← Back to the plan

Kora Power, due diligence

Prepared 2026-09-06 for The Recording Club / Tiny Room Studios LLC, 1660 9th St, Santa Monica CA 90404 (SCE). Question on the table: TRC holds a ~May 2026 reservation with Kora Power to "catch power spikes and put us in a lower demand tier." Ship dates have slipped from ~June 2026 to August-September 2026. Should TRC wait, cancel, or replace?

Headline: Kora sells a residential, split-phase, whole-home panel-replacement system that has never shipped to any customer. It is not a commercial peak-shaving device, it has no documented demand-limit setpoint, and its 120/240 V split-phase architecture is a poor match for a commercial building. Cancel it as the plan. The $100 is fully refundable and immaterial either way.

1. What the product actually is

1.1 Architecture

Kora sells the Founders Edition Home Energy System , three components sold together, all of which must be installed for the system to work as advertised:

Sources: Tech Specs, Founders Edition · FAQ · Founders Edition product page

1.2 Capacity and discharge

From Kora's own tech-spec page:

The 112 kWh / 91.2 kW figures in Kora's press release are the maximum four-stack configuration, not a shipping product.

1.3 How it decides to discharge, no demand-limit setpoint is documented anywhere

This is the single most important finding for TRC's use case.

Nothing on Kora's site, in its FAQ, in its tech specs, or in its press materials describes a kW demand setpoint, the control that a peak-shaving device needs, where you tell it "never let grid import exceed X kW" and it discharges automatically to hold that ceiling.

What Kora actually advertises is:

• Time-of-use load shifting, "Charge your battery when power is cheap. Discharge it when rates spike." That is a clock-driven rate arbitrage feature, not a demand-limiting one.
• AI-powered load shifting, undefined; no control document published.
• Circuit-level load control, the Smart Panel can shed circuits. Kora's Utilities Manager page pitches this to utilities as transformer-overload prevention, which is the closest thing to demand management in their materials, but it is load shedding (turning things off), not battery-backed demand limiting.
• Energy Trading (VPP), Kora's own product page says this "launches in late 2026." It does not exist yet.

Assessment (high confidence): Kora has not published a demand-limiting control, and TOU shifting will not reliably hold a facility under a 20 kW demand ceiling. A facility whose peak is driven by coincident equipment operation needs a metered setpoint with current transformers on the service conductors, continuously comparing measured grid import against a target. That is a specific, documented feature in competing products (see §6) and its absence from every Kora document is meaningful.

1.4 Split-phase only, a probable disqualifier for TRC

Kora's inverter output is 120/240 V split phase, and the Smart Panel's grid input is specified as "Standard Split Phase 120 V / 240 V at 60 Hz. Compatible with nearly all U.S. homes."

TRC's building has at least one three-phase circuit (the sauna, per Farzad, 2026-09-04). A Santa Monica commercial building of this type is commonly served at 120/208 V three-phase. If TRC's service is three-phase, Kora's equipment cannot be installed on it without a transformer and a fundamentally different design than Kora sells or supports.

This needs to be resolved by reading the meter and main panel, and it should be resolved before any battery of any brand is purchased. Until then, treat it as a live risk, not a settled fact.

1.5 Certifications and listings, Kora's own claims are internally inconsistent

Kora's tech-spec page claims a long list of certifications as accomplished fact, including "UL9540 Certified," "UL1741 SA/SB Certified," "UL1973 Certified," "UL 1998 Certified," "UL 1699B Certified."

Three things contradict that:

1. Kora's own LinkedIn post of 2026-02-26 describes only "UL 9540A testing." UL 9540A is a fire-propagation test method, not a product listing. Claiming 9540A on social while claiming full 9540 certification on the spec page is the kind of gap that usually means the listing is not finished.
2. Kora emailed TRC a "Manufacturing Update: Certification Progress" on 2026-08-07. A company does not send certification-progress updates about certifications it already holds.
3. I could not independently verify any Kora listing. UL's Product iQ database returned HTTP 403 to automated access and requires an account. This is unverified, not disproven, someone with a Product iQ login should search "Kora Power" directly before drawing a final conclusion.

