According to Fortune, China has substantially increased energy subsidies that could cut electricity bills by up to half for data centers using domestic chips from Huawei and Cambricon, specifically excluding facilities using foreign processors. In parallel, OpenAI announced a $38 billion, seven-year cloud computing deal with Amazon Web Services to train AI models and process ChatGPT queries, with AWS expecting to meet OpenAI’s capacity needs by late 2025. Meanwhile, Palantir reported blockbuster Q3 earnings with $1.2 billion revenue (up 63% year-over-year) and $476 million net income (up 40%), while CEO Alex Karp dismissed skeptics including short-seller Michael Burry. These developments signal an accelerating global AI infrastructure competition with profound implications.
Sponsored content — provided for informational and promotional purposes.
The Coming Subsidy Escalation
China’s targeted energy subsidies represent a new front in the AI arms race that goes beyond traditional industrial policy. By specifically favoring data centers using domestic chips, Beijing is creating a protected ecosystem where Chinese AI companies can train models at significantly lower operational costs. This isn’t just about catching up with Nvidia—it’s about building an entire stack where efficiency concerns become secondary to strategic autonomy. The Palantir earnings letter showing 121% growth in U.S. commercial business indicates American companies are already responding to this competitive pressure, but government support remains fragmented compared to China’s coordinated approach.
The Cloud Capacity Crunch
OpenAI’s $38 billion Amazon commitment—following even larger deals with Microsoft and Oracle—reveals the staggering infrastructure requirements of frontier AI models. We’re witnessing the emergence of a two-tier cloud market where a handful of AI giants command unprecedented negotiating power and capacity guarantees. AWS promising to meet OpenAI’s needs by end of 2025 suggests we’re 12-18 months from the next infrastructure bottleneck, as current data center construction timelines struggle to keep pace with AI compute demand. This creates a winner-take-most dynamic where only the best-capitalized AI companies can afford to train cutting-edge models.
The Great AI Valuation Divergence
Palantir’s explosive growth alongside Michael Burry’s short position represents the central tension in AI investing today. While Karp’s confidence reflects genuine commercial adoption, the skepticism highlights concerns about whether current valuations account for the massive capital expenditure required to stay competitive. The divergence between companies building foundational AI infrastructure (Amazon, Microsoft) and those applying AI (Palantir) will become more pronounced as compute costs escalate. We’re likely to see a shakeout where only companies with either massive scale or highly defensible niche applications survive the coming capital intensity wave.
Geopolitical Friction Ahead