Skip to content
YEGFacts.ca
On this page

Something wrong on this page?

Corrections are published, never quietly edited.

Fact check ·

What happened with Edmonton's electric buses?

What happened with Edmonton's electric buses?

Edmonton's electric buses fell well short of the contract, and the $82 million is a claim the City filed in court, not a loss it has shown.

TL;DR

  • Edmonton told the bankruptcy court its electric buses managed about 165 km in winter against the 268 km the contract promised.
  • At times more than half of Edmonton's 60 electric buses were off the road, by the City's own account to the court.
  • Proterra went bankrupt, leaving Edmonton without reliable parts or a warranty it could count on.
  • The $82 million is what Edmonton asked for in Proterra's bankruptcy, and what it will get back is still not public.
  • Toronto ran electric buses from three makers through its winters and kept some of them on the road 95 per cent of the time.

Verdicts by a Three AI models from three vendors (Claude, GPT, Gemini) researched this claim in a blind first round, then read one another’s findings in a second round that documented errors. This label describes a run frozen before September 23, 2026.More in the methodology under a A fixed synthesis rule, published in advance, turns the three verdicts into one finding word, so no person chooses the finding. The rule and every change to it are versioned in the methodology changelog.More in the methodology · The date the accounting window closes. Anything that happened after it is outside this check, and the date is fixed in the brief before any model runs.More in the methodology · The date we last re-read the cited sources and confirmed the page still matches them. It is not a claim that the records themselves changed.More in the methodology · When this is due for a fresh check of its sources. Past that date, treat it as unverified until it has been re-reviewed.More in the methodology · The version of the published method that produced these findings. Every change to the prompts, the merge and synthesis rules, the vocabulary or the validation bumps it.See what changed

Common forms of the claim

Composite paraphrases written by us to show the shapes the claim takes; not captured posts, not quotations, not a measure of how common each is.

  • Paraphrase — not a real post

    The city blew $82 million on electric buses that don't even work in winter. Total fiasco.

  • Paraphrase — not a real post

    Didn't Edmonton lose like $82M on those Proterra buses? And they knew the company was failing?

  • Paraphrase — not a real post

    Edmonton's e-bus disaster proves EVs can't handle real winters. Stop buying them.

What actually happened

Edmonton bought 60 electric buses, they fell short, the maker went bankrupt, and a number from the bankruptcy court became the number everyone repeats. This page follows the buses and the money.

The order and the 350 km announcement

Edmonton made its battery-electric bet in the late 2010s. In 2018, the federal, provincial and municipal governments announced joint funding of more than C$43 million for the first tranche. The City ordered 40 buses from Proterra Inc. — the U.S. manufacturer that built Edmonton’s 60 battery-electric buses. It filed for bankruptcy in August 2023., a U.S. electric-bus manufacturer, later expanded the order to 60, and publicly announced them with up to 350 km of range on ETS routes when the first entered service in August 2020.

Breakdowns, then Proterra’s bankruptcy

The fleet’s troubles became public in stages. Transit-union representatives described mechanical issues, battery problems and missing parts. In August 2023, Proterra filed for A U.S. bankruptcy process in which a company reorganizes under court supervision while continuing to operate. bankruptcy in the United States. Edmonton’s contract was not among those transferred to the buyer of Proterra’s transit business, which left the City expecting its warranty provisions would not be honoured and without a reliable parts channel for a fleet it had paid US$58,761,600 for.

What the City told the bankruptcy court

In October 2023 the City went to the bankruptcy court with an itemized filing: it alleged substantially shorter range than contracted, and put its actual costs to that point at C$1,352,655.58 — repair labour, parts, battery blankets and an outstanding receivable. The specific numbers came in February 2024, when the City filed a A formal document a creditor files in a bankruptcy stating how much it believes it is owed. It is an assertion, not a court ruling or an audited amount. seeking more than $82 million for breach of contract and negligence: the buses were contracted to deliver 328 km of range in normal conditions and 268 km in extreme cold, but managed roughly 165 km on average in winter — 38% short of the cold-weather guarantee — and at best about 250 km in warmer weather, short of even the 350 km the City had announced at launch. That proof of claim matches the “$82 million” number that circulates on social media.

