Watt Guard
WattGuard
Decode the AI-driven spike on your power bill, find the 15–30% you can recover, and turn it into action — supplier shopping, time-of-use shifts, and a one-click rate-case comment.
Date: 2026-05-08 Form factor: Web app (single-page; mobile-friendly) Status: Prototype
What it is
WattGuard is a single-page workspace for households whose electric bills have spiked because of hyperscale data-center demand. It decomposes the user's current bill into its real cost components, calls out the share that is data-center-attributable, and walks the user through the three real consumer levers: switching supplier in retail-choice states, shifting flexible loads onto a TOU rate, and filing a public-comment letter on an active state-PUC rate case. The prototype demonstrates the end-to-end flow against three sample households (Loudoun County VA on Dominion, Houston TX on a competitive REP, and Pittsburgh PA on Duquesne Light's default Price-to-Compare).
Who it serves
Owner-occupants and renters in the PJM, ERCOT, and CAISO footprints — the territories that have absorbed the largest data-center-driven cost increases since 2023. Specifically:
- Households in the Mid-Atlantic seeing PJM capacity-charge pass-throughs (Virginia, Maryland, NJ, OH, PA, IL, IN, KY, NC, WV, DC).
- Texas households on default or expiring REP plans who haven't shopped PowerToChoose in 12+ months.
- Pennsylvania, Ohio, and Illinois households still sitting on the utility default Price-to-Compare instead of a fixed 12-month plan.
- People who saw a $50–$200/month winter-bill spike and don't know whether to blame the weather, the wholesale market, or themselves.
Why it could be profitable
Monetization model:
- Free decoder + savings estimate (acquisition).
- $5/mo Watchdog subscription — monthly bill audits with anomaly alerts and rate-shop reminders 30 days before contract expiry.
- $25 one-time Switch Concierge — a guided supplier-switch flow, including the broker handoff in TX, PA, OH, and IL (we collect a $40–$80 acquisition fee from the supplier per switched household).
- $1.50/comment Civic Voice — pre-filled, evidence-backed PUC rate-case comments delivered via the right portal, with proof of filing.
- B2B/white-label: utility consumer-advocate offices, state attorneys general, and nonprofits (Citizens Utility Board, AARP, Sierra Club) can license the decoder + comment generator.
Demand rationale (why now):
- The PJM 2026/2027 capacity auction cleared at the FERC-approved cap of $329.17/MW-day — about a 10x jump from two auctions prior, with data-center demand identified as the dominant driver. (Enel North America, Citizens Utility Board, 2026-01-30)
- US residential electricity has risen ~36% from 2020 to Feb 2026 (12.76¢/kWh → 17.44¢/kWh), and a January 2026 Bloom Energy report projects US data-center demand to nearly double from 80 GW to 150 GW by 2028. (CNBC, 2026-02-12, Consumer Reports)
- Areas with dense data-center load saw electricity prices jump 267% over five years. One Virginia resident's bill went from ~$100 in December 2025 to $281 in January 2026. (Yale Climate Connections, 2026-01)
- Switching suppliers in deregulated markets routinely saves 15–30%, but participation lags the opportunity — Texas leads at 87% switching, but PA sits at ~35% and Ohio at ~57%. (Electric Choice)
- The political environment is finally moving — 238 data-center-related bills were introduced in 50 states in 2025 with 40+ enacted in 21 states; Oregon became the first state to create a dedicated data-center electricity rate class. The case for ratepayer-side advocacy has never been more concrete. (Introl, 2026)
- FERC has explicitly directed PJM to write new rules for large-load interconnection while protecting consumers — meaning rate-case comments now have a real federal hook. (FERC)
A consumer tool that turns an opaque, one-page utility bill into a clear "you can recover $X by doing Y" workflow has obvious price-disruption headroom against the existing options (call your utility, hire an energy broker, or do nothing).
Form factor & scope
A single-page web app, mobile-friendly. The prototype demonstrates a 4-panel audit flow against three sample households. A production version would integrate Green Button Connect (utility data download), live retail-supplier feeds (PowerToChoose, PA Power Switch, Apples-to-Apples Ohio), state PUC docket APIs, and a managed comment-filing pipeline.
Minimum viable scope demonstrated: household selection → bill-component decoder with data-center attribution → supplier shop with best-fit highlighting → load-shifter slider (live monthly + annual savings) → rate-case voice with pre-filled comment letter, copy + download.
How to run it
- Open
index.htmlin any modern browser (Chrome, Edge, Firefox, Safari). - The first sample household (Loudoun County, VA) loads automatically. Click another household in the left rail to see the same flow against an ERCOT household with retail choice or a PA household on the utility default.
- Click any line item in the Bill decoder for a plain-English explainer of what the charge pays for and why it is moving.
- In Supplier shop, the row tagged
BEST FITis the highest-savings plan; awarn-styled row flags expensive default/variable plans to avoid. - Drag the Load shifter slider to see live monthly and annual savings change as you simulate moving more load to off-peak hours. The topbar's annual-savings pill updates as you do.
