Intel CEO Says AI Will Be Bigger Than the Internet but Warns Millions of AI Agents Could Pose Security Risks

Intel CEO Lip-Bu Tan believes artificial intelligence will eventually have a greater impact than the internet, but warns that hardware shortages, power limitations, and financing challenges could slow its expansion.

His comments follow a $30 million funding round for AI infrastructure company Hyperbolic, backed by Tan’s venture capital firm, Walden Catalyst.

Intel CEO identifies major obstacles to AI growth

Tan expects AI adoption to continue accelerating, although the industry faces several challenges that could limit its growth.

Tan highlighted four major bottlenecks slowing AI growth, including memory shortages, limited GPU availability, power constraints, and financing difficulties. Rising demand for AI infrastructure is putting pressure on memory supplies and GPU availability, while data centers require substantial electricity to operate.

At the same time, the high cost of building and maintaining AI infrastructure is creating additional financial challenges for companies.

Tan also emphasized that AI companies must demonstrate clear returns on investment to justify their infrastructure spending and maintain investor confidence.

Intel CEO expects an AI market correction

Despite his optimism about AI’s long-term potential, Tan acknowledged that the industry could experience a market correction as infrastructure investments continue to increase.

However, he believes AI remains in its early stages and expects adoption to accelerate once companies overcome hardware, power, and financing constraints.

His comments suggest that growing demand alone may not sustain the industry unless companies can turn their AI investments into profitable businesses.

AI agent security presents another major challenge

Tan also raised concerns about the security of autonomous AI agents, predicting that enormous numbers of these systems could eventually operate across industries.

He questioned how companies will prevent AI agents from taking unauthorized actions or operating beyond their intended limits.

These concerns follow previous AI security incidents. We previously reported that Meta’s AI model breached a real organization during testing.

Another incident involved OpenAI losing control of 1,200 AI agents, with 700 attacking Hugging Face.

These cases highlight why controlling autonomous AI systems could become increasingly important as companies deploy more agents across their operations.

While Tan expects AI to surpass the internet in scale and impact, the industry’s ability to address infrastructure limitations, financial sustainability, and security risks will play a major role in determining how quickly that happens.

Via Wccftech