DeepSeek's recent announcement of its V3 Large Language Model (LLM) has stirred significant discussions in the technology and finance sectors. Contrary to rumors, DeepSeek did not develop OpenAI for $5 million. Instead, the focus has shifted to the potential impacts of its V3 LLM on compute demand. Concerns have risen over a possible peak in spending on Graphics Processing Units (GPUs), igniting debates among analysts and investors about the future trajectory of the AI industry.
Analysts have weighed in on these developments with varying perspectives. Atif Malik from Citi highlights that while there could be challenges to the dominance of U.S. companies in advanced AI models, their access to more advanced chips provides a competitive edge. This factor, Malik suggests, will likely prevent leading AI companies from veering away from investing in cutting-edge GPUs.
"While the dominance of the US companies on the most advanced AI models could be potentially challenged, that said, we estimate that in an inevitably more restrictive environment, US' access to more advanced chips is an advantage. Thus, we don't expect leading AI companies would move away from more advanced GPUs." – Atif Malik
The introduction of DeepSeek's V3 LLM has fueled concerns regarding compute demand and potential constraints on GPU spending. However, analyst C.J. Muse views this as a bullish indicator for the industry. Muse argues that with Artificial General Intelligence (AGI) appearing closer to reality, the Jevons Paradox will likely result in heightened demand for compute power, rather than a reduction.
"Following release of DeepSeek's V3 LLM, there has been great angst as to the impact for compute demand, and therefore, fears of peak spending on GPUs. We think this view is farthest from the truth and that the announcement is actually very bullish with AGI seemingly closer to reality and Jevons Paradox almost certainly leading to the AI industry wanting more compute, not less." – C.J. Muse
Amidst these discussions, several analysts have reiterated their confidence in Nvidia's potential. C.J. Muse recommends buying Nvidia during any market weakness, while Srini Pajjuri has reaffirmed buy ratings on Nvidia, ASML, and Broadcom. Pajjuri advises against succumbing to doomsday scenarios and emphasizes that DeepSeek's innovations may lower model training costs, especially for U.S. hyperscalers who can capitalize on their GPU access advantage.
"If DeepSeek's innovations are adopted broadly, an argument can be made that model training costs could come down significantly even at U.S. hyperscalers, potentially raising questions about the need for 1-million XPU/GPU clusters as projected by some…: A more logical implication is that DeepSeek will drive even more urgency among U.S. hyperscalers to leverage their key advantage (access to GPUs) to distance themselves from cheaper alternatives." – Srini Pajjuri