The U.S. administration’s decision to permit Nvidia to export its advanced H200 artificial intelligence accelerators to approved customers in China marks a consequential moment for the global technology and investment ecosystem, according to Nigel Green, CEO of international financial advisory and asset management firm deVere Group.
The approval opens a controlled channel for one of Nvidia’s highest-performance AI chips—hardware that has until now remained restricted due to national security considerations. The policy shift, analysts say, has implications that stretch far beyond semiconductor revenues and into the structural evolution of global AI competitiveness.
“This move changes how capital markets should think about future AI leadership, competitive dynamics, and long-term value creation across sectors,” Green says. “It alters the speed and scale at which AI capability can spread. It matters for investors far beyond the chipmakers themselves.”
The H200 is built for large-scale training and inference, supporting trillion-parameter models and advanced multimodal systems. Its availability in China reduces a constraint that has shaped the global AI landscape for nearly two years. Limited access to top-tier accelerators pushed Chinese developers to pursue alternative strategies: algorithmic efficiency, large-scale dataset utilisation, and hardware-aware optimisation.
Over the past year, China’s AI sector demonstrated resilience under these limitations. Breakthroughs such as DeepSeek—trained on constrained hardware like Nvidia’s H20—showed that competitive AI performance could be achieved through optimisation rather than raw compute power.
“DeepSeek proved that hardware limits did not stop progress,” Green notes. “They simply forced a different approach.”
With the introduction of H200-class computing, those constraints begin to ease. For investors, Green argues, the core story is acceleration: shorter development timelines, greater iteration speed, and more direct competition with global platforms.
In the near term, markets are expected to focus on earnings impact. Expanded access to China could provide a revenue uplift for semiconductor vendors and adjacent sectors linked to data center buildouts, cloud infrastructure, and industrial automation.
However, the medium- and long-term outlook introduces more complexity. “Greater availability of advanced compute increases the number of serious AI competitors across industries such as autonomous vehicles, advanced manufacturing, logistics optimisation, healthcare analytics, and defence-linked technologies,” Green says. “When AI capability broadens, competitive advantages compress. It changes how investors price leaders versus challengers.”
Green cautions against interpreting the decision through a narrow geopolitical lens. Instead, he argues, it reflects the economics of AI: “Whoever can combine compute, data and capital most effectively will advance fastest.”
China brings unique characteristics to that equation, including massive domestic datasets, rapid scaling capacity, and a willingness to prioritise capability even at lower efficiency. These dynamics may produce different competitive trajectories compared with U.S. and European markets.
Over longer horizons, Green says investors must factor in how wider diffusion of AI capability affects valuation frameworks. As AI becomes a baseline enabler of productivity, automation, and service scalability, the differentiators shift.
“When more players can access similar tools, excess returns shrink. Markets then reward execution rather than exclusivity,” he adds.
This creates new considerations for portfolio construction. Concentration risk increases if investors assume a narrow cluster of dominant AI winners. Broader access to advanced compute could expand the field of competitive companies.
“Investors need to prepare for a market where AI leadership is contested, not assumed,” Green says.
While some policymakers argue that export controls safeguard national advantage, and others contend they hinder innovation, Green maintains that investors should focus on practical outcomes.
“The approval of H200 exports accelerates the transition already under way: AI becoming a widely deployable industrial technology rather than a tightly held advantage,” he concludes.
