AMD CEO Lisa Su Rejects AI Bubble Theory Amidst Industry Concerns

At the 2025 WIRED “Big Interview” event in San Francisco, Lisa Su, the CEO of AMD, rejected the speculations of concerns about the “AI bubble.” When she was asked whether the current AI boom was a bubble, she replied, “emphatically, from my perspective, no.”

Her remarks come as artificial intelligence continues to attract massive investment, raising alarm among industry specialists. “AI bubble” critics argue that the current AI boom might mirror previous episodes like the “dotcom bubble” of the late 1990s, marked by irrational exuberance, inflated valuations, and a stark disconnection between hype and real returns. Most notably, a senior market analyst noted, the AI is a bubble 17 times higher than the dotcom one.

A bursting bubble could jeopardize not just high-flying startup valuations but the broader ecosystem, from chipmakers to data-centre operators and cloud providers. Even Sundar Pichai, Google’s CEO, warned against the bubble bursting, noting, “No company is going to be immune, including us.”

Against such a backdrop, Su’s blunt rejection of bubble warnings signals both confidence in long-term hardware demand and a counter-narrative to growing market skepticism. Her view is quite similar to AMD rival NVIDIA CEO Jensen Huang’s rejection of the AI bubble theory.

Su’s Rejection of an AI Bubble: Rationale, Market Stakes, and Counterpoints

Su framed her view around the premise of demand for AI chips needed for AI training, and inference remains structural and far from speculative. She argued that AI adoption is accelerating across multiple sectors and that hardware will remain in demand for years. 

The AI-infrastructure boom is visible in the recent performance of both AMD and NVIDIA. Both companies' share values have risen since 2019. NVIDIA’s stock price increased from around $3 in 2019 to over $180, while AMD’s stock price rose from 2019’s $30-40 to roughly $215. Both companies have secured long-term deals from OpenAI in 2025. Additionally, AMD will start shipping chips to China again, accelerating the company’s current momentum.

NVIDIA’s chief, Jensen Huang, also argued that workloads are migrating from CPUs to GPU-based infrastructure and that AI demand reflects a long-term tech transformation. Despite AMD and NVIDIA’s claims, representatives of multiple tech giants, such as IBM's CEO Arvind Krishna and Ex-Intel's CEO Pat Gilsenger, raise concerns. Additionally, Krishna pointed out that only 20%-30% of companies claiming the AI market is booming might achieve the success they want, while the rest will face the backlash. 

On the other hand, a 2025 case study using a “Capability Realization Rate” (CRR) model suggests that many companies commanding high AI-related valuations have yet to deliver proportionate performance. However, whether businesses using AI can truly match the possibilities they are marketing to customers and whether broader acceptance is feasible in the long run would determine stability in the AI infrastructure market.