Asheville’s executed Axon contract has never been released (Promise Tracker row 12). But the contract form is national, and its record is documented. This page collects what the same paper did elsewhere, so every “we don’t know what we signed” has a “but here’s what cities like us got.”
The money: escalation is the business model
- Baltimore: $11.7M over 6 years (2016) → $35M (“price triples,” 2020) → a 10-year, $153.2M sole-source deal (May 2026): $10M/year rising toward $16M/year with AI services. Motorola formally warned the city it was overpaying by at least $50M by skipping competitive bidding. The council president voted no; it passed anyway. (Baltimore Banner · Baltimore Brew)
- Mesa, AZ: per-officer cost 2.4× in six years; the current contract allows annual costs to rise from $2.1M to $2.5M with no further council approval. (Mesa Tribune)
- Reading, PA: a surprise $525,000 mid-year budget transfer to cover the Axon renewal; police training postponed to pay for it. The city’s own finance director: “With Axon they lure you in with the initial contract and sweeteners, and then on the back end, the data storage is where they really make their money.” (via Yahoo News)
- The antitrust case: after Axon bought its only real competitor (Vievu, 2018), its average bodycam price went $255 → $490. Baltimore and other cities are suing; a federal judge refused to dismiss the damages claims in January 2025. (Cohen Milstein, In re Axon VieVu Antitrust Litigation)
- Axon’s own investor pitch: net revenue retention 124–126%: existing customers pay ~25% more each year; large customers upgrade at 140–300%. Local confirmation from our own record: a resident documented Asheville’s 2020 bodycam contract jumping 86%, with $345K billed “in error.” (Axon Q2 2026 8-K · 2026-05-12 City Council)
The lock-in: documented, not theoretical
- Fontana, CA: Axon’s rep told the city its only exit was non-appropriation and that leaving “could tarnish the city’s credit rating.” The contract contained a termination-for-convenience clause the whole time. The city also kept paying for Evidence.com storage it wasn’t using. (MuckRock) Proof both that Axon steers cities away from their own exit rights, and that convenience termination has existed in Axon paper. Demand it.
- Manchester, NH: paid $300,000 for nothing (eating the remainder of its contract) just to escape to another vendor.
- Evanston, IL: told to sign its 7-year, $5.8M renewal by December 12 or pay 8% more (~$470K), a deadline discount used to rush the decision. The “bodycam renewal” quietly grew to include fleet cameras, drones, and a real-time crime center. (Daily Northwestern)
- And it’s recruiting here now: Brevard is currently weighing an Axon offer of three “free” LPR cameras bundled into an ~$80K five-year package (the same pattern: free hardware first, escalating subscription after), one county over. (Transylvania County)
The oversight record
- Denver got its Axon contract five days before the committee vote; its own Surveillance Task Force got it after. Denver still negotiated a 1-year term, 50-camera cap, 21-day retention, no vendor national database, and per-query audit trails: every protection was winnable; Asheville asked for none. (ACLU-CO)
- Draft One, Axon’s AI report-writer (in the ecosystem Asheville joined): EFF found it’s built to delete the AI’s draft so no one can audit machine vs. officer; police chiefs couldn’t tell which of their own reports were AI-written; the King County prosecutor refuses AI-drafted reports; a randomized trial found no time savings at all. (EFF · KOMO · J. Experimental Criminology)
- Facial recognition: Axon’s 2019 pledge not to put it on bodycams is dead: a live watch-list pilot is running on the same hardware Asheville bought, with the former ethics-board chair objecting. The board itself resigned en masse in 2022; Axon replaced it with an in-house council whose reports aren’t public, then bought a military drone maker. (EFF · Policing Project · The Markup)
- When its own hometown pushed back: 26,000 Scottsdale residents petitioned a referendum on Axon’s HQ project; Axon had the Arizona legislature nullify the local vote. That is this vendor’s documented posture toward local democratic oversight. (AZ Mirror)
The data program you’re enrolled in by default
Buried in the Master Services and Purchasing Agreement (the same national baseline paper Asheville signed onto) is the Axon Customer Experience Improvement Program (ACEIP), and its first structural fact is the one that matters: “By default, Customer will be a participant in ACEIP Tier 1.” Not opt-in. Enrolled unless the agency affirmatively revokes.
What Tier 1 grants, per the MSA’s own text (v24, August 2025): Axon “may make limited use of Customer Content from all of its customers to provide, develop, improve, and support current and future Axon products.” The privacy protection is de-identification by “commercially reasonable efforts” (a business-diligence standard, not a guarantee), after which the extracted content “may then be further modified, analyzed, and used to create derivative works,” while remaining “linked indirectly, with an attribution, to the Customer.” Tier 2, a checkbox, opens custom agreements for uses like AI model training.
In plain terms: the footage and data Asheville’s taxpayers fund (bodycams, in-car video, the RTIC’s feeds) is, by contract default, raw material for Axon’s future products. The city can revoke Tier 1 at any time in writing. So the questions for APD are a records request away, and they’re yes-or-no: Has Asheville revoked ACEIP Tier 1? Is the Tier 2 box checked? (Documents request list · primary text: Axon MSPA v24, ACEIP Appendix, same baseline as the public Durham copy)
The rider checklist: all proven gettable
Term cap (Denver: 1 year) · device cap (Denver: 50) · retention cap (Denver: 21 days) · no vendor national database (Denver) · per-query audit trails (Denver) · reworked data-ownership terms (Syracuse) · termination for convenience (it was in Fontana’s contract) · AI-report disclosure and draft preservation (now state law in Utah and California; an ordinance can do the same here) · no auto-renewal (a one-line strike). Asheville’s Sourcewell baseline has none of these. Every future appropriation vote is a chance to demand them.
Twelve talking points, sourced
Ready for the podium. Full citations above; the frame is always: Asheville signed this sight-unseen; here’s what the same paper did elsewhere.
- Baltimore’s Axon bill tripled once, then went from $5.9M to a path toward $16M a year, under a 10-year AI bundle like ours.
- Motorola told Baltimore in writing that skipping competitive bidding cost at least $50M. Asheville used the same no-bid shortcut.
- Axon tells Wall Street its customers pay ~25% more every year. Cost growth isn’t a risk in this contract; it’s the product.
- Mesa’s costs rise without further council approval; Reading raided its general fund mid-year, and its finance director called the model what it is: lure, then bill.
- Federal courts are hearing cities’ claims that Axon tripled prices after buying its only competitor; the judge refused to dismiss.
- Fontana was told it had no exit clause. It did. If our council hasn’t read our contract, neither had theirs.
- Denver’s council got the contract 5 days out and still won a 1-year term, 21-day retention, and mandatory audits. Asheville got none of it, and our contract’s public release was promised and never happened.
- The King County prosecutor won’t accept AI-written police reports; our bundle pre-authorizes the tool with no rules at all.
- Draft One deletes its own drafts so nobody can audit it. California now requires preservation by law. Asheville could too, by ordinance.
- Axon’s ethics board quit over Taser drones; its replacement reports to an Axon executive; facial recognition is being piloted on our exact hardware while our paperwork never mentions the words.
- Axon supplies ICE and DHS, and is negotiating a ~$220M ICE deal; a 7.5-year contract welds Asheville’s data infrastructure to that vendor through 2034.
- Evanston was told: sign in 72 hours or pay $470K more. A deadline discount is a sales tactic, not a reason to sign a 7.5-year contract nobody has read.
Related: Getting Flock Out · Documents · Their Claims vs The Record · Follow the Money