The Problem

GPU access shouldn't be this hard

IT Procurement

3-month cycles for cloud access approvals. Purchase orders, vendor evaluations, security reviews — before a single student writes a line of code.

Colab Limitations

Random disconnections mid-training. GPU queues during peak hours. No real GPUs for serious workloads. Not suitable for a structured course.

HPC Complexity

Batch queues, module systems, no container isolation. Students spend more time fighting infrastructure than learning machine learning.

Cost Comparison

A fraction of the cost of Colab Pro subscriptions

40 students, 12 weekly 3-hour labs, L40S GPUs

Platform	Cost / Semester	Setup	Notes
razorBridge	~€104	Minutes	Pay-per-use, real GPUs, SSH access
Colab Pro	~$2,880	None	40 × $12/mo × 6 months. Disconnections, GPU queues.
AWS Academy	Variable	Complex	IAM, VPCs, quotas. Requires dedicated IT support.
University HPC	Capital	6+ months	Procurement, installation, sysadmin staff.

Estimate based on L40S at €1,44/hr (pricing as of June 2026) shared across 20 users per GPU. Actual costs depend on GPU type and utilization.

For Professors

Four steps. Zero DevOps.

1

Create Event

Set GPU type, start/end dates, and participant limit. Takes 2 minutes.

2

Share Code

Students join with rb event join and a join code.

3

Monitor

See who's connected, GPU utilization, and credit usage in real time from the dashboard.

4

Done

Auto-teardown when the event ends. Download a cost report for your department.

For Students

Three commands. Real GPU.

1

Install CLI

$ pipx install rb-cli

2

Join Event

$ rb auth login
$ rb event join CUDA-2026
✓ Joined "ML Workshop" — GPU: L40S (48GB)

3

SSH In

$ rb blade ssh
Connected to gpu-blade-abc123 (L40S, 48GB VRAM)
student@gpu-blade:~$ nvidia-smi

Features

Built for Academic Workflows

Pre-Configured Environments

PyTorch, TensorFlow, CUDA pre-installed. Students start coding, not configuring.

Roster CSV Upload

Import your student list. Everyone gets access when the event starts.

Usage Monitoring

Real-time dashboard showing who's connected, GPU utilization, and credit usage.

Event Extension

Need more time? Extend events with one click. No re-provisioning needed.

Budget Control

Set semester budgets with hard caps. Invoice billing for university procurement.

Multi-User Isolation

Each student gets an isolated container on shared GPU hardware. No interference between users.

Security & Compliance

Built for academic environments

SSH Container Isolation

Each user runs in a separate container with resource limits.

Hashed API Tokens

Credentials are hashed at rest. No plaintext storage.

EU Data Centers

All infrastructure runs in European data centers. GDPR compliant.

No Data Retained

Student data and workspaces are purged after the session ends.

Share with Your Department

Request our university one-pager with pricing, security details, and setup instructions.

Request University Materials

Ready to equip your ML course with real GPUs?

From €500/semester. Invoice billing. Zero IT overhead.

Request a Demo Request a Quote

GPU compute for your entire ML course