Range per charge: promised against observed
Announced at launch, 2020350 kmContracted, normal conditions328 kmContracted, extreme cold268 kmObserved, best warm weather250 kmObserved, winter average165 km

Announced range from the 2020 launch announcement; contracted and observed ranges as stated in the City's own bankruptcy filings. Grey bars are figures the City was promised or contracted for, dark bars what it reported measuring. YF-EV-0012, YF-EV-0003

What other cold cities found

Edmonton’s fleet is one manufacturer’s buses in one city, and the record from elsewhere does not show the technology failing in the cold. Toronto’s transit agency tested 60 buses from three manufacturers side by side, and its evaluation found winter range loss for all three, from 13 to 30 per cent depending on the manufacturer, while some fleets were still available 95 per cent of the time. A later update from the same agency put battery-electric propulsion at under 4 per cent of recorded defects and identified no fundamental technology barrier. Winnipeg, with winters as severe as Edmonton’s, ran its own cold-climate evaluation and recommended going ahead with a test fleet. Edmonton’s trouble sat with one supplier whose bankruptcy cut off its parts, which is a vendor failure and not by itself a result for the technology.

Where the claim and the fleet stand now

Later developments bear directly on the money question.

  • January 2024 — Proterra’s transit business was sold to Phoenix Motor, which said it planned to restock spare parts.
  • March 2024 — Proterra’s reorganization plan took effect.
  • May 2026 — the claims register still lists Edmonton’s claim as A bankruptcy claim with no collateral behind it. General unsecured creditors are paid after secured creditors, often only partially.; no allowed or paid amount has been published.

The City’s own project page still lists all 60 electric buses in the ETS fleet, operating from two garages.

Claims checked

Did Edmonton lose $82 million on its electric buses?

The public record can’t back this up — not the same as proven false.More in the methodology· All three seats reached the same verdict. Panel agreement describes seats, not vendors; two OpenAI seats agreeing are not two independent vendor confirmations.More in the methodology panel

What kind of record answered the question — an audited statement, a council report, a dataset. It is reported apart from the finding, because a claim can be Supported on a thin basis.More in the methodology: Direct Edmonton evidence

  • The $82 million figure is a proof of claim the City filed on February 2, 2024 in Proterra's Chapter 11 bankruptcy, seeking damages for breach of contract and negligence. A damages claim is not an audited loss. YF-EV-0003, YF-EV-0002
  • The City's October 2023 court filing says it paid US$58,761,600 under the 60-bus contract, and itemized C$1,352,655.58 in repair labour, parts, battery blankets and an outstanding receivable at that time. YF-EV-0001
  • A May 2026 claims-register entry lists Edmonton's claim as general unsecured, with no allowed or paid amount shown — how much Edmonton will ever recover is not public. YF-EV-0002
  • Governments announced more than C$43 million in joint funding for the original project: more than C$21.5M federal, about C$10.8M provincial and more than C$10.8M municipal — the payer mix matters to any 'Edmonton lost' framing. YF-EV-0010
1 more fact
  • The buses entered service in 2020, were reported still in daily service in early 2024, and the City still lists all 60 in the ETS fleet — service was delivered and residual value exists, neither of which the loss framing counts. YF-EV-0005, YF-EV-0004, YF-EV-0003

Limitations

  • Not established does not mean nothing was lost. Real public costs are on the record and they are not small. It means no public record found so far shows a net loss of about $82 million.
Not established

Established background

Did Edmonton's Proterra electric buses fall well short of what the City's contract called for?

The evidence affirmatively establishes it.More in the methodology

Yes. Edmonton told the bankruptcy court that its Proterra electric buses averaged about 165 kilometres in winter even though the contract guaranteed 268 in extreme cold.

