What happens when a robot company shuts down its servers?

Direct answer: When a maker’s cloud dies, owners usually face one of three outcomes — the robot bricks completely, it keeps moving but loses voice/maps/apps, or a community/self-host rescue keeps a subset of features alive. The outcome is determined by how much of the product’s intelligence lived on those servers, not by the sticker price. Our robot graveyard documents nine sourced shutdown and sunset events; the pattern is consistent enough to buy against.

Most shopping guides stop at “does it work today.” Best Robot Match exists because consumer robots are long-lived hardware glued to short-lived startups. A children’s companion can become a paperweight overnight if conversation AI only ran in the cloud. That is not hypothetical: it is the Moxie case, verified in our evidence ledger and graveyard. The same failure mode hit Jibo in 2019, and Vector has now survived two corporate deaths. Understanding the shutdown sequence is how you avoid joining the next Death Toll counter. This page is not a list of offline robots — that lives at robots that work offline. Here we walk failure modes, sourced cases, and the checks that show which path a given SKU is exposed to before you buy.

Three outcomes when the cloud goes dark

Outcome	What owners experience	Graveyard examples
Full brick	Core intelligence or activation fails; unit is inert or limited to a goodbye sequence	Moxie (2025), Jibo (2019)
Degrade + possible rescue	Motors/sensors may still run; cloud voice, faces, or app control fail; community self-host may restore some features	Anki/DDL Vector eras; Neato cloud sunset (smart features)
Never reached owners / soft kill	Company folds before volume shipping, or a product line is disabled with refunds	Kuri cancelled; Amazon Astro for Business bricked with refunds

What the graveyard cases actually show

Moxie (Embodied, died 2025-01-30): company-shutdown, bricked: true. Conversational AI depended entirely on Embodied’s servers; families had to explain a dead robot friend. Community OpenMoxie later enabled re-flashing — after the fact.
Jibo (2019): cloud-sunset, bricked. Final dance, then dark. No durable rescue in our record.
Vector (Anki 2019 → DDL 2026): two corporate collapses. Cloud voice unreliable; Escape Pod self-host is the practical survival path. We list Vector 2.0 as discontinued/unsupported.
Neato (Vorwerk wind-down): cloud-sunset without always bricking the motors — app/Wi-Fi features degraded while some units still cleaned. Proof that “vacuum still rolls” is not the same as “product survived.”
Original Aibo support end (2014): discontinuation of parts/repair produced robot “funerals” in Japan — longevity risk is not only cloud, it is also parts and maker commitment. Sony later relaunched Aibo; the ERS-1000 still rates full server dependency on our profile.

Even a trillion-dollar maker will brick a line: Amazon discontinued Astro for Business and disabled units (refunds offered). Consumer Astro remains invite-only with full cloud dependency and longevity grade D on our tables — company size is not immunity.

How to predict your robot’s shutdown path

Read server dependency on the profile: none (cannot be cloud-bricked), partial (core motion/clean often continues; apps/voice/maps may die), full (high brick risk).
Check longevity grade and company-stability risk — formula on methodology.
Ask whether a mandatory subscription gates intelligence (see subscription lock-in checklist).
Look for a documented self-host or offline mode before purchase — rescues after bankruptcy are charity, not a warranty.

Prices and scores cited elsewhere on this site come from our July 2026 verified dataset (29 robots). We do not invent street prices here; open data is on /data.

FAQ

Do all robots brick when the maker shuts down?

No. Outcomes split three ways in our graveyard: full brick (Moxie, Jibo), feature degradation with local rescue paths (Vector via Escape Pod), or products that never shipped volume (Kuri, Laundroid). Server dependency on each profile predicts which path is likely.

How fast do servers go dark after a company fails?

There is no fixed timeline. Embodied ceased operations in December 2024 and Moxie’s servers went dark on 30 January 2025. Anki collapsed in April 2019 and cloud features limped until Digital Dream Labs later acquired assets — then failed again. Treat any full-cloud robot as able to die on short notice.

Can community projects save a bricked robot?

Sometimes. OpenMoxie re-flashes Moxie units; Escape Pod self-hosts Vector. Rescue is uneven, technical, and never guaranteed at purchase — our graveyard records which events have a documented rescue path.

How does Best Robot Match score this risk?

Every robot gets a server-dependency rating (none / partial / full), a longevity grade, and company-stability judgement. Formula and limits are on /methodology; sourced shutdown events live on /graveyard and /evidence.

Match on longevity and shutdown risk — not marketing claims

Open the match engine →