Corporate America’s AI Bill Is Pushing It Toward Open Models

Corporate America’s AI Bill Is Pushing It Toward Open Models
AT&T spent the first years of the generative-AI boom using paid, closed systems from Anthropic and OpenAI for customer service, call transcription and coding. The appeal was capability; the trade-off was cost and little control over the underlying technology.

This year, as those costs climbed, the telecom group pivoted to models that can be downloaded and modified without payment or permission. By May, open models made up 20% of AT&T’s AI use; that figure has since reached 40% and could hit 60% within months, chief data and AI officer Andy Marcus said. The company has cut AI costs by as much as 80% compared with earlier this year. “We think it could be higher,” Marcus said of the open-model share.

AT&T is part of a wider corporate shift. Airbnb and Deloitte are also moving toward models that are cheaper and easier to tailor, while open models accounted for 58% of U.S. usage on OpenRouter last month, up from 10% a year earlier. Nvidia’s $12.9 billion purchase of the open-model library Hugging Face underscored how seriously the industry is taking that turn.

The economics are compelling, particularly as Chinese developers including DeepSeek, Moonshot AI and Alibaba release increasingly capable open systems. Atlas Cloud chief executive Jerry Tang said some can deliver 80% to 90% of the performance of OpenAI and Anthropic models at roughly 20% of the cost. “A Mazda can get you to your destination just fine, and a Maserati doesn’t have a lot of investment value,” he said.

Yet the switch is not absolute. Closed models still lead on demanding work such as coding and image or video generation, while open models are strongest in simpler, specialized jobs. And AT&T, wary of regulatory and data-privacy risks, is studying Chinese models but not deploying them; it is instead customizing U.S.-developed alternatives including Google’s Gemma and Meta’s Llama.

That leaves a broader political and commercial divide. Anthropic’s Dario Amodei argues for tight AI controls over national-security concerns, while Nvidia’s Jensen Huang and Meta’s Mark Zuckerberg frame openness as essential to a balanced AI ecosystem. Zuckerberg’s answer to foreign competition: “Our goal should be to make U.S. open-source models the best in the world.”

Continue reading https://foxvector.com/stories/01a07bad-98f6-3919-7368-36981fa77b75