Presentation Components

Compact components for PowerPoint/Keynote slides

1. Pricing Cards (Compact)

Version A: Horizontal (3 Cards)

BASIC
RTX PRO 4000
24 GB
VRAM • 46.9 TFLOPS
from €4,299
  • ✓ Models up to 20B parameters
  • ✓ Plug-and-Play
  • ✓ Open WebUI pre-installed
PRO
RTX PRO 6000
96 GB
VRAM • 125 TFLOPS
from €13,599
  • ✓ Models up to 120B+ parameters
  • ✓ On-site installation
  • ✓ Premium Support
CUSTOM
By request
Multi-GPU
NVLink • Cluster
On request
  • ✓ Rack-Mount Servers
  • ✓ Cluster Solutions
  • ✓ Custom configuration

Version B: Mini (for small slides)

BASIC
24 GB
46.9 TFLOPS
€4,299
PRO
96 GB
125 TFLOPS
€13,599
CUSTOM
Multi-GPU
NVLink
Request

Version C: Table

ModelGPUVRAMTFLOPSPrice
AI Cube Pro TOPRTX PRO 600096 GB125from €13,599
AI Cube CustomMulti-GPU∞∞On request

2. USPs / Benefits

Version A: 4-Grid

🔒
GDPR compliant
100% local
💰
No token costs
One-time payment
⚡
Plug & Play
Ready to use
🇩🇪
Made in Germany
DE/EN Support

Version B: Badges (for header)

GDPR compliantNo CloudNo token costsup to 96GB VRAMMade in Germany

Version C: Checklist

✓100% GDPR compliant
✓No cloud dependency
✓No ongoing costs
✓German & English support
✓Open WebUI pre-installed
✓Plug & Play setup

3. Comparison: AI Cube vs. Cloud

Cloud vs. On-Premise

☁️ Cloud (ChatGPT etc.)🖥️ AI Cube
Data Privacy❌ Data with third party✓ 100% local
GDPR❌ Problematic✓ Fully compliant
Costs❌ Per token / month✓ One-time
Availability❌ Internet dependent✓ Always available
Vendor Lock-in❌ Yes✓ No

4. Technical Specs

AI Cube Pro - Specs Card

AI CUBE PRO
RTX PRO 6000
NVIDIA Blackwell
96 GB
GDDR7 VRAM
125
TFLOPS
120B+
Parameters
PCIe 5
Interface

Specs Comparison Table

SpecificationBasicPro
GPURTX PRO 4000RTX PRO 6000
VRAM24 GB GDDR796 GB GDDR7
TFLOPS46.9125
Max. Model Size~20B Parameters120B+ Parameters
SSD1 TB NVMe2 TB NVMe
Pricefrom €4,299from €13,599

5. Use Cases

Use Case Icons

📝
Text Generation
💬
Chatbot
🔍
Document Analysis
💻
Code Assistant
🎯
RAG / Knowledge Base

6. Reseller Benefits

Reseller Advantages

🏷️
White-Label
Sell under your own brand
💵
Attractive Margins
Tiered pricing for partners
🤝
DE/EN Support
Also for your end customers

White-Label Process

1
Request
2
Manufacturing
3
Branding
4
Delivery

7. Hero / Title Slides

Version A: Dark Hero

GDPR compliantMade in Germany

AI Cube

On-Premise AI Server for Enterprises

Local LLM Inference • No token costs • 100% data privacy

Version B: Light Hero

WZ-IT SOLUTIONS

AI Cube

The On-Premise AI Server for GDPR-compliant enterprises

✓ Up to 96GB VRAM✓ Open WebUI✓ Plug & Play

8. Cost Comparison

ROI / Break-Even

AI Cube Pro pays for itself after
~6 months
when used like ChatGPT Enterprise
~€2,500/month
ChatGPT Enterprise (10 users)
€13,599 one-time
AI Cube Pro

8b. ROI Calculator - GPT-5 mini vs. AI Cube

Version A: Compact ROI Card

ROI at 500 Tokens/s (24/7)

GPT-5 mini • Input:Output = 3:1

ROI < 4 Months
OpenAI Cloud
$42,768
per year
$3,564 / month
5.18 billion tokens/month
AI Cube Pro
€13,599
one-time
€0 ongoing costs
Unlimited tokens
<4 Mo.
Break-Even
€30,000+
Savings/Year
∞
Token Limit

Version B: Detailed Calculation

Cloud Cost Calculation

GPT-5 mini @ 500 Output Tokens/s (24/7 Operation)

