-centers/
  17. ERCOT targets December 10 for the Abbott data-center audit; Batch Zero April 2027 — mgrid — 2026-09-05 — https://mgrid.org/2026/09/05/ercot-december-10-abbott-data-center-audit-batch-zero-april-2027/
  18. Order Adopting New 16 TAC §25.194, Project 58481 — PUCT — 2026-09-18 — https://interchange.puc.texas.gov/Documents/58481_218_1684654.PDF
  19. West Texas ranch could host over 3,400 gas-fired engines for power project — Oil & Gas Watch — 2026-07-29 — https://news.oilandgaswatch.org/template/brief/west-texas-ranch-could-host-over-3-400-gas-fired-engines-for-power-project
  20. Texas Electric Utility Only Considered Gas to Power $10 Billion Meta Data Center — Inside Climate News — 2026-09-24 — https://insideclimatenews.org/news/24092026/el-paso-electric-only-considered-gas-to-power-meta-data-center/
  21. Caterpillar 2Q 2026 Earnings Call Transcript — Caterpillar IR — 2026-08-04 — https://s25.q4cdn.com/358376879/files/doc_financials/2026/q2/2Q-2026-Earnings-Call-Transcript.pdf
  22. Generac (GNRC) Q2 2026 Earnings Call Transcript — Motley Fool — 2026-07-29 (published 2026-08-07) — https://fool.com/earnings/call-transcripts/2026/08/07/generac-gnrc-q2-2026-earnings-call-transcript
  23. Cummins Q2 earnings call highlights (summary only) — MarketBeat — 2026-08-04 — https://www.marketbeat.com/instant-alerts/cummins-q2-earnings-call-highlights-2026-08-04/
  24. Wärtsilä Half-year Financial Report January–June 2026 — Wärtsilä — 2026-07 — https://www.wartsila.com/fwd/result-presentations/interim-report-q2-2026
  25. INNIO Secures Additional Major Order from VoltaGrid: 1.5 GW — INNIO — 2026-02-11 — https://www.innio.com/en/news-media/press-releases/innio-secures-additional-major-order-from-voltagrid-1-5-gw-for-behind-the-meter-power-generation/
  26. GE Vernova Q2 2026 results press release (8-K) — SEC EDGAR — 2026-07-22 — https://www.sec.gov/Archives/edgar/data/0001996810/000199681026000147/gevpressrelease2q26.htm
  27. EquipmentShare and Cummins announce collaboration to deploy 1 GW of mobile power — Business Wire (via Finviz) — 2026-09-29 — https://finviz.com/news/396492/equipmentshare-and-cummins-announce-landmark-collaboration-to-deploy-1-gigawatt-of-mobile-power
  28. Generac Holdings FY2020 Form 10-K — SEC EDGAR — 2021-02-23 — https://www.sec.gov/Archives/edgar/data/1474735/000143774921003793/gnrc20201231_10k.htm
  29. Cummins FY2020 Form 10-K (snippet only) — SEC EDGAR — 2021 — https://www.sec.gov/Archives/edgar/data/26172/000002617221000012/cmi-20201231.htm
  30. Power Solutions International FY2023 Form 10-K (snippet only) — SEC EDGAR — 2024 — https://www.sec.gov/Archives/edgar/data/1137091/000113709124000007/psix-20231231.htm
  31. Elevate GA Trump Tariff Tracker — IWLA — 2026-08-24 — https://iwla.com/wp-content/uploads/2026/08/Elevate-GA-Trump-Tariff-Tracker-updated-8.24.26.pdf
  32. EROC daily price history — Yahoo Finance — 2026-06-10 to 2026-10-09 — https://query1.finance.yahoo.com/v8/finance/chart/EROC?range=6mo&interval=1d
  33. EROC short interest history — MarketBeat — retrieved 2026-10-11 — https://www.marketbeat.com/stocks/NYSE/EROC/short-interest/
  34. EROC shorts double in a month — ORTEX — 2026-09-28 — https://ortex.news/articles/174709/eroc-shorts-double-in-a-month-as-activist-looms-large
  35. ERock forecast — StockAnalysis — retrieved 2026-10-10/11 — https://stockanalysis.com/stocks/eroc/forecast/
  36. EROC stock forecast — WallStreetZen — retrieved 2026-10-11 — https://www.wallstreetzen.com/stocks/us/nyse/eroc/stock-forecast
  37. EROC forecast — TipRanks — retrieved 2026-10-11 — https://www.tipranks.com/stocks/eroc/forecast
  38. EROC price target and ratings — MarketBeat — retrieved 2026-10-11 — https://www.marketbeat.com/stocks/NYSE/EROC/forecast/
  39. Barclays initiates ERock with Overweight on power scarcity — Investing.com — 2026-07-06 — https://www.investing.com/news/analyst-ratings/barclays-initiates-erock-stock-with-overweight-on-power-scarcity-93CH-4776181
  40. Wolfe Research initiates ERock with Outperform — Investing.com — 2026-07-06 — https://za.investing.com/news/stock-market-news/wolfe-research-initiates-erock-stock-with-outperform-rating-93CH-4355841
  41. Bank of America sees growth potential in power company tied to Anthropic IPO (secondary) — Crypto Briefing — 2026-08-13 — https://cryptobriefing.com/bank-of-america-erock-anthropic-ipo-power/
  42. ERock (EROC) investigation (law-firm solicitation; search text only) — Hagens Berman — 2026-08 — https://www.hbsslaw.com/cases/erock-inc-eroc-investigation
  43. Texas Data Center Tax Exemption Faces Serious Legislative Risk — Holland & Knight — 2026-07-30 — https://www.hklaw.com/en/insights/publications/2026/07/texas-data-center-tax-exemption-faces-serious-legislative-risk
  44. How AI Labs Are Solving the Power Crisis: The Onsite Gas Deep Dive (snippet only) — SemiAnalysis — undated — https://newsletter.semianalysis.com/p/how-ai-labs-are-solving-the-power
  45. Riot discloses $9.1B AI data center lease (secondary) — LetsDataScience — 2026-08 — https://letsdatascience.com/news/riot-discloses-91b-ai-data-center-lease-report-identifies-an-20525fa3
  46. Stem, Inc. Form 10-K for FY2025 — SEC EDGAR — 2026 — https://www.sec.gov/Archives/edgar/data/1758766/000175876626000016/stem-20251231.htm
All EROC Bulls vs Bears reports
Disclaimer. AI-generated research and podcast (Claude by Anthropic, voices by Google Gemini) for educational purposes — not financial advice and not a recommendation to buy or sell. Willem (MR.INVEST) owns this stock. Check the company’s own filings before making any decision.
← All Bulls vs Bears reports
EROC Bulls vs Bears — Stock Analysis — MR.INVEST