DRAFT RESEARCH · 19 AUG 2026
The Computational State
NEW MEXICO COMPUTATIONAL COMMONS · HPC + AI + SCIENCE + PUBLIC MISSION
New Mexico already possesses many of the ingredients of a nationally significant advanced-computing ecosystem. The strategic opportunity is not simply to buy another supercomputer. It is to connect the constellation.
Research thesis
New Mexico's national laboratories, research universities, New Mexico Consortium, scientific workforce, advanced-computing infrastructure and distinctive public challenges create an unusual opportunity: a federated computational commons that helps move appropriate workloads from development environments to university HPC, regional and national research infrastructure, and, where mission, partnership and authorization permit, national-laboratory collaboration.
This is deliberately a draft hypothesis, not a claim that these independently governed resources can simply be pooled. Eligibility, security boundaries, funding restrictions, export controls, data governance, mission requirements and institutional authority remain first-class design constraints.
The corridor
NMC as connective tissue
The New Mexico Consortium is structurally interesting because its advanced-computing work already spans HPC, data-intensive computing, artificial intelligence and emerging technologies, and because its institutional mission connects research, universities, laboratories and industry. Rather than attempting to duplicate LANL, Sandia or university computing centers, NMC could expand its role as an on-ramp, broker, testbed and translation layer.
A computational ladder
Tier 0: workstation and local development. Tier 1: accessible development, notebook, container and modest GPU environments. Tier 2: university HPC at institutions such as UNM, NMT and NMSU subject to their policies. Tier 3: regional, NSF and DOE research infrastructure. Tier 4: national-laboratory collaboration where sponsorship, mission and authorization justify access.
The ladder turns HPC from an intimidating destination into a progression: Python → Linux → containers → schedulers → GPU → MPI → distributed AI → scientific computing at scale.
Flagship mission: a Digital Twin of New Mexico Water
Water is an unusually strong demonstration problem because it demands geospatial analysis, hydrology, historical records, remote sensing, simulation, uncertainty analysis, climate data and increasingly AI. A research program could integrate authoritative datasets and models to explore drought, groundwater drawdown, aquifer recharge, surface-water interaction, reservoir operations, wildfire impacts, infrastructure vulnerability and long-range scenarios.
The governance principle matters: the twin should augment scientific and public decision-making, preserve provenance and uncertainty, and never disguise model output as legal or scientific authority.
Other grand challenges
- Fire: high-resolution scenario modeling using terrain, fuels, weather, drought and infrastructure.
- Energy: grid resilience, renewables, storage, microgrids and water-energy interactions.
- Trusted AI: controlled evaluation of model robustness, agent security, data poisoning, supply-chain risk and human-accountable AI.
- Quantum + HPC: training and experimentation around hybrid scientific-computing workflows as the technology matures.
The economic-development layer
A computational commons can become infrastructure for companies as well as papers. A startup or public-interest team could enter with a problem and receive research translation, workforce, secure development patterns, data engineering and an appropriate compute path. The scarce asset is not a GPU by itself. The durable advantage is the network of people, institutions, expertise, data and pathways surrounding it.
180-day hypothesis test
- Inventory meaningful HPC, GPU, storage, network and AI capabilities, including eligibility and funding constraints.
- Convene a technical and governance council across willing institutions.
- Prototype a small federated portal and resource catalog without pretending institutional boundaries disappear.
- Select one grand challenge, preferably water, and define a bounded research workload.
- Run a computational challenge with researchers, students and partners.
- Publish results, unmet demand, governance lessons and a defensible expansion case before seeking major capital.
Destination
NEW MEXICO: THE COMPUTATIONAL STATE. Science at the scale of the landscape.
The proposition is not Silicon Valley in the desert. It is something native to New Mexico's strengths: national science, difficult landscapes, public mission, advanced computing and a culture capable of seeing enormous systems at once. The desert already contains extraordinary compute. The research question is how responsibly, securely and productively we wire the constellation together.
Research sources
Research status: DRAFT. Architecture, access assumptions, benchmark positions, funding paths and institutional roles require validation with the participating organizations before this becomes a proposal.