NVIDIA is giving away powerful AI for free — and making more money than its competitors from it
On June 4, 2026, NVIDIA released Nemotron 3 Ultra, the largest open-source artificial intelligence model in the Nemotron 3 line. The model weights, training data, and training methodologies have been made publicly available under a permissive license. The model is designed for long-lived autonomous agents and complex reasoning.
Unlike closed flagship models such as ChatGPT or Claude, Nemotron 3 Ultra can be downloaded, fine-tuned on proprietary data, and run on your own infrastructure. The focus here is not on maximum intelligence, but on openness, efficiency, and control.
What makes the architecture unique
Nemotron 3 Ultra is not just a "scaled-up transformer." It is based on a hybrid architecture combining three approaches: Mamba-2 layers, Attention layers, and Latent Mixture of Experts (Latent MoE). The Latent MoE mechanism directs each request only to the necessary "specialists" within the model, compressing data before transmission.
Mamba-2 layers process long texts quickly and efficiently: costs grow linearly, not exponentially. Attention layers accurately retain large volumes of text in memory. The model has approximately 550 billion parameters in total, but only about 55 billion are activated for processing each token. With a context window of 1 million tokens and a speed exceeding 300 tokens per second, this provides 5-6 times greater throughput and roughly 30% lower task costs.
NVIDIA's strategy: betting on the ecosystem
The main value of the release is not the model itself, but the ecosystem that NVIDIA is building around its hardware. The logic is simple: whoever runs Nemotron almost certainly does so on NVIDIA graphics cards, fine-tunes it using its software tools, and deploys it on its software. Openness here is not charity, but a way to bring developers back to purchasing the company's hardware.
With a market capitalization exceeding $5 trillion, the cost of training Nemotron 3 Ultra, which ran into hundreds of millions of dollars, is a nearly negligible expense for the company. Graphics card sales more than cover the research, so NVIDIA can give the model away for free and still earn more than closed competitors with paid access.
The political context adds further weight to the release. An open American model can be inspected, modified, and run on private servers—making it attractive for countries building independent national AI, from Europe to Southeast Asia. Such a model cannot be remotely disabled, which is especially valuable in light of recent restrictions surrounding closed models.
Where the model falls short and what's next
Despite all its advantages, Nemotron 3 Ultra is not the smartest model on the market. In the independent Artificial Analysis Intelligence Index, it scored 48 points—the best result among open US models, but trailing leaders like Kimi K2.6 (54 points) and DeepSeek. Open models, according to analysts, lag behind closed ones by three to seven months.
But this gap matters less and less if an open model is simply sufficient for real-world tasks. A bank deploying Nemotron 3 Ultra to process loans on its own servers doesn't need flagship-level intelligence—it needs a model that can be fine-tuned on private data, kept within a secure perimeter, and not share confidential information with outsiders.
My expert opinion: NVIDIA's bet on efficiency rather than benchmark records may prove more farsighted. With mass AI adoption, the cost of running a model comes to the forefront, and one that is nearly as capable but five times cheaper wins in real-world deployment. The open ecosystem will only strengthen: NVIDIA has the resources, motivation, and distribution channels to release increasingly powerful open models faster than any other company.