The Allen Institute for AI (Ai2) is leading a $152 million public-private effort to build open artificial-intelligence infrastructure for scientific research. Announced August 14, 2025, the project combines a $75 million NSF award with a $77 million NVIDIA contribution. It is research and infrastructure support—not a venture-capital round, acquisition or unrestricted $152 million cash payment to Ai2.
The formal project is the Open Multimodal AI Infrastructure to Accelerate Science, or OMAI. Its goal is to provide open models, datasets, software, evaluations and computing resources for work in fields such as materials science, biology and energy.
What was announced
The National Science Foundation and NVIDIA announced support for Ai2 to develop OMAI, a national-scale ecosystem for scientific AI. NSF is providing $75 million through its Mid-Scale Research Infrastructure program, while NVIDIA is contributing $77 million. Ai2 and NSF describe the combined total as $152 million.
The announcement came from both organizations on August 14, 2025. The NSF description calls the arrangement a partnership to develop advanced AI models and infrastructure for the U.S. scientific community. Ai2’s account is available in its announcement.
Free tools Windows power users keep installed
One-click scans. No signup required.
#1 Best Overall
That wording matters. Public descriptions do not characterize the arrangement as equity financing, a startup fundraising round or an unrestricted donation. “Combined support” or “public-private research and infrastructure investment” is more precise than saying Ai2 raised $152 million from investors.
What OMAI is intended to build
OMAI is designed as an open scientific AI stack rather than a single chatbot. Planned components include:
- Open multimodal foundation models that can work with combinations of text, images and other scientific information.
- Scientific AI tools and applications for research workflows.
- Training, evaluation and data infrastructure.
- Computing resources for developing models and conducting research with them.
- Software and documentation intended to let researchers inspect, adapt and extend the systems.
- Research into the science and engineering of AI itself.
NSF says these tools could help scientists process and analyze research, generate code and visualizations, and connect new findings with earlier work. Materials science, biology and energy are among the potential application areas cited by NSF. The project’s OMAI page describes its mission and related work.
The public announcements do not provide a complete line-item budget, so readers should not assume that each dollar maps to a particular model, GPU cluster or university award.
Recommended Free Tools
Rank #2
- NVIDIA Volta GV100 Architecture — 4,608 CUDA Cores, 640 1st-Gen Tensor Cores delivering 14 TFLOPS FP32 and 112 TFLOPS deep learning performance for AI training, inference, HPC, and scientific computing workloads
- 32GB HBM2 ECC Memory — 900 GB/s Bandwidth — High-bandwidth memory on a 4096-bit bus with ECC error correction provides the memory capacity and throughput required for the largest AI models, simulations, and datasets
- PCIe 3.0 x16 Interface — 250W TDP — Standard PCIe Gen3 connectivity with passive cooling designed for enterprise rack server deployment in HPE ProLiant, Dell PowerEdge, and Supermicro platforms with adequate chassis airflow
- NVLink — Scale to 96GB Unified Memory — Connect two V100 GPUs via NVLink at 300 GB/s bi-directional bandwidth to scale GPU memory from 32GB to 96GB for larger AI training and HPC workloads
- Multi-Precision Computing — Supports FP64 (7 TFLOPS), FP32 (14 TFLOPS), FP16 (112 TFLOPS) and INT8 precision modes for flexible deployment across training, inference, and scientific simulation workloads
Why Ai2 is leading the effort
Ai2 is a nonprofit research institute, not a conventional commercial AI company. Its existing open-model work includes the OLMo language-model family and Molmo multimodal models. OMAI builds on that strategy by attempting to make the surrounding data, software, evaluation and computing environment useful to outside researchers as well as Ai2 scientists.
The project is also part of NSF’s Mid-Scale Research Infrastructure portfolio. That makes its purpose closer to building shared national research capacity than to financing a product launch.
Who is involved
OMAI is a consortium rather than an Ai2-only program. Listed academic partners are:
- University of Washington
- University of Hawaiʻi at Hilo
- University of New Hampshire
- University of New Mexico
Ai2’s principal investigator is Noah A. Smith, Ai2’s senior director of NLP research and an Amazon Professor at the University of Washington’s Paul G. Allen School. The University of Washington’s listed co-principal investigator is Hanna Hajishirzi. Other named co-principal investigators are Travis Mandel, Samuel Carton and Sarah Dreier. The university’s announcement provides additional institutional and investigator details at the Allen School website.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →What “fully open” should mean in practice
“Fully open” is an ambition and a description of the project’s approach, not a guarantee that every future release will expose every artifact under unrestricted terms. Ai2 says its model approach is meant to provide the components needed to analyze models, fine-tune them and train them from scratch.
