O.ATLAS



Atlas Datacenters: Pioneering AI Infrastructure

Our Atlas Datacenters will represent the cutting edge of AI infrastructure, powered by the groundbreaking Cerebras CS-3 systems featuring WSE-3 chips—the world’s largest and fastest AI processors. Each Atlas datacenter will host a cluster of 64 CS-3 servers, with 900,000 AI-optimized cores and 44GB of on-chip memory per chip, enabling processing capabilities that will outperform entire traditional supercomputing installations.


Community Ownership and Sovereignty

These datacenters will ensure O sovereignty while being community-owned through our innovative RWA NFT program. From Atlas DC1 to DC2, we will build a decentralized network of compute power that will scale to house the most advanced AI infrastructure in the world. Unlike traditional cloud providers or tech giants' datacenters, our Atlas facilities will be owned by our community through fractional NFTs, with rewards distributed in $O tokens over a 3-year period. This unique model will ensure that our computational resources remain independent and aligned with our community's interests rather than corporate objectives.


Unmatched Performance Capabilities

The performance capabilities will be extraordinary - we will achieve inference speeds 20 times faster than traditional cloud providers and 3 times faster than even Groq's LPU solutions. This won't just be about raw speed though - it will be about building an unstoppable foundation for truly sovereign artificial intelligence that will serve humanity's collective interests.


Why WSE-3?

Cerebras will provide significant advantages over model-to-model routing systems and server-less inference layers, even though it will not yet offer model-to-model communication on the same chip (a feature expected in Q2 2025).

Unified Processing Power:WSE-3’s Scale: With 900,000 AI-optimized cores and 4 trillion transistors on a single wafer-scale chip, Cerebras will deliver unmatched computational power. This will eliminate the need for complex orchestration across multiple GPUs or servers that routing systems and serverless layers will require. Instead of managing distributed resources, Cerebras will offer a single, massively powerful chip that can handle large-scale AI workloads.

Scalability and Simplicity:No Complex Orchestration: Traditional systems will often require intricate routing logic and distributed programming to connect models across servers. In contrast, Cerebras will scale from 1 billion to 24 trillion parameters without changing code, making it easier to manage large AI models. While it currently lacks model-to-model communication on-chip, the simplicity of scaling within the WSE-3 architecture will offer a clear operational advantage.

Memory Bandwidth and Efficiency:High Bandwidth, Low Latency: Cerebras’s 21 PB/s memory bandwidth will far exceed that of traditional GPUs, allowing for fast, efficient processing of AI tasks. This level of bandwidth, combined with 44 GB of on-chip memory, will ensure that data flows smoothly within the chip, avoiding the latency issues that distributed systems face when moving data between nodes.

Energy Efficiency and Cost:Lower Total Cost of Ownership: While the initial investment in Cerebras might be higher, the efficiency gained through reduced power consumption, simplified infrastructure, and lower operational complexity will lead to lower overall costs. Serverless systems, while flexible, will often incur ongoing costs and resource management challenges that Cerebras will mitigate with its integrated design.

Simplified Deployment:Expert Installation and Support: Cerebras will provide white-glove installation and continuous software upgrades, reducing the burden on in-house teams and ensuring optimal performance over time. This will contrast with the self-managed nature of model-to-model routing systems and server-less layers, which will be resource-intensive to maintain and scale.

AI & HPC Chip Comparison: Cerebras WSE-3 vs. Groq LPU vs. NVIDIA H100

Feature
Cerebras WSE-3
Groq LPU Chip
NVIDIA H100
Architecture
Wafer-Scale Engine (WSE-3)
Deterministic Processor (SIMD)
Traditional GPU
Chip Size
46,225 mm$^2$ (56x larger than a GPU)
Not specified
Standard GPU size (814 mm$^2$)
Core Count
900,000 AI-optimized cores
Not specified
16,896 FP32 + Tensor Cores
Transistors
4 Trillion
Not specified
80 Billion
Peak Performance
125,000 TFLOPs
Up to 750 TOPs, 188 TFLOPs
3,958 TFLOPs
Memory
44 GB on-chip SRAM
230 MB on-die SRAM
0.05 GB
Memory Bandwidth
21 Pb/s
Up to 80 TB/s on-die memory bandwidth
3.35 TB/s
Interconnect Bandwidth
214 Pb/s
16 integrated RealScale interconnects
0.0576 Pb/s
Power Consumption
23 kW
Max: 300W, TDP: 215W, Avg: 185W
700W
Process Node
5nm
14nm
Not specified
Scalability
Scales from 1B to 24T parameters with no code changes
Scalable with chip-to-chip interconnects
Requires multiple GPUs
Precision Levels
Optimized for sparse matrix computations
INT8, INT16, INT32, FP32, FP16
FP8, FP16, and more
Deployment
White-glove installation and support services
Simplified integration
Requires complex setup
Cooling
Custom cooling solutions integrated
Not specified
Standard data center cooling solutions
Use Cases
Large Language Models, multimodal support, Supercomputer Clusters
AI,ML, and HPC workloads with ultra-low latency
AI model training, inference, HPC
Cost
$1.5M per unit; $1.218M with bulk purchase (64 units)
Not specified
$30,000 - $40,000 per GPU
Installation
Expert installation and validation testing
Easy-to-use software suite for fast integration
Requires user or integrator setup
Support
Continuous software upgrades and managed services
End-to-end on-chip protection, error-correction code (ECC)
Standard support options

Cost

Offering
Price
Quantity
Total Cost
64-Node Wafer Scale Cluster (supports GPT model tasks (pre-training, fine-tuning, inference) with one year of included support and software upgrades)
$1,219,000 per WSE3 node
64-nodes
$78m
Installation (Installation at Facilities)
$40,000
One Time
$40,000
Delivery (Delivery of hardware to the installation site)
$80,000
One Time
$80,000
Professional Services (Consulting Services for Machine Learning and/or
Datacenter Facility readiness; Contracted in 100-hour
blocks; example contracting configuration shown)
$350/hr
100 hours
$35,000
SOC2 Datacenter Hosting Ops
$5,000 x 64 node = $320,000
48 months
$15,360,000
Datacenter Fiberline Cost
$17,000/Month
48 months
$816,000
2.5 MW Electricity Cost
$20,000/Month
48 months
$9,600,000
TOTAL
$ 103,851,000
Total $ 103,851,000