- In Rate-case voice, edit the pre-filled letter, then Copy letter or Download .txt to file with the appropriate state PUC.
Note:
sample-data.jsonis also embedded inline as a<script type="application/json">block insideindex.html, so the prototype runs directly fromfile://without needing a local web server.
What's in this prototype
- Sticky topbar with a live PJM auction-clearing pill (
$329.17/MW-day) and a running "Est. annual savings opportunity" pill that updates as you change households or move the load slider. - Household rail: three sample profiles with distinct ISO/regulatory contexts (regulated PJM-VA, deregulated ERCOT-TX, deregulated PJM-PA), each with their own usage profile and persona.
- Persona card: current month, current bill, average monthly kWh, and a 12-month YoY trend tinted red (up) or green (down).
- Step 1 — Bill decoder: each line item rendered as a stacked bar (baseline vs. data-center-attributable share). Click a row to see the explainer; the KPIs show the data-center share of the whole bill in both percentage and dollars.
- Step 2 — Supplier shop: alternative-plan table for retail-choice states, with
BEST FITandwarnrow styling. For regulated states, this panel honestly tells the user there is no shopping lever and routes them to the other panels. - Step 3 — Load shifter: an interactive slider (range tied to ~50% of the household's monthly kWh) with peak/off-peak rate cards and live monthly + annual savings.
- Step 4 — Rate-case voice: docket metadata (ID, title, deadline with days-left, filing portal URL), bullet list of evidence-based talking points, and a pre-filled letter that pulls in the household's actual data-center-attributable dollar amount. Copy-to-clipboard and
.txtdownload both work.
Roadmap
- Green Button Connect ingest: pull the user's actual 13-month kWh + bill PDF, decode each line item against utility tariff schedules so no manual entry is needed.
- Live supplier feeds: ingest PowerToChoose (TX), PA Power Switch (PA), and Apples-to-Apples (OH) so plans are real-time, with broker-handoff revenue share.
- Auto-rate-shop on contract expiry: detect the 30-day-out window in the user's bill data and re-shop automatically.
- Filed-comment service: file the generated rate-case comment with the correct state PUC portal and return a stamped receipt.
- Multi-household / portfolio mode: for landlords, property managers, and small-business owners with multiple meters across a state.
- TOU enrollment concierge: walk the user through opting into the utility's existing TOU rate (Dominion's TOU-D, ComEd's Hourly Pricing, etc.) and verify enrollment.
- Notification stack: push alerts when a new rate case opens in the user's territory, when the PJM auction clears, when a default Price-to-Compare changes.
Sources
- PJM 2026/2027 Capacity Auction Results (Enel North America) — supports the $329.17/MW-day clearing price and 22% YoY hike.
- PJM Rate Shock: $100B Data Centers vs Ratepayers (Introl, 2026) — supports the 238 state bills, Oregon dedicated rate class, and political-backlash framing.
- The devil is in the details: PJM data-center policy analysis (Citizens Utility Board, 2026-01-30) — supports the 63% data-center share of the auction increase and the $9.3B added capacity cost figure.
- Home electricity bills are skyrocketing — for data centers, not so much (Yale Climate Connections, 2026-01) — supports the 267% five-year price jump in data-center-dense regions.
- AI Data Centers: Big Tech's Impact on Electric Bills (Consumer Reports) — supports the $100→$281 Virginia residential bill anecdote.
- Electricity prices rising at 2x inflation; data center demand means no relief (CNBC, 2026-02-12) — supports the 36% residential price increase from 2020 to Feb 2026.
- Electricity Rates by State, May 2026 (Electric Choice) — supports state-by-state rate averages, switching rates (TX 87%, OH 57%, PA 35%), and the 6.69¢/kWh PA fixed-rate floor.
- FERC directs PJM to create new large-load rules (FERC press release) — supports the federal regulatory hook for state-level rate-case comments.
Requirements
WattGuard — Requirements
Goals
- Make a household's monthly electric bill legible — show the user, in plain English and visually, what each line item pays for and how much of it is being driven by data-center load growth.
- Quantify the recoverable portion of the bill across three real consumer levers (supplier switch, load shift, rate-case advocacy) and surface a single annual-savings number.
- Convert "I'm angry about my bill" into a concrete next action within one session — a switch decision, a TOU enrollment, or a filed comment.
- Honestly distinguish between regulated and deregulated markets; never offer fake levers that the user's state does not allow.
- Ground every claim in cited public-record evidence (PJM auction filings, FERC orders, state PUC dockets, federal price stats).
Primary user
A US residential ratepayer in PJM, ERCOT, or CAISO who has just received a 20%+ higher bill than the same month last year. They have read a news headline about AI data centers and electricity costs, suspect they are overpaying, and have ~30 minutes of patience to do something about it before they close the tab. They are not a power-systems expert; they own a phone or a laptop and a recent bill PDF.