Assessment (moderate confidence): Kora does not yet hold a completed UL 9540 listing for the Founders Edition, and the website overstates its certification status. UL 9540 certification for a novel integrated system typically runs many months and requires the final production hardware; a company still sending progress updates in August is unlikely to have a listing in hand.

Why this matters concretely: Santa Monica's building department will require a listed ESS for permit under NEC Article 706 and NFPA 855, and SCE will require the inverter to appear on the CEC certified equipment list for a Rule 21 interconnection (SCE SGIP/ERB accelerated PTO checklist, requirement 3). No listing means no permit and no permission to operate.

1.6 Installation, permitting, and interconnection burden

Kora markets the install as "less than two hours with a qualified electrician." That is the panel swap only. The real path for TRC:

1. Licensed electrician, panel replacement or subpanel addition, concrete pad for the Powerblocks (Kora specifies a ≥700 lb static-load foundation).
2. City of Santa Monica electrical permit; NEC 706 plan set; NFPA 855 fire-area compliance for the ESS.
3. SCE Rule 21 Non-Export interconnection application , required. A behind-the-meter battery that never exports still needs this. Use Form 14-732, submitted through SCE's Grid Interconnection Processing Tool, with an $800 non-refundable review fee, a single-line diagram, a site plan, and a certificate of liability insurance. (SCE Rule 21 non-export page)
4. There is an accelerated PTO path, but it caps total inverter nameplate at 10 kVA/kW and requires stand-alone storage used for "peak shaving and backup mode only," a CEC-listed inverter, load-side-of-main connection, and Protection Option 3 or 6. A single Kora Powerblocks stack at 11.4 kW continuous exceeds that 10 kW cap and would fall to the standard, slower track.

Realistic interconnection timeline for a standard-track Rule 21 non-export project: weeks to a few months after equipment is on site and inspected.

1.7 Price and financing

• Reservation: $100, fully refundable.
• Founders Edition system: $3,800 listed on the product page. The Smart Panel alone is listed at $2,000, with a "Total Cost: $9,600" and a claim of "potential tax incentives of $16,205" that is not explained and does not reconcile with the stated prices. Treat Kora's pricing display as unreliable.
• Financing: Kora's support FAQ lists a "How do I finance my purchase?" entry, but no terms, lender, or program is published.
• The reservation locks in the price shown at checkout.

2. The company

2.1 Founder and team

Greg Connolly, Founder/CEO. Co-founded Trifecta Nutrition (organic meal delivery, Sacramento) with his sister Elizabeth Connolly in 2015; ran it to >$500M cumulative revenue and ~$110M ARR, stepped down to a board role in May 2025. Prior: Amara Beverage, SmartTurn, ADP. (LinkedIn · Equilar)

This is a real operator with a real prior exit-scale outcome , in direct-to-consumer food, not in power electronics or in any UL-listed hardware category. That is the relevant caveat: he is a first-time hardware founder in one of the most heavily regulated consumer product categories in the US.

Team listed on korapower.com/pages/about:

• Carlos Restrepo, CTO, described as former CTO of Sonnen "from product development to exit to Shell for $500M." This is the strongest technical signal on the page and would be worth verifying independently.
• Jerremy Spillman, CRO, prior exit to JLL; built "AI HVAC HERO."
• Tyler Thomas, Chief AI Officer, former CTO of Trifecta.
• Abe Fechter, Power Electrical Engineer, 20+ years.
• Murray Baker, Principal Firmware Engineer. Peter Macdonald, Systems Engineer (ex-Domatic).
• Kyle Pate, Director of Marketing (former Trifecta).
• Paige Tranchina, Customer Success Manager, the person TRC has been corresponding with. Her title on Kora's site is "Customer Success Manager," described as "Experienced DTC customer experience leader." She is a CX manager, not an engineering or program contact. Escalation on a go/no-go should go to Connolly directly.

