Anton du Plooy, Global Tech Sector Analyst, Ninety One, discusses how DeepSeek’s artificial intelligence (AI) breakthrough is challenging U.S. dominance, bypassing restrictions and reshaping global AI investment and geopolitics.
17 March 2024. DeepSeek, the AI lab affiliated with the Chinese quant hedge fund High-Flyer, has made waves with its recent release of a highly efficient large language model (LLM). Unlike previous Chinese AI models that have lagged Western competitors, DeepSeek’s models are delivering performance close to top-tier U.S. offerings, such as OpenAI’s GPT-4, at a fraction of the cost.
This breakthrough has reignited global discussions about AI dominance, technological independence, and the economic consequences of cutting-edge machine learning innovations. Given the backdrop of intensifying U.S.-China geopolitical tensions, DeepSeek’s advances carry far-reaching implications for the semiconductor industry, the broader AI ecosystem and geopolitics.
Market Reaction and Investment Turmoil
The immediate response to DeepSeek’s announcement was seismic. NVIDIA, the dominant force in AI computing hardware, saw its stock plummet, wiping out over $500 billion in market capitalisation. The reason? DeepSeek reportedly achieved comparable AI model performance while circumventing U.S. export restrictions on advanced NVIDIA Graphics Processing Unit (GPU) chips.
However, the initial panic was followed by a reassessment. It became clear that DeepSeek’s cost efficiencies were partly due to leveraging pre-existing frontier models via distillation techniques, rather than building an entirely novel foundation. Additionally, full training and operational costs were later estimated to be significantly higher than initial reports suggested.
The Geopolitical AI Chess Game: U.S. vs. China
DeepSeek’s success story is also one of resilience in the face of U.S. restrictions on AI chip exports to China. By employing H800 chips—procured before the latest sanctions—the company sidestepped trade limitations while proving that AI innovation in China isn’t wholly dependent on NVIDIA’s most powerful hardware.
This raises a key question: will the U.S. tighten restrictions even further? If history is any indication, policymakers in Washington may respond with additional curbs on chip exports, reinforcing a bifurcation of AI capabilities between the West and China. Yet, such restrictions often have unintended consequences—forcing China to accelerate its domestic semiconductor development and rethink its reliance on Western AI ecosystems.
DeepSeek also underscores the broader failure of U.S. containment policies. The Biden administration’s “small yard, high fence” approach to technology export controls aimed at restricting China’s rise as a technology superpower has not only fallen short but may have accelerated China’s AI advancements. Huawei’s resurgence, following similar sanctions, is a case in point. Initially crippled by U.S. blacklisting, Huawei rebounded through indigenous semiconductor innovation, proprietary software, and diversification into new sectors like electric vehicles. DeepSeek’s success follows a similar trajectory, reinforcing China’s broader strategic goal of AI leadership by 2030.
In addition, China’s approach to AI is deeply embedded in state-driven industrial policies, such as the “New Generation AI Development Plan” and “Made in China 2025.” Unlike Western counterparts that rely on private sector funding, China’s AI ambitions benefit from a “whole of government” strategy, integrating AI into industrial, economic, and national security domains. The U.S. faces an uncomfortable reality: its export controls may be accelerating, rather than hindering, China’s AI capabilities
DeepSeek’s Impact on AI Investment and Infrastructure
Despite initial concerns that DeepSeek’s cost efficiencies would lead to reduced spending on AI infrastructure, the opposite has occurred. Hyperscaler cloud providers, including major players in both the U.S. and China, have re-accelerated capital expenditures (CAPEX) on AI infrastructure. This aligns with Jevon’s Paradox—the economic principle that increasing production efficiency of a resource can drive up overall demand.
The Future of AI and the investment landscape
DeepSeek’s rise challenges the notion of US dominance, underscoring that AI development is no longer in the exclusive domain of Silicon Valley. As technological advancements continue despite geopolitical constraints, the AI sector is entering a more multipolar era—where open-source models, proprietary innovations, and geopolitical manoeuvring shape the competitive landscape.
For investors, the implications are significant. The falling cost of AI model development and deployment suggests the opportunity set is diversifying across industries, expanding the range of companies positioned to benefit. AI is no longer a U.S.-centric phenomenon— and accelerated adoption creates opportunities far beyond traditional infrastructure plays in this next phase of AI-driven transformation.
