NVIDIA is giving away powerful AI for free: a genius move or hidden monopolization?
On June 4, 2026, NVIDIA released its largest AI model to the public—Nemotron 3 Ultra. The weights, training data, and methodologies were published under a free license. But don't get too excited about the "freebie": behind this generosity lies a well-thought-out strategy that brings the company more profit than any paid competitors.
Nemotron 3 Ultra was not created to break records in intelligence tests. Its goal is to be efficient, controllable, and open for customization. Unlike "black boxes" like ChatGPT or Claude, this model can be downloaded, fine-tuned on your own data, and run on your own infrastructure. The bet is on openness and control, not on maximum performance at any cost.
Hybrid Architecture: Mamba-2, Attention, and Latent MoE
At the core of Nemotron 3 Ultra lies an innovative hybrid architecture consisting of three components. Mamba-2 layers process long texts quickly and efficiently—their computational costs grow linearly, not exponentially like a standard attention mechanism. Attention layers, in turn, accurately retain large amounts of context in memory. And Latent MoE (Mixture of Experts) compresses data before passing it to "specialists," making each of their tasks narrow and precise without unnecessary computations.
The model has approximately 550 billion parameters in total, but only about 55 billion are used to process each token. This allows it to "think" like a giant system while behaving like a much more compact one in terms of cost. A context window of 1 million tokens and a speed of over 300 tokens per second provide 5-6 times greater throughput and roughly 30% lower task execution costs compared to alternatives.
NVIDIA's Strategy: A Free Model as the Key to Selling Hardware
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 own software. Openness here is not charity, but a way to bring developers back to purchasing the company's hardware.
NVIDIA can afford this. With a market capitalization of over $5 trillion, training Nemotron 3 Ultra, which likely cost hundreds of millions of dollars, is almost a negligible expense for the company. Graphics card sales more than cover research costs, so NVIDIA can give the model away for free and still earn more than closed competitors do with paid access.
The political context also adds weight to the release. An open American model can be inspected, modified, and run on your own servers—this makes it extremely attractive for countries building independent national AI, from Europe to Southeast Asia. No one can remotely disable such a model, 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 globally it lags behind leaders like Kimi K2.6 (54 points) and DeepSeek. Open models, according to analysts, trail closed ones by three to seven months.
However, 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 its own secure perimeter, and not share confidential information with outsiders.
My expert opinion: NVIDIA's bet on efficiency rather than test records may prove more far-sighted than it seems. With mass AI adoption, the cost of running a model comes to the forefront, and one that is almost as smart but five times cheaper wins in real-world operation. The open ecosystem will only grow stronger: NVIDIA has the resources, motivation, and distribution channels to release increasingly powerful open models faster than any other company. This is not charity—it's a new round of market monopolization, this time through open source.