Advisors listed include a SMUD CEO/GSEC board member, the Executive Director of the UC Davis Energy and Efficiency Institute, the CEO of the Connectivity Standards Alliance, the CEO of Ansync Labs (named as "R&D Partner and investor"), and a 20-year CAISO Director of State Affairs. Advisors are cheap; the Ansync Labs relationship is the interesting one, since it suggests Kora's R&D and possibly build is outsourced.

2.2 Funding, hard numbers from the SEC

Kora filed a Form D and Form D/A on 2026-04-10 (filing index):

Named investors from the September 2025 launch: Moneta Ventures and Growth Factory Ventures, a $2.6M SAFE. (Kora press release · LinkedIn announcement)

Assessment (high confidence): ~$5.3M raised total, with a round still open. For context, bringing a novel integrated ESS through UL 9540 listing, tooling, contract manufacturing, and first production is routinely a $20M-$50M undertaking. Enphase, Tesla, FranklinWH, and Sonnen each spent orders of magnitude more to get to first shipment. $5.3M and 9 people is a seed-stage company attempting a capital-intensive regulated hardware product. That does not make failure certain, but it makes the published timelines implausible.

2.3 Ship-date history, five slips, zero deliveries

The website is currently displaying three mutually contradictory ship dates simultaneously, "early 2026," "later July," and "2026." That is not a company running a tight launch program.

2.4 Evidence of real installations, none found

I searched for customer installs, install photos, YouTube walkthroughs, and third-party reviews. Findings:

• No installed-unit photos, no install videos, no third-party review, no independent teardown, no installer forum discussion anywhere I could find.
• Kora's own documentation library (support.korapower.com/documents) contains install manuals and single-line-diagram templates revised April through July 2026, consistent with a company preparing for first installs, not one that has done them.
• Kora's marketing claim that "Early adopters have already cut $1,500 monthly bills to zero" appears in the press release with no named customer, site, or data. Given that no product has shipped, this claim is not credible as written.
• Solar Power World's own comment section flags the coverage as unexamined: "This looks more like a press release from Kora than a review or analysis of the product... Did you see the product in an installation?"

2.5 Complaints, BBB, Reddit, social

I found no BBB complaints, no Trustpilot reviews, no Reddit threads, and no consumer complaints about Kora Power. This is not reassurance, it is a consequence of having no customers. A company with zero deliveries generates zero complaints.

Kora maintains Facebook, Instagram, YouTube, and LinkedIn accounts (linked from their About page). LinkedIn activity is regular but low-engagement, posts in Feb, May, and Aug 2026 drawing 12 to 67 reactions against ~340 followers. Their public content is product marketing and explainer material, not install documentation.

Assessment (high confidence): there is no public evidence that Kora has delivered, installed, or energized a single unit for any customer, anywhere, as of 2026-09-06.

3. What the August 7 "Certification Progress" update most likely means

TRC received an email titled "Manufacturing Update: Certification Progress" from Greg Connolly on 2026-08-07. Based on the surrounding evidence, the most probable readings, in order:

1. Most likely (moderate confidence): they are mid-way through UL 9540 / UL 1741 SB evaluation and it is the binding constraint on shipping. A full UL 9540 system certification requires the final production hardware, component-level listings underneath it (UL 1973 for the battery, UL 1741 SB for the inverter, UL 1998 for the safety software), plus UL 9540A fire-propagation test data to satisfy NFPA 855. Realistic elapsed time from submission to listing for a first-time integrator with novel hardware is 6 to 18 months, and revisions to hardware restart portions of it. Their February 2026 LinkedIn post referencing only 9540A testing is consistent with being roughly a year into that pipeline.
2. Also plausible: a component-level milestone was reached and framed as progress on the whole. The Smart Panel's UL 67 and UL 916 claims are for a much simpler product than the ESS and would plausibly land first. Note the Smart Panel is listed as a subpanel, which constrains how it can be installed.
3. Contract manufacturer: unknown. Kora's workforce is "distributed across United States and China" per LinkedIn, and Ansync Labs is described on Kora's own About page as "R&D Partner and investor." A reasonable inference is Chinese cell and possibly module supply with US integration, but I could not confirm a named CM and will not guess one.

