Businesses Turn to Open-Weight AI Models to Rein In Soaring Costs
Tinder, PNC, CH Robinson and Siemens are among firms shifting queries to cheaper open-weight models as bills from OpenAI and Anthropic climb.
What to know
- Mentions of "open weight" or "open source" AI models in corporate earnings calls rose sixfold in Aug-Sept 2026 versus the same period a year earlier, per AlphaSense data cited by the FT.
- Adoption is spreading beyond tech firms to banking, logistics and industrials, driven by rising per-query costs as AI use shifts from chatbots to agents and to token-based billing.
- Chinese open-weight models offer lower costs but raise data-privacy concerns for some clients; building on open weights also carries higher upfront infrastructure and talent costs.
- Frontier labs like OpenAI and Anthropic still get used when top-tier capability is needed, suggesting a hybrid rather than wholesale replacement of proprietary models.
“The frontier models like OpenAI’s Astra and Claude Fable already have enough intelligence for 90% of the tasks we’re trying to do. If open-weights models catch up, I may not need to use them anymore.”
Vinay Kuruvila, CTO, Tinder · Financial Times (via PYMNTS) ↗ · Sep 26
Vinay Kuruvila Chief Technology Officer at TinderNvidia AI chipmakerPNC Financial Services BankCH Robinson Logistics companySiemens Industrial conglomerate
How it unfolded 2 developments, newest first · click a bar or a number to jump articles
-
2
FTBusinesses across sectors embrace open-weight AI amid cost surge
The Financial Times reported that companies including PNC Financial Services, CH Robinson and Siemens have discussed using open-weight models, and that references to "open weight" or "open source" AI in earnings calls jumped sixfold in August-September 2026 versus a year earlier.
“In January we were spending at the rate of $1 million per year and by July it had climbed to $10 million … I don't want another 10X increase.”
— Vinay Kuruvila -
first by PYMNTS, 1d ago
-
- 5 days quiet
-
1
BloombergDevelopers shift to open-weight AI, including Chinese models
Bloomberg reported that software companies are increasingly adopting open-weight AI models, including Chinese ones, to cut costs and reduce dependence on proprietary U.S. vendors, though some firms cite higher upfront costs, talent needs and data-privacy concerns with Chinese models.
-
1 outlet first by PYMNTS, 7d ago · read ↗
-
-
background
AT&T cuts AI coding costs 56% by routing to cheaper models — AT&T reported cutting costs of coding and other advanced AI tasks by as much as 56% using tools that route employee queries to cheaper models when appropriate, with only a 2% drop in performance quality; it aims to raise the share of queries handled by open-source models from 40% to 60-70%.
-
background
PYMNTS frames open-vs-closed AI as a cost tradeoff for CFOs — A PYMNTS report described the choice facing middle-market CFOs as whether savings, flexibility and control from open-weight models justify taking on more infrastructure responsibility versus paying for bundled proprietary platforms.
-
background
Rising AI costs tied to shift from chatbots to agents — An earlier report found that AI costs from large labs have climbed as usage shifts from chatbots to agents, which consume more computing power, and as labs move from flat subscriptions to token-based billing.