| Mobile Web

Search results for Devin

AI & Enterprise

Database firms push data tech to cut token waste by AI agents

As AI companies shift from flat-rate plans to usage-based billing, operating costs for AI agents are rising sharply. Database companies are pushing infrastructure solutions designed to reduce token consumption, The Register reported. Vector database firm Pinecone launched Nexus, a knowledge engine that pre-arranges data structure and context. TigerData, developer of TimescaleDB, introduced Ghost, a database platform for AI agents billed by compute time rather than database count.

AI & Enterprise

Even at high prices, users stay as monthly AI subscriptions reach 1,500,000 won

More users are spending hundreds of dollars a month on premium generative AI subscriptions such as Anthropic’s Claude Max, ChatGPT Pro and Google AI Ultra, treating them as work infrastructure rather than consumer services. Some pay more than $1,000 a month to combine token limits and strengths across models. Time savings is a key factor, though complaints about high prices and weaker performance are also growing, including around Anthropic’s Opus 4.7 model.

AI & Enterprise

Companies tighten AI budgets as model routing gains traction, pressuring OpenAI and Anthropic

U.S. companies are adopting “model routing” to cut AI spending by assigning different models to different tasks, moving routine work to cheaper, faster options and using costly models only for complex jobs. Executives and boards are tightening oversight as AI bills outstrip budgets. Cognition CEO Scott Wu said cost efficiency can improve sharply on repetitive work. Glean CEO Arvind Jain estimated 95 percent of enterprise AI usage still runs on expensive frontier models.