Assumptions
500 t/s
Output
1,500 t/s
Input (3:1)
24/7
Operation
5.18B
Tokens/Month
Input Cost
3,888M × $0.25= $972
Output Cost
1,296M × $2.00= $2,592
Total OpenAI Cost
$3,564 /month
$42,768 /year
AI Cube Pro
One-time purchase
€13,599
ROI in ~3.5 months

Version C: Visual Cost Comparison (12 Months)

Cost Development Over 12 Months

GPT-5 mini vs. AI Cube Pro

OpenAI GPT-5 mini$42,768
AI Cube Pro~€13,000
First Year Savings
~€27,000+
At 500 tokens/s continuous load

Version D: Sales-Ready (very compact)

ROI CALCULATOR

GPT-5 mini vs. AI Cube Pro

Cloud (500 t/s, 24/7)$42,768/year
AI Cube Pro€13,599 one-time
Break-Even<4 Months
From ~300 t/s, cloud is more expensive than on-prem

Version E: Infographic with Pricing Details

500 Output Tokens/s @ 24/7

When Does On-Premise AI Pay Off?

2,000
Total Tokens/s
500 out + 1,500 in
5.18B
Tokens/Month
GPT-5 mini Pricing
Input: $0.25/1M
Output: $2.00/1M
$3,564
per month
Month 0
~3.5 Mo.
Break-Even
12 Months
$42,768
Cloud / Year
€13,599
AI Cube Pro one-time
~€27,000+
Savings / Year

Key Takeaways for Sales

  • From ~300 tokens/s continuous load, cloud is more expensive than on-prem
  • At 500 t/s: GPT-5 mini costs >$42,000 per year
  • AI Cube Pro: ~€13,000 one-time → ROI in <4 months
  • AI Cube Pro pays for itself in under 4 months at comparable usage

13. What We Do

Service Overview

What We Do

Full-stack IT services for modern businesses

Cloud Infrastructure
AWS, Azure, Hetzner
Software Development
React, Node.js, Python
Hardware
AI Servers, GPU Systems
IT Security
Audits, Pentesting, Compliance
Networking
VPN, Firewalls, IDS/IPS
Managed Services
24/7 Monitoring & Support

Alternative: Horizontal Strip

Cloud
Software
Hardware
Security
Networking
Managed

12. Reseller Program (Optional)

Partner Benefits

For partners only

Reseller Program

Become an AI hardware provider

White-label hardware
Sell under your brand
Attractive margins
Tiered pricing model
Priority support
Dedicated partner channel
Custom configurations
Tailored to your customers
Direct-to-customer delivery
Optional drop-shipping
Trade-in upgrade program
Upsell existing customers

11. Delivery & Setup

Full-Service Delivery

Delivery & Setup

End-to-end full-service

Preinstalled & tested
Before shipping
Optional on-site delivery
DE & NL available
Professional installation
Expert setup on-site
GPU driver & benchmarking
Optimized configuration
Container environment
Docker ready
VPN & security config
Network integration
Optional managed service
Ongoing support, updates & monitoring available

10. Security & Compliance

Security Features

Security & Compliance

Designed for environments with strict requirements

100% GDPR compliant
Full regulatory compliance
No external data transfer
Data processed on local hardware only
Secure Boot, TPM 2.0
Hardware-level security
Optional encryption
Full disk encryption available
No telemetry
Zero data collection
On-prem integration
VPN & firewall compatible

9. Real-World Scenarios

Use Cases by Industry

Real-world scenarios powered by the AI Cube

Legal
Document search, RAG, contract review
Healthcare
SOP knowledge bases, protocol retrieval
IT Teams
Local code assistant, automation pipelines
Enterprises
Internal GPT-like assistants
Government
Sensitive data, on-prem compliance
Research
Model experimentation & fine-tuning

9. Performance Benchmarks

Blackwell Generation (Current)

Reference:ChatGPT-4o outputs ~50 tokens/s max
ModelAI Cube Pro (96GB)AI Cube Pro (96GB)
GPT-OSS 20B50 t/s1 User200 t/s1 User
GPT-OSS 120B—not enough VRAM
150 t/s 1 User
1050 t/s 20 Users, small ctx
300-500 t/s 20 Users, mixed ctx
Qwen3-30B-A3B FP8—not enough VRAM
90-150 t/s 1 User, 1k ctx
650 t/s 10 Users, 1k ctx
22 t/s 1 User, 256k ctx
115 t/s 10 Users, 256k ctx
311 t/s 10 Users, 64k ctx
413 t/s 10 Users, 32k ctx

Previous Generation (Ada GPUs) - LLaMA 3

ModelAI Cube Pro
LLaMA 3 (8B Q4_K_M)60 t/s1 User130 t/s1 User

Screenshots of this page can be used directly in presentations.

Tip: Use browser DevTools (F12) → Device Toolbar for exact sizes