Monday, November 23, 2026

"U.S. Companies Are Turning to Open AI Models to Cut Costs"

Input
2026-09-28 10:09:53
Updated
2026-09-28 10:09:53
"U.S. companies are turning to open AI models to reduce costs," reports said. Low-cost open-weight AI models from Chinese companies. Yonhap News

[Financial News] The Financial Times (FT) reported on the 27th, local time, that U.S. companies are turning to low-cost open-weight AI models as they seek alternatives to Anthropic and OpenAI to reduce their enormous cost burdens.
The trend is spreading throughout the U.S. industrial sector, beyond Silicon Valley, which is home to many IT companies.
According to research platform AlphaSense, the number of times executives mentioned open-weight and open-source models during corporate earnings calls and investor conferences in August and September surged sixfold from the same period a year earlier.
Companies are seeking open-weight models because of their cost competitiveness.
Open-weight models are a type of open model whose publicly available AI model weights can be downloaded directly by companies and customized to run on their own hardware.
They cost far less than paying to use premium versions of ChatGPT or Claude. Companies can also offer open-weight models for free to lock users into specific hardware or support advertising sales.
Share of tokens processed by open-weight models rose from 7% in December last year to 56% in August this year

Chinese companies such as DeepSeek are dominating this market. In the open-weight market, Chinese firms including DeepSeek and Z.ai are gaining influence. France's Mistral AI and U.S. companies such as NVIDIA, Reflection AI and Thinking Machines Lab have also released models using similar approaches, but FT reported that Chinese companies currently lead the market.
The adoption of open-weight models is spreading beyond the technology industry to telecommunications, finance, logistics and manufacturing. AT&T currently handles about 40% of its AI workloads with open models and plans to raise that share to 70% within the next year.
AT&T explained that the cost savings are substantial because it processes 45 billion tokens a day. By fine-tuning open models with its own data, the company can achieve performance comparable to or better than that of closed models for certain tasks.
Dating app Tinder is also shifting some tasks to open-weight models because of the cost burden. Its annualized AI spending rose tenfold, from about $1 million in January to $10 million in July.
Security, in addition to cost, is cited as a reason for choosing open-weight models. Operating models on internal corporate infrastructure allows companies to process customer information and sensitive internal data without sending it to external AI providers.
The shift is expected to create new competitive pressure for OpenAI and Anthropic. Although a significant portion of corporate AI spending still goes toward the two companies' cutting-edge models, the rapid improvement of less expensive open-weight models could reduce the need for companies to use high-priced models.
FT reported that companies are increasingly reducing AI costs by using high-performance models for complex tasks and less expensive open-weight models for repetitive, large-scale workloads.
OpenAI and Anthropic join price-cutting competition by releasing lower-priced models

As a result, OpenAI and Anthropic have entered the competition by urgently releasing lower-priced models.
In fact, open-weight models accounted for 7% of all tokens processed through AI developer platform Vercel's AI Gateway in December last year, but that share surged to 56% in August this year.
Vinay Kurvira, Tinder's chief technology officer, said, "Our annual AI spending soared from $1 million in January to $10 million in July, so we began shifting some inquiries from nontechnical users to open models," adding, "If the performance of open-weight models catches up with high-performance frontier models such as OpenAI's Astra and Anthropic's Claude Fable, there will be no reason to use expensive closed models."
Andy Marcus, AT&T's chief data and AI officer, also said, "We currently process 40% of our AI workloads with open models and plan to increase that figure to 70% within a year."
FT pointed out that data sovereignty and security concerns, in addition to cost savings, are important factors pushing companies toward open models.
Scott Wallace, a senior director at global data center operator Digital Realty, said, "We never input customer data or sensitive, complex internal confidential information into external frontier models," explaining, "We resolved the security concerns by building an internal chat interface on private infrastructure based on an open-weight model."
FT assessed, "The rapid progress of low-cost AI systems could threaten the revenues of OpenAI and Anthropic, which have driven the data center boom with massive investments."
[email protected] Lee Seok-woo, international affairs specialist Reporter