Tech stocks tumble on concerns over AI spending
https://www.bbc.com/news/articles/c802gvxpyxzoJust one article of many out there
This wasn’t all that hard to predict. It can do cute things, but it’s not living up to the claims and returns.
Another explination
Studies indicate that 95% of companies deploying AI have struggled to see measurable returns. The shortfall largely stems from massive computing costs, poor workflow integration, and a gap between generalized chatbots and actual production needs.
Why AI Returns Are Lacking Soaring Infrastructure Costs:
AI is proving to be far more expensive than many companies anticipated, with some businesses burning through yearly AI budgets well ahead of schedule
The "Treadmill" Effect:
Many out-of-the-box tools provide an initial productivity boost but fail to scale or adapt to company-specific workflows, leading to abandoned deployments.
Workflow Congestion:
Simply increasing the output of a single worker often causes bottlenecks if the rest of the organization's processes cannot accommodate the surge in data or analysis.
How Organizations Are Adapting
To actually generate value, the strategy is shifting from broad, expensive experimentation to targeted efficiency:
Model Routing:
Companies are routing easier tasks to cheaper, open-source models rather than relying exclusively on expensive frontier models, saving resources for complex tasks.
Operational Redesign:
AI is most effective when it is deeply embedded into how decisions are governed and made, requiring companies to redesign their operating models.