What would settle it: ask Kora, in writing, for the UL file number and the issuing NRTL for the Powerblocks ESS, and the CEC equipment-list entry for the inverter. Both are public, verifiable, and instantly dispositive. A company that holds them will produce them in an hour. A company that will not produce them does not hold them.

4. Probability they ship a working, listed, installable unit to a Santa Monica commercial customer

I am splitting this because the questions are very different, and because "ship a box" and "legally operating at 1660 9th St" are separated by months of permitting.

Reasoning.

By Oct 1 (3.5 weeks away): Even in the best case where a listing lands tomorrow, TRC still needs a licensed electrician, a Santa Monica permit, an NEC 706 plan set, NFPA 855 review, an $800 SCE Rule 21 non-export application on the standard track (the 11.4 kW stack exceeds the 10 kW accelerated-PTO cap), and permission to operate. That sequence does not compress into three weeks. Near zero.

By Dec 31: Shipping a first article is genuinely possible , the install manuals and SLD templates dated April-July 2026 suggest they are preparing. But first shipments from seed-stage hardware companies slip, and Kora has already slipped five times in twelve months while leaving three contradictory dates on their own website. Then add 6-14 weeks of permitting and interconnection on top. Low.

Ever: Connolly is a credible operator with real capital, a plausibly strong CTO hire, and a round still open. Most likely long-run outcomes, roughly: they ship a residential product later than promised (~55%), they pivot or get acquired for the smart-panel IP (~20%), or they run out of money before first shipment (~25%). But note that "ever ships to TRC" is much lower than "ever ships," because of the split-phase problem: even a fully successful Kora may simply not be installable on TRC's service, and Kora has shown no intent to build a commercial three-phase product.

Confidence in these numbers: moderate. They rest on public information only. The largest uncertainties I could not resolve are (i) whether a UL listing already exists (Product iQ is login-gated) and (ii) whether TRC's service is single-phase or three-phase.

Strongest argument against my own conclusion, stated fairly: Connolly has actually built and scaled a company to nine figures, the CTO claim (ex-Sonnen through a $500M exit) is exactly the right background, the install documentation is real and recently revised, and the SEC filing shows a live round with $2.25M still available. A company one quarter from first shipment looks a lot like this. If the UL listing turns out to be issued, my probabilities for (b) roughly double. That is why the verification question in §3 is the one that matters, and why cancelling a refundable $100 costs TRC nothing even if I am wrong.

5. Refund and cancellation terms

From Reservation Terms & Conditions and the Refund Policy:

• The $100 reservation fee is fully refundable at any time before the ship date. No obligation to purchase until the system ships. Cancel by emailing support@korapower.com with the order number, or call 916-245-6649.
• Pre-shipping cancellations are refunded in full "provided manufacturing or assembly has not begun." That qualifier is a real hook, once Kora says it has started building TRC's unit, the clean exit narrows.
• After shipping, TRC would have 30 days to request a refund for any reason per the reservation terms. But the refund policy contradicts this in part: it says returns "are not accepted for installed or commissioned equipment, custom configurations, or energy storage systems (Powerblocks modules, Smart Panels) once shipped," with unopened standard configurations returnable only "at Kora's discretion, subject to inspection and restocking fees," and return shipping at the customer's expense under hazmat rules for lithium batteries.
• Note the reservation terms reference products called the "Keystone System" and "Keystone Plus System", names that appear nowhere else on Kora's site. Their legal documents are stale relative to their product line.
• Governing law: California. Liability capped at the reservation fee.