For each release, researchers should check which of these are actually available:
- Weights: the trained parameters.
- Data: training data, sources or meaningful documentation about what was used.
- Code: training and inference software, including preprocessing.
- Evaluations: tests, methods and results that others can inspect or reproduce.
- Infrastructure: access to the compute, tools and instructions needed to run or extend the work.
- Licensing: explicit terms covering research, modification and any commercial use.
Open weights alone do not make a project reproducible if the data, preprocessing pipeline or required hardware remains inaccessible. Scientific data can also involve privacy, copyright, export-control or other restrictions. The public announcement does not establish a complete model-by-model licensing policy, universal access policy or release schedule for every planned system.
OMAI is not NAIRR
OMAI and the National AI Research Resource (NAIRR) are related to the broader U.S. effort to expand research access, but they are different initiatives.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Rank #4
- Memory Size: 16 GB GDDR6 ECC.
- Memory Bus Width: 128-bit.
- Memory Bandwidth: 200 GB/s.
- CUDA Cores: 1280.
- Peak Single Precision floating point performance: 18 Tflops (GPU Boost Clocks).
| OMAI | NAIRR |
|---|---|
| Specific project led by Ai2 | Broader NSF-led national resource |
| Focuses on open multimodal AI infrastructure and models for science | Coordinates access to computing, models, datasets, software and expertise |
| Supported by $75 million from NSF plus $77 million from NVIDIA | Built through a multi-agency and public-private framework |
| Develops models and infrastructure | Connects researchers to a wider resource ecosystem |
NSF’s NAIRR page describes the broader resource. Ai2 is separately a founding NAIRR partner, contributing open models, training software, evaluation software and data resources, as explained in its NAIRR announcement. OMAI is not a replacement for NAIRR or another name for it.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Where the project stood in 2026
The August 2025 funding announcement did not mean the entire ecosystem was immediately operational. In an update dated May 7, 2026, Ai2 said OMAI compute had begun coming online, including systems powered by NVIDIA Blackwell Ultra hardware. The update indicates a move from award announcement toward deployment, but it does not establish that every promised model, tool or researcher-access mechanism is complete. Ai2 reported that status at OMAI compute comes online.
Why the project matters—and what could limit it
Potential benefits
- Reproducibility: Researchers can inspect and build on systems instead of relying only on black-box APIs.
- Scientific specialization: Models and tools can be adapted to domain-specific data and workflows.
- Broader participation: Shared infrastructure could help universities that lack hyperscale corporate resources.
- U.S. research capacity: NSF and Ai2 frame the project as part of maintaining leadership in AI-enabled science.
These are intended benefits, not results already demonstrated across science. A meaningful assessment will require peer-reviewed work, replication, useful benchmarks and evidence of actual scientific discoveries.
Trade-offs and open questions
- More transparency can improve scrutiny and innovation while also making capable systems easier to misuse.
- National compute can broaden access, but high-end GPU capacity will remain scarce and may require applications, quotas or prioritization.
- Scientific models may need domain expertise and may behave differently from consumer chatbots.
- The project seeks open resources while relying substantially on NVIDIA hardware, creating a potential concentration concern.
- Releasing complete data and training details can be slower and harder than operating a closed system.
Researchers evaluating OMAI should ask who can apply for access, where users are located, how compute is allocated, what each release licenses, whether results can be reproduced and how privacy, security and conflicts of interest are governed. The size of the award alone cannot answer those questions.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →The bottom line for readers
Ai2 has not announced a conventional $152 million financing round. It is leading OMAI, a federally supported and NVIDIA-supported attempt to build open AI models and shared infrastructure for science. NSF is supplying $75 million and NVIDIA $77 million. Compute deployment had started by May 2026, but the complete ecosystem and its access rules were still being developed. The project’s significance will ultimately depend on how open the releases are in practice and whether researchers outside the largest institutions can use them to produce verifiable scientific results.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
