250K hotel fog rooms · 5M in-room edge GPUs — one global AI network.
Two tiers, one program: back-of-house fog nodes across 250,000 chain hotels for nightly batch AI, plus 5,000,000 in-room GPU edge servers auto-released on check-in. No new data hall, no civil works — hospitality operations are never touched.
Back-of-house rack node
Off-peak batch inference, fine-tuning micro-jobs, render-farm bursts. Rack-level micro-liquid keeps the mechanical room quiet and cool.
- ›Hours: 22:00 → 07:00 local, auto-scaled
- ›Typical yield: 6–12 GPU-hrs / node / night
In-room GPU node
Lightweight AI image gen, cloud-esports and local low-latency inference. Revenue upside comes from premium “esports rooms” guests actively want to book.
- ›Instant release on check-in
- ›Esports-room ADR uplift retained by hotel
Two partnership models — pick your capex appetite
Aether funds the retrofit
Net revenue (after incremental power) split 50% Aether · 50% Hotel. Cloud-esports room upside is 100% yours; Aether keeps only a 5% tech-service fee.
Hotel owns the hardware
Aether provides the scheduler, demand channel and compliance layer. Net revenue split 70% Hotel · 30% Aether.
From survey to first revenue in ~30 days.
- D1Site surveyRack count, power, cooling headroom.
- D7RetrofitMicro-liquid loops & scheduler agent installed.
- D14CertificationResidency, ISO & fire-safety sign-off.
- D21Traffic rampAI workloads onboarded from global queue.
- D30First payoutMonthly revenue-share statement issued.
Share your city, brand and rack count — we'll come back with a modeled revenue estimate within 5 business days.
Hotel partner questions
No. In-room GPUs run only when the room is vacant and release compute the moment a guest checks in. Back-of-house fog nodes run on off-peak power at night. HVAC, noise and power draw stay within the property's normal envelope.
Standard rollout is 30 days: site survey (week 1), micro liquid-cooling retrofit and network commissioning (weeks 2–3), scheduler onboarding and pilot workloads (week 4). The hotel provides rack space or in-room slots, power and network; Aether handles hardware, cooling, ops and compliance.
Model A: Aether funds hardware and retrofit, revenue splits 50/50 — best for hotels that want zero capex. Model B: the hotel self-invests in hardware and takes 70% of GPU-hour revenue while Aether keeps 30% for scheduling, ops and demand — best for hotel groups that want asset ownership and higher upside.
12–18 months for the hotel side under Model B, driven by high-utilization nightly batch inference and cloud-esports co-selling. Model A hotels earn from day one with no capex.
Never. Guest network and Aether compute network are physically and logically isolated. All workloads are anonymized, tenant-scoped and audit-logged, satisfying PIPL, GDPR and PCI-DSS boundary controls.
Bring your hotel into the global compute network.
Tell us about your property portfolio and we'll come back within two business days with a site-survey plan and revenue model.