Practical read: TRC's financial exposure today is $100 and it is recoverable with one email. The exposure that matters is not the deposit, it is the invoice for the full system value that arrives on the confirmed ship date, at which point the return path becomes narrow and expensive. Do not let a Kora invoice be paid without a UL file number in hand.

6. Would a battery like Kora's even solve TRC's problem?

6.1 First, a problem with the premise, and it applies to the soft starters too

SCE bills demand as the maximum 15-minute average kW in the billing period. A compressor's locked-rotor inrush lasts a fraction of a second to a few seconds. Averaged across a 900-second window, a 2-second inrush contributes roughly 0.2% of that interval's value. Inrush is effectively invisible to a demand meter.

That means Ari's 19.5 kW reading on 2026-01-22 was almost certainly produced by sustained coincident running load over a full 15 minutes, several AC units running simultaneously alongside the video wall, studio gear, sauna, refrigeration, and any EV charging, not by a startup spike.

This is the same conclusion Farzad reached twice and that Victor stated on 2026-08-28 ("compressor savers help with startup amps but not running amps so they won't do major changes to your electric bill"). They are right, and the $4,500 soft-starter quote is very likely money spent on the wrong problem. Greg's reply, "It's just the startup amps that are the problem for us," is the premise that needs to be tested before any hardware is bought.

(Flagging this here because it reframes the Kora question. It overlaps the soft-starters and SCE-tariff workstreams; those agents should confirm independently against actual interval data.)

Consequence for the battery question: a battery will work for TRC where soft starters will not, because a battery does exactly the right thing, it holds sustained grid import below a setpoint for the length of the peak. But it only works if it is sized for the sustained kW and the duration, and if it has a real demand-limit control.

6.2 Sizing for TRC

Working from the known data points (19.5 kW peak 2026-01-22; 17.6 kW on 2026-02-28; 19.5 kW internal ceiling; ~20 kW tariff threshold):

• Power needed: enough to cover the gap between the uncontrolled peak and the target ceiling, plus margin. If uncontrolled peaks reach 25 kW and the target is a hard 18 kW, that is 7 kW continuous. If peaks are closer to 21-22 kW, 4-5 kW suffices.
• Energy needed: the peak window duration times the shaved kW. A 3-hour evening peak at 5 kW is 15 kWh delivered, which at ~90% round-trip efficiency and a 20% reserve floor means **~20 kWh usable installed**. A 4-hour window at 7 kW pushes that to ~35 kWh.
• Recommendation: size for ~8-10 kW continuous and 25-30 kWh usable, and do not commit to a number until interval data is in hand.

This is why the electrical tracker Greg wants comes first, not last. SCE 15-minute interval data (Green Button download) plus a permanent monitor will tell TRC exactly what the peak shape is, how long it lasts, and what is driving it. That single input determines whether the answer is a $0 operational change (staggering equipment, which TRC already started on 2026-03-17), a $2k monitoring install, or a $25k battery. Buying a battery before measuring is the same mistake as buying soft starters before measuring.

6.3 Alternatives available now

The critical column is demand-limit setpoint, a user- or installer-settable kW threshold that the system holds by discharging. Time-of-use scheduling is not the same thing and will not reliably hold a ceiling.

Sources: Sol-Ark Grid Peak Shaving app note · Sol-Ark 30K-3P-208V manual · Sol-Ark 30K-3P product page · Powerwall 3 datasheet · Powerwall 3 design considerations · Powerwall 3P · Enphase IQ Battery 10 datasheet · FranklinWH operating modes · FranklinWH user manual

Two corrections to the assumptions in the brief, both stated with high confidence:

1. FranklinWH does not have a demand-limit / peak-shaving setpoint. Its documented modes are Emergency Backup, Self-Consumption, and Time of Use. FranklinWH's marketing blog uses the phrase "peak shaving" generically about batteries; their actual product documentation exposes no kW threshold. Do not specify FranklinWH for this job.
2. Tesla Powerwall 3 does not expose a non-zero grid-import ceiling in the standard residential product. Its behavior targets minimizing grid import, not holding a specific kW. Third-party commercial marketing that claims demand-charge reduction with Powerwall is describing an effect, not a control.