Nobody is arguing about this one, so the site does not carry it as a finding of its own. It is the ground the claim above stands on, and it is what makes that claim worth asking.

Do Edmonton's problems with Proterra buses prove that electric buses don't work in cold cities?

The evidence affirmatively points against it.More in the methodology· All three seats reached the same verdict. Panel agreement describes seats, not vendors; two OpenAI seats agreeing are not two independent vendor confirmations.More in the methodology panel

What kind of record answered the question — an audited statement, a council report, a dataset. It is reported apart from the finding, because a claim can be Supported on a thin basis.More in the methodology: Edmonton + comparable cities

  • Toronto tested 60 buses from three manufacturers head-to-head. Winter range loss was real for all three, but outcomes diverged sharply by manufacturer — New Flyer reached a 70,000 km mean distance between failures against a 30,000 km target, with 95% fleet availability. YF-EV-0006
  • A later TTC program update reported battery-electric propulsion accounted for less than 4% of recorded defects and identified no fundamental technology barrier. YF-EV-0007
  • Winnipeg — a city with winters as severe as Edmonton's — ran a 213-page cold-climate technical evaluation and recommended proceeding to a zero-emission test fleet. YF-EV-0008
  • Edmonton's own problems were concentrated in one manufacturer whose bankruptcy cut off its parts channel and left warranty coverage unlikely to be honoured — a vendor failure, not by itself a category-wide result. YF-EV-0004, YF-EV-0003

Limitations

  • Cold-weather range reduction is real for every manufacturer — in Toronto's data, 13% to 30% less winter range depending on vendor; the contradicted proposition is 'unsuitable for cold cities,' not 'unaffected by cold.'
  • Toronto's winters are milder than Edmonton's; transferability of exact figures is moderate.
Contradicted

What the City promised

  • Edmonton's new Proterra electric buses will deliver up to 350 km of range on ETS routes.

    City of Edmonton / Edmonton Transit Service (as publicly announced at launch) · · Assessed · YF-EV-0012

Edmonton evidence

Comparable cities

  • Toronto

    Controlled three-manufacturer comparison in one system: same conditions, sharply different outcomes — evidence that manufacturer and support, not propulsion, drove the differences. Toronto's own later struggles (second-generation dispatch ranges, availability dips) also caution against overclaiming success.

    Transfers to Edmonton? Moderate for exact figures (milder winter); high for the manufacturer-variation point.

  • Winnipeg

    A prairie city as cold as Edmonton evaluated zero-emission options in detail and proceeded to procurement; its modelling put long-range depot-charged buses at about 75% coverage of conventional runs under maximum heating demand.

    Transfers to Edmonton? High for climate; procurement scale still small.

What remains unknown

  • Independent (non-litigation) measurements of fleet availability and range by year.
  • How much performance improved or degraded after Phoenix Motor reopened a parts channel in 2024.
  • Edmonton's actual recovery, if any, from the Proterra bankruptcy estate.
  • Lifecycle cost of the Proterra fleet vs a diesel counterfactual (maintenance, energy, availability, early replacement).
  • Whether any battery-electric fleet has yet matched diesel lifecycle costs in a city as cold as Edmonton.

Missing evidence

Records we asked for and have not received, or that do not appear to exist publicly.

  • ETS per-year fleet availability, maintenance cost and range data for the Proterra fleet (holder: Edmonton Transit Service).
  • Bankruptcy distribution records for claim 1281 (holder: City of Edmonton / claims agent).
  • ETS lifecycle cost data sufficient to compute net public cost (holder: Edmonton Transit Service).

AI review

Earlier findings used three AI models from three vendors (Claude, GPT, Gemini). The canonical finding is computed from the blind round-one verdicts by a deterministic rule published in advance; the models never vote and never decide. Cross-review documents their errors, and what it changed is shown below rather than folded into the finding. Authority comes from the methodology, not from the models.

Each reviewer's verdict and confidence are from round one, the round in which the reviewers worked independently.

