|
Inference at the edge of ambition.
Nine nodes. Eleven GPUs. Five architecture generations. Free AI compute for the communities that need it most.
Data Hub
Aggregate metrics across the ECO-Foundry local cluster.
Cluster Architecture
Nine local nodes connected via Cat 8 Ethernet with elastic burst capacity on the cloud.
H100, H200, A100, B200, GB300
Elastic burst capacity
Multi-archBlackwell
DGX Spark · GB10 Grace Blackwell
128 GB Unified · 1 PFLOPS
Blackwell + GraceAmpere
RTX 3080 Ti
12 GB GDDR6X
Ampere GA102RTX 3080 Ti
12 GB GDDR6X
Ampere GA102RTX 3080 Ti
12 GB GDDR6X
Ampere GA102Turing
RTX 2080 Ti
11 GB GDDR6
Turing TU102Pascal
2x TITAN Xp
24 GB GDDR5X
Pascal GP102GTX 1080
8 GB GDDR5X
Pascal GP104GTX 1080
8 GB GDDR5X
Pascal GP104Apple Silicon
M4 Max 40-core
64 GB Unified
Apple Silicon 3nmECO-Foundry
Energy-conscious, always-on compute serving local communities. Nine nodes spanning five architecture generations, open for community access.
Feynman
NVIDIA DGX Spark · Grace Blackwell
“What I cannot create, I do not understand.”
NVFP4 · FP8 · INT8 · BF16 · FP16 · TF32 · FP32
Laplace
NVIDIA GeForce RTX 3080 Ti
“The weight of evidence for an extraordinary claim must be proportioned to its strangeness.”
Euler
NVIDIA GeForce RTX 3080 Ti
“Read Euler, read Euler, he is the master of us all.”
Riemann
NVIDIA GeForce RTX 3080 Ti
“If only I had the theorems! Then I should find the proofs easily enough.”
Noether
Dual NVIDIA TITAN Xp
“My methods are really methods of working and thinking; this is why they have crept in everywhere anonymously.”
Dirac
NVIDIA GeForce RTX 2080 Ti
“The aim of science is to make difficult things understandable in a simpler way.”
Gauss
NVIDIA GeForce GTX 1080
“Mathematics is the queen of the sciences, and number theory is the queen of mathematics.”
Planck
NVIDIA GeForce GTX 1080
“Science cannot solve the ultimate mystery of nature. And that is because we ourselves are a part of the mystery.”
Turing
Apple M4 Max · MacBook Pro
“The sciences do not try to explain, they hardly even try to interpret, they mainly make models.”
Exo-Scale
Elastic GPU capacity via the cloud. When community projects need more than the local cluster can provide, we scale to the cloud.
GPU Cloud
The Superintelligence Cloud
Provider
Cloud Partner
GPU Fleet
H100, H200, A100, B200, GB300
Instances
1x to 8x GPU
Clusters
16 to 2,000+ GPUs
Software
PyTorch, CUDA, cuDNN
Status
Active
Network
Interconnect topology across the ECO-Foundry and affiliated networks.
Cat 8 Ethernet
40 Gbps rated
S/FTP, 7 Gbps effective
eero Max 7
WiFi 7
Tri-band, 4.3 Gbps
Fiber Uplink
1 Gbps symmetric
UCLA eduroam
WPA2-Enterprise
Internet2, 802.1x
DGX Spark
200 Gbps
2x QSFP
LADWP Power
LADWP
1.75 kW peak GPU draw
Built for Impact
Purpose-built infrastructure powering free AI tools for local communities.
Community Inference
Free inference for community partners. Models up to 200B parameters on Feynman. 128 GB unified memory, 1.5+ PFLOPS cluster peak. No API limits, no rate caps, no barriers.
Applied AI Deployment
From climate models to educational tools. Local fine-tuning on the ECO-Foundry, cloud burst via Exo-Scale for larger workloads. We help partners go from idea to deployed model.
Open Compute
52,992 GPU cores across nine nodes available to nonprofits, educators, and community developers. Always available. No cloud dependency.