The product that actually does what TRC needs is Sol-Ark, in either the split-phase 15K or the three-phase 30K-3P-208V, because it has an explicit, documented, installer-settable grid consumption threshold with current transformers on the service conductors. The three-phase model also carries UL 1741 SB, IEEE 1547-2018, CEC, and SGIP listings on its datasheet, which is exactly what SCE's Rule 21 process asks for.

6.4 SGIP eligibility, probably closed to TRC

• The General Market Large-Scale Storage budget (the non-residential category TRC would use) shows $0.25/Wh at Step 5 with N/A for Steps 6 and 7 in the SGIP incentive tables, and the CPUC lists it as "available through 2025." One 2026 industry source states it closed to new applications after December 30, 2025, with remaining budgets waitlisted.
• Equity Resiliency ($1,000/kWh) requires a Tier 2/3 High Fire-Threat District or repeated PSPS events. Santa Monica does not qualify.
• Non-Residential Equity ($850/kWh) requires a disadvantaged/low-income community location or eligible public agency status.

Assessment (moderate confidence): TRC should not build SGIP into the financial case. Verify directly with SCE before assuming otherwise. The 30% federal ITC for standalone commercial storage is the more reliable incentive and should be confirmed with TRC's accountant.

Verdict: CANCEL, and do not replace it with another battery yet

Confidence: high on cancelling Kora. High on the sequencing recommendation. Moderate on the specific replacement product.

On Kora specifically. It is a residential, split-phase, whole-home panel-replacement product from a 9-person, $5.3M seed-stage company that has never delivered a unit, whose website displays three contradictory ship dates at once, whose certification claims contradict its own social posts and its own August email, and which has no documented demand-limit control, the one feature TRC actually needs. Even in the good scenario where Kora succeeds, it is very likely the wrong shape of product for a commercial building. TRC reserved a home battery to solve a commercial demand-charge problem.

But the deposit is not the decision. $100 refundable is noise. The real cost of the Kora reservation has been six months of it occupying the slot marked "our plan for the electrical problem" while the actual problem went unmeasured. That is what to cancel.

Do this, in order

1. Email Kora today requesting the UL file number and issuing NRTL for the Powerblocks ESS, and the CEC equipment-list entry for the inverter, plus written confirmation of whether any product supports 120/208 V three-phase and whether any product exposes a kW demand setpoint. Send to Greg Connolly, not to Paige Tranchina. Cancel the reservation and take the $100 back regardless of the reply, the reservation can be re-placed later if the answers are good, and the "manufacturing has begun" clause only gets harder to exit.
2. Determine TRC's service configuration. Read the meter and the main panel: single-phase 120/240, or three-phase 120/208. This one fact eliminates roughly half the product options and must precede any purchase.
3. Get the data before buying anything. Download 12 months of SCE 15-minute interval data for account [account], and install the electrical tracker Greg wants. Find out whether the 19.5 kW peak is a sustained 15-minute plateau (battery territory) or something else entirely.
4. Put the $4,500 soft-starter quote on hold pending step 3. Two independent contractors have now said it will not move the bill, and the 15-minute averaging math supports them.
5. If and only if the data confirms a sustained peak that operational changes cannot fix, spec a Sol-Ark system, 15K if the service is split-phase, 30K-3P-208V if three-phase , sized to the measured peak shape, with the grid peak-shaving threshold set below the tariff boundary. Budget $22k-$50k installed depending on phase and battery size, plus an $800 SCE Rule 21 non-export application, and 3-6 months from order to permission to operate. Confirm the federal ITC with TRC's accountant; do not count on SGIP.

Steps 1 through 4 cost almost nothing and can start this week. Step 5 should not start until step 3 is finished.