Open-Weight AI Models Surge as Enterprises Seek Control and Cost Savings
The era of proprietary AI dominance may be waning as open-weight models race ahead, according to analysts and enterprise leaders. These free-to-download, customizable alternatives to large language models (LLMs) are gaining traction due to the flexibility, governance, and cost benefits they offer.
“It’s almost like these blank canvases are available now and then you can paint it on your own,” said Deepak Seth, senior director analyst at Gartner. “You don’t have to make the canvas itself. So you’re not starting from scratch, even when you’re building your own model.”
Open models—sometimes called open-weight models—allow IT leaders to fine-tune AI to their specific needs without vendor lock-in. Analysts say this addresses two critical pain points: visibility into internal AI use and economic control.
The shift is accelerating as more practical applications emerge. “Open source is more flexible and can be used in ways that proprietary models … in some cases can’t be trusted to operate,” said Jesse Williams, cofounder and COO at Jozu, an AI tooling company. Still, he stressed that proprietary models like ChatGPT and Gemini remain popular and aren’t slowing down.
Recent outages at closed-model providers such as Anthropic and OpenAI have further pushed CIOs to consider open alternatives for resiliency. “It is still early in the AI race,” said Max Goss, senior research director at Gartner, adding, “CIOs do need to be mindful of what they’re putting AI to work for and what is the alternative.”
Popular open models include Meta’s Llama, Mistral, DeepSeek, and Minimax. Even proprietary vendors are releasing open versions: Google’s Gemma, OpenAI’s GPT-OSS, and Microsoft’s Phi. However, these lightweights are trained on smaller datasets and may not match the intelligence of full-scale LLMs.
“What is this model good at? You have to figure that out. None of them are truly general purpose models,” said Max Leaming, head of data science and AI solutions at ManpowerGroup. Enterprises must experiment to find the right fit, he added.
Companies like ServiceNow, Microsoft, HubSpot, and RWS argue that open models integrate more easily into existing AI infrastructures, lower computing costs, and support agentic AI workflows. This combination is driving adoption across industries.
Background
Proprietary LLMs like OpenAI’s ChatGPT and Google Gemini have dominated headlines, but open-weight models have quietly gained ground. The concept is similar to Linux: free to download, tweak, and deploy. Analysts say this gives enterprises unprecedented control over AI governance and economics.
The shift is not a backlash against proprietary models, Williams noted, but rather a recognition that different use cases demand different approaches. Open models now fill gaps that closed systems cannot safely or cost-effectively address.
What This Means
For CIOs and IT leaders, the rise of open-weight models signals a strategic choice: they can now build custom AI solutions without depending on a single vendor. This reduces risk of lock-in and outage-related disruptions.
However, open models require more upfront experimentation and may lack the breadth of training data that proprietary systems offer. Enterprises must balance flexibility with the need for intelligent, general-purpose tools. The AI race is far from over, and open models are now a key player in the ecosystem.
— Reporting contributed by industry analysts and enterprise practitioners.