Sovere S4.
For voice.
For more teams.
S4 brings four GPUs to Voice, coding assistance and staff questions. A larger starting point when several teams share the system.
NVIDIA RTX PRO 6000 Blackwell
96 GB per GPU. Up to four isolated partitions per card.
- People, approx.
- 300–800
- Total GPU memory
- 384 GB
- NVIDIA GPUs
- 4
- Complete system · year 1 included
- $225,000
Approximate team size. Capacity depends on model choice, task size and simultaneous use.
S4’s main uses.
For 300–800 people, approximately.
Consider S4 when several teams share AI or you want to introduce Voice. We size it around the calls, questions and coding tasks running at your busiest times.
Put your call workflow on S4.
Use Voice on S4 to answer customer calls, follow your policies and pass requests to your systems or a person.
Coding help for a wider team.
Choose Code on S4 when more developers need help writing and reviewing changes using your repositories and internal libraries.
Company answers for more teams.
Use Knowledge on S4 to answer questions across teams, drawing on your guides, tickets and decision notes.
S4 specifications.
| Graphics cards | 4 × NVIDIA RTX PRO 6000 Blackwell Server Edition |
|---|---|
| GPU memory | 96 GB per GPU · 384 GB total |
| Maximum GPU power | 2.4 kW for the 4 GPUs |
| CPU, RAM, storage & networking | Confirmed in your quote |
S4 pricing.
- HardwareOne-time payment
- $90,000
- Software, model setup & first-year serviceIncluded in your initial purchase
- $135,000
- Total initial purchaseComplete system + first service year
- $225,000
- Annual managed AI serviceModel improvements, optimisation & support
- $144,000per year, from year 2
Your first 12 months of managed AI service start at system acceptance and are included in the initial purchase. Annual renewal starts in year 2. Prices in USD, excluding tax. Your written quote confirms the configuration, service scope and any additional costs. See what’s included.
Would S2 be enough?
S2 is a starting point for one team’s lighter work. Choose S4 when Voice or shared use calls for more capacity.
| At a glance | Sovere S2 | Sovere S4Currently viewing |
|---|---|---|
| People, approx. | 20–150 | 300–800 |
| Starting workload | One team’s lighter coding, document or support work | Voice or shared coding and staff support |
| Suggested AI models | Code · Documents · Knowledge | Voice · Code · Knowledge |
| GPU capacity | 2 GPUs · 192 GB total memory | 4 GPUs · 384 GB total memory |
| Hardware price | $50,000 | $90,000 |
| Total initial purchase · first service year included | $159,000 | $225,000 |
| Annual managed AI service · from year 2 | $120,000 | $144,000 |
| Next step | Explore S2 |
Questions about S4.
Call length, response speed, model choice and other work on the server affect capacity. We assess your call workflow before agreeing a simultaneous-call target.
We confirm dimensions, electrical supply, network and maintenance clearance for your installation.
Yes, if the combined workload fits. We plan GPU allocation and peak demand together so the chosen applications have the capacity they need.