Functional requirements
- FR1: Render a left-rail picker of sample households spanning regulated PJM, deregulated ERCOT, and deregulated PJM. Switching households re-renders all four panels with that household's data.
- FR2: Render a persona card with current month, current bill amount, average monthly kWh, and a 12-month YoY trend tinted red (up) or green (down).
- FR3: Bill decoder — render every component of the user's current bill as a stacked horizontal bar (baseline vs. data-center-attributable share), labeled with the dollar amount and the data-center percentage.
- FR4: Bill decoder — clicking a component shows a plain-English explainer of what that line item pays for and why it is moving.
- FR5: Bill decoder — show two KPIs: total data-center-attributable share of the bill (percentage) and total data-center-attributable dollars per month.
- FR6: Supplier shop — for deregulated households, render an alternative-suppliers table (rate, type, term, ETF, monthly savings) with the highest-savings plan tagged
BEST FITand any negative-savings plan styled as a warning. - FR7: Supplier shop — for regulated households, suppress the table and explicitly tell the user there is no shopping lever in their state, routing them to load shifting + rate-case comments.
- FR8: Load shifter — provide a slider scaled to ~50% of the household's monthly kWh, plus peak/off-peak rate cards. Update the live monthly + annual savings as the user drags.
- FR9: Rate-case voice — render docket metadata (ID, title, deadline + days-left, filing portal URL) and an evidence-based talking-points list for the household's state.
- FR10: Rate-case voice — pre-fill a comment letter that pulls in the user's current bill amount and computed data-center-attributable dollars; expose Copy and Download .txt actions.
- FR11: A topbar pill must show a running "estimated annual savings opportunity" that combines the best supplier-switch monthly savings with the current load-shift monthly savings, multiplied by 12.
- FR12: Run from
file://with no build step, no remote API calls, and no auth —sample-data.jsonis also embedded inline as a<script type="application/json">block. - FR13: Layout must collapse cleanly to a single column on screens narrower than 1100px so that the prototype is presentable on a phone.
User stories
- As a Loudoun County, VA homeowner, I want to see why my bill jumped from $215 to $281 in one month, so that I can tell whether to call the utility or wait for the season to change.
- As a Houston REP customer, I want to compare my current rate to the best fixed plan available, so that I can decide whether to switch before my contract ends.
- As a Pittsburgh renter on the Price-to-Compare default, I want to see the lowest 12-month fixed rate in PA Power Switch, so that I can stop overpaying every month.
- As any household, I want to drag a slider and see exactly how much I save per month by shifting laundry and dishwasher to off-peak hours.
- As a household angry about data-center pass-through costs, I want a one-click pre-filled comment letter to my state PUC, so that my voice is in the docket without spending an evening on the wording.
- As a user who lives in a regulated state, I want the app to honestly tell me supplier shopping is not available, instead of pretending it is.
- As a user, I want my session-level changes (selected household, slider value, selected line item) to persist across panels within the same session.
- As a non-expert, I want every dollar number explained in one sentence I can paste to a friend.
Non-functional requirements
- No build step — opens directly via
file://from the folder, no Node, no npm. - No external API keys — the prototype uses bundled JSON only.
- No remote scripts — all CSS/JS is local; no CDN dependency that would break offline.
- Accessibility — every interactive control is keyboard reachable; color is not the sole signal (we pair color with text on best-fit and warning rows, and on the up/down YoY tint).
- Privacy — the production version must use Green Button Connect's authorized data flow only; no scraping of utility logins. The prototype handles only synthetic data.
- Citation discipline — every load-bearing dollar/percentage in the README is hyperlinked to a public source; the in-app letter cites federal precedents (FERC large-load order, Oregon rate class).
- Latency — initial render under 200ms on a mid-tier laptop; slider updates at 60fps.
Out of scope (for the prototype)
- Live utility data ingestion (Green Button Connect, OAuth flows).
- Live retail-supplier feeds (PowerToChoose, PA Power Switch, Apples-to-Apples).
- A real comment-filing pipeline to state PUC portals.
- TOU enrollment APIs.
- Solar/storage ROI modeling (acknowledged as out of scope; users with rooftop PV need a different tool).
- Mobile-app form factor (we'll do a PWA wrapper after we validate web).
- Spanish-language and accessibility-AA polish (planned but not in this slice).
Open questions
- What is the right monetization split between consumer subscription ($5/mo Watchdog) and supplier-acquisition fee revenue? The two pull in opposite directions on the recommended-plan ranking.
- Should the supplier-acquisition fee be disclosed line-item next to each plan, or only in a global footer? Disclosure helps trust but may suppress conversion.
- For regulated states, is there a viable "demand-response enrollment" lever we are missing in the prototype? Some utilities pay $50–$200/year for enrollment in conservation programs.
- Do we want to support multifamily / submetered households in v2, where the resident's bill comes from a building owner rather than the utility directly? Different legal regime entirely.
- For rate-case comments, do we batch-submit on behalf of users (with their consent), or hand them off to the portal? Batch submission could push us into unauthorized practice of law in some states.