ClaimClaude Fable 5GPT-5.6 SolGemini 3.1 ProPanel finding
Did Edmonton's Proterra electric buses fall well short of what the City's contract called for?SupportedModerateSupportedModerateSupportedModerateSupportedUnanimous panel
Did Edmonton lose $82 million on its electric buses?Not establishedHighNot establishedHighNot establishedHighNot establishedUnanimous panel
Do Edmonton's problems with Proterra buses prove that electric buses don't work in cold cities?ContradictedHighContradictedHighContradictedHighContradictedUnanimous panel
Reviewer notes — Did Edmonton's Proterra electric buses fall well short of what the City's contract called for?

Claude Fable 5

  • Every element of the normalized proposition — range, reliability, durability, vendor support — is established; under the v1.4 rule the litigation-sourced precision of the figures is a confidence consideration, not a partial-support one.

After cross-review: Re-decided 2026-09-01 under the tightened v1.4 verdict rule, evidence-locked to the original archived set. Original panel: Adjacent (2 Partially supported, 1 Supported).

GPT-5.6 Sol

  • Material shortfalls established across the proposition: 165 km winter vs 268 km contracted, 250 km best vs 328 km, recurring outages and hardware failures, and the loss of vendor support. Confidence Moderate because the sharpest figures are City court filings.

After cross-review: Moved from Partially supported to Supported once the tightened rule required naming a failing part of the proposition — none could be named.

Gemini 3.1 Pro

  • Shortfalls established in all specified areas; continued daily service does not negate failure to meet contracted performance. Evidence quality (untested allegations) is carried in confidence.

After cross-review: Same move as the GPT seat, independently: with no failing part identifiable, the verdict is Supported and the evidence-quality doubt lives in Moderate confidence.

Reviewer notes — Did Edmonton lose $82 million on its electric buses?

Claude Fable 5

  • The $82M is a legal damages claim including negligence; treating it as cash lost conflates a litigation position with an accounting result.

GPT-5.6 Sol

  • Purchase price was US$58.76M — the $82M headline exceeds even the gross purchase price in USD terms; the loss framing counts no recoveries, grants, service delivered or residual value.

Gemini 3.1 Pro

  • As an unsecured creditor Edmonton was expected to recover little, but the final net loss cannot be computed from public records.
Reviewer notes — Do Edmonton's problems with Proterra buses prove that electric buses don't work in cold cities?

Claude Fable 5

  • One manufacturer's failure in one city, during that manufacturer's bankruptcy, cannot carry a category-wide generalization that measured multi-manufacturer data contradicts.

GPT-5.6 Sol

  • Same-market contrast: St. Albert's different-manufacturer struggles and Toronto's manufacturer spread show implementation variance, not uniform category failure.

Gemini 3.1 Pro

  • Cold degrades range for every manufacturer, but degraded is not unusable; several cold-climate systems operate battery-electric fleets in regular service.
reviews/electric-buses/2026-09-01-rerun
reviews/electric-buses/2026-08-31

That directory also holds the raw round-one and round-two JSON from every reviewer, the merged evidence and the synthesis output.

Reasoning effort is not recorded in this run’s manifest. Before methodology v1.6 only the GPT and Gemini commands pinned it; the Claude seat used an unpublished local default.

How it was registered

Registered

How far the work has got
Gate complete
Whether it is worth checking
Going ahead
What a reader can see
Published

