Real AI Challenge Lies in Computing Power Shortage, Not a Bubble, States Daniel Newman: 'People Claiming There's a Bubble Fail to Grasp the Reality'

AI Market Faces Compute Power Shortage, Not a Bubble, Says Futurum CEO

During his address at CES 2026, Daniel Newman, CEO of Futurum Group, challenged the notion that artificial intelligence is experiencing a speculative bubble. Instead, he emphasized a more pressing issue: a significant worldwide deficit in the computing resources required to support what he describes as a “multi-decade super-cycle” for AI advancement.

The Infrastructure Challenge

In a recent interview shared on his X account, Newman argued that those warning of an AI bubble are missing the bigger picture. He believes the current excitement is only the beginning of a much larger technological transformation.

Newman pointed out that the emergence of advanced “agentic” AI—systems capable of carrying out complex, independent tasks—will soon push the demand for computing power to unsustainable heights.

“We simply don’t have the infrastructure to meet today’s needs, let alone tomorrow’s,” Newman cautioned. He forecasts that as demand accelerates over the next five to ten years, the tech industry will confront a major hardware bottleneck.

2026: A Turning Point for Enterprise AI

Newman identified 2026 as a pivotal year when enterprise-level AI will become prominent, shifting the focus from consumer-facing chatbots to tangible business returns and improved ROI for corporations.

He observed that only a fraction of the available trained data is currently being used, with most valuable datasets still locked within proprietary systems for applications like drug development, manufacturing, and supply chain management.

This trend reflects a broader industry movement from the costly “build phase” of AI model training to the “monetization phase,” where AI begins to deliver measurable efficiency gains and profit growth.

Scaling Up Efficiency

To highlight the scale of AI operations, Newman referenced Google’s Gemini AI, which now processes an astonishing 10 trillion tokens every day.

On a practical note, he shared that his own company has leveraged AI to reduce market research timelines from six months to just two weeks.

“We’re still at the very beginning,” Newman concluded, stressing that the era of robust AI profitability and longevity is just starting, despite ongoing skepticism.

AI-Focused ETFs to Watch

For those interested in investing in the AI sector, here are several ETFs linked to AI technologies and their recent performance:

Note: This article was created in part with the assistance of AI and reviewed by Benzinga’s editorial team.

Image credit: Shutterstock