New York — Altimeter Capital CEO and tech investor Brad Gerstner is urging caution amid the excitement surrounding OpenAI’s new partnerships with AMD and Nvidia, warning that the headline-grabbing deals are “announcements, not deployments.”
Speaking to CNBC on Monday, Gerstner said the true test will be in execution.
“Now we will see what gets delivered,” he said. “Ultimately, the best chips will win.”
Compute Power Is the New Currency
The comments come as OpenAI continues to expand its infrastructure partnerships to fuel its next generation of large-scale AI models. The company’s recent megadeal with AMD adds to its deep ties with Nvidia, which already dominates the global market for AI processors.
Gerstner noted that the partnerships highlight one thing above all else — the global shortage of compute power.
“These deals provide more evidence that the world will remain compute-constrained despite best efforts to bring massive supply online,” he said.
The observation echoes what many in the AI industry have warned for months: as demand for AI workloads grows exponentially, even the biggest chipmakers will struggle to keep pace.
OpenAI’s expansion plans aren’t happening in a vacuum. The U.S. and China are locked in an AI arms race — one defined as much by hardware as by algorithms.
Chinese AI firm DeepSeek has become one of the most closely watched players after unveiling a lower-cost, open-sourced model that it claims rivals OpenAI’s flagship GPT models. Unlike OpenAI, DeepSeek builds its systems using domestically manufactured AI chips, bypassing U.S. export controls.
But the rise of DeepSeek has also raised security red flags in Washington.
A recent report by the National Institute of Standards and Technology’s (NIST) Center for AI Standards and Innovation warned that DeepSeek’s outputs tend to amplify Chinese Communist Party narratives more frequently than American AI models — prompting renewed scrutiny over AI’s geopolitical implications.
AMD Partnership Signals OpenAI’s Next Phase
OpenAI’s new partnership with AMD is being viewed as a strategic move to diversify supply chains and accelerate model development beyond Nvidia’s hardware ecosystem. AMD’s chips are expected to help train and deploy multimodal AI systems at lower cost and higher scale.
“We’re really seeing a world of compute scarcity because of surging demand for AI services — not just from OpenAI, but across the entire ecosystem,” said an OpenAI spokesperson on CNBC’s Squawk on the Street. “That’s why it’s so important for this industry to come together.”
With chip capacity and supply chains now at the heart of global competition, Gerstner’s words offer a pragmatic reminder: AI innovation will depend as much on silicon as on software.