What the states mean

Article history

  • updatedPresentation, under the layers rule in docs/DESIGN.md section 12 (methodology v1.32): the standfirst is now one sentence and the TL;DR five plain sentences with one fact each. The 38 per cent shortfall, the February 2024 proof of claim, the exact purchase price, the $43 million in joint funding and Toronto's 13 to 30 per cent range losses all moved into the explanation. The Edmonton figures already sat there beside their sources. The Toronto and Winnipeg evidence behind the cold-cities claim had only ever been in the TL;DR and in the claim detail, so the explanation gained a short section, What other cold cities found, carrying it with its three sources. No finding, figure, hedge or source changed. A plain-speech read from another vendor covered the standfirst and every bullet before merge; its report and the editor's dispositions are at reviews/2026-09-16-readable-layers/plain-speech/gpt-1.md. After the critique, a two-sentence opening sits before the first subheading, so the explanation no longer starts with two headings back to back. The independent critique of the rendered page and the editor's dispositions are at reviews/2026-09-16-readable-layers/critique-1.md. Two further outside comparisons of the old and new pages, run by the founder, are at reviews/2026-09-16-readable-layers/compare-1.md and compare-2.md with dispositions; what they changed on this page: the standfirst no longer says the $82 million is not money the City has lost, which was more than Not established allows, and says it is a claim filed in court, not a loss shown; the two performance bullets say the City told the court, which is the attribution the old TL;DR carried; the $82 million is back in the TL;DR in place of the purchase price, which is in the explanation; and the new cold-cities section links its three sources as words rather than as archive numbers.
  • updatedHeld afterwards against the court filings and City pages the page rests on, the new wording came back with two problems. One is "one failed purchase from one supplier": the City's filings establish that the buses fell well short of the contract, and the City's own page records that they went on carrying passengers, so calling the purchase itself failed says more than the record does. The standfirst now says one supplier's buses falling short here. The sentence under the $82 million claim gained a bound as well, because no public record showing a loss that size has been found, which is not the same as none existing. No finding, figure or hedge moved.
  • updatedA reading seat at another vendor worked through the new wording, and three of its rewrites landed here. The heading asks what happened with the buses instead of what went wrong, since one claim on this page is Not established and another is Contradicted, and a heading has no business settling either before a reader gets there. "One bad contract" reads as a contract with bad terms, so the standfirst now says one failed purchase from one supplier. The cold-cities question names the Proterra buses, which gives somebody arriving cold the premise it runs on. Findings, figures, hedges and sources are untouched.
  • updatedPresentation, under the plain-speech standard in docs/DESIGN.md section 12: the title now asks the question a person actually asks, and the standfirst, each claim's question and each claim's answer were rewritten into the words people use. No finding, figure, hedge or source changed. Every string that changed is listed beside its old wording in docs/plain-speech-rewrite-2026-09-03.md. A different-vendor plain-speech read went over the new wording the next day, and every claim on this page names it.
  • updatedPresentation under methodology v1.19: the procurement claim now renders as established background beneath the $82 million claim it is the premise of. The two keep their separate findings, pages and evidence; nothing was merged and no verdict moved. The $82 million claim stays on the board, and it reopens when the Proterra bankruptcy estate reports what Edmonton actually recovers.
  • updatedPresentation (methodology v1.18): each claim now carries its own one-sentence answer, in the words a person would say, and the standfirst was rewritten to hold one idea per sentence. Every fact the old standfirst carried is still on the page, in the TL;DR and the body. No change to findings, facts or evidence.
  • updatedPresentation: one-line answer shortened to a single sentence; body given subheadings for scanning. No change to findings, facts or evidence.
  • verdict-changeThe procurement-failure claim moved from Partially supported (Adjacent panel) to Supported (Unanimous, Moderate confidence) after a re-decision under the methodology v1.4 verdict rule: evidence-quality doubts now affect confidence, not the finding, and no reviewer could name a part of the proposition that fails. Evidence set unchanged; re-run artifacts in reviews/electric-buses/2026-09-01-rerun/.
  • correctionThe 350 km commitment's measurable form wrongly added 'including Edmonton winter conditions' — the launch announcement did not specify winter range. Corrected; the alleged contractual cold-weather specification is tracked in the claims. Source date for the launch article corrected to 2020-07-24.
  • publishedInitial publication. Founder delegated the publication gate to the AI audit for this story (source verification: 22 statements checked against archived bytes, 0 unsupported, 7 imprecisions fixed; privacy/release check: clean).

Every revision is in the public commit history of this file.