Tag
Gartner predicts that AI inference costs per agentic workflow will increase more than fivefold by 2028, driven by efficiency gains that enable more powerful models and applications, paradoxically raising overall costs.
Gergely Orosz criticizes Gartner's AI reports as pay-to-play, noting that AWS and Microsoft pay for favorable rankings while Anthropic, OpenAI, and Cursor do not, leading to skewed results like AWS being ranked above Anthropic and OpenAI omitted.
The author reflects on Gartner's prediction that 40% of agentic AI projects will be canceled by 2027, emphasizing that the real failure is not model incompetence but quiet failures in production due to bad data or API issues, and that most teams measure single task completion rather than reliability over hundreds of runs.
VentureBeat survey finds 71% of enterprises admit that most of their so-called AI agents are actually simple single-prompt wrappers, not true multi-step workflows, making the term 'agent' diluted and adoption statistics unreliable.
Gartner forecasts that AI servers will consume more power than all conventional data center hardware combined by 2027, with global data center electricity consumption set to grow 26% this year and reach over 1,200 TWh by 2030.
Gartner named OpenAI as an Emerging Leader in its 2025 Innovation Guide for Generative AI Model Providers, recognizing the company's progress in supporting over 1 million companies deploying AI safely at scale. The recognition reflects OpenAI's enterprise momentum, including 9x year-over-year growth in ChatGPT Enterprise seats and strong customer adoption across major organizations.