Descriptions:
JoAnne Feeney of Advisors Capital Management joins Bloomberg Tech to argue that the recent AI chip stock selloff—triggered by Kimi K3’s release and fears that more efficient AI models would reduce hardware demand—represents a buying opportunity rather than a structural break. Feeney has long-held positions in Nvidia and Broadcom for clients and contends that these companies’ competitive advantages are more durable than those of the model layer.
Her central thesis: falling inference costs drive higher usage volume, which in turn creates more demand for data center infrastructure, Nvidia GPUs, and Broadcom networking chips. Even as foundation model providers face margin compression from competition, the infrastructure layer benefits from rising overall workload volume. She argues that data center operators and chip suppliers also offer differentiated value—security, reliability, stability, and data handling—that model providers cannot easily replicate.
Feeney flags the memory sector as a specific risk, warning that Micron’s low valuation (approximately 6x forward earnings at the time of recording) reflects a market that understands the HBM price increases driving recent profits are not sustainable. With three leading-edge memory manufacturers simultaneously expanding capacity, she anticipates the classic semiconductor cycle of oversupply and price compression will eventually reassert itself. The conversation provides a grounded, historically-informed counterweight to both the extreme bull and bear narratives dominating AI chip sentiment in the wake of Kimi K3’s market debut.
📺 Source: Bloomberg Tech · Published July 21, 2026
🏷️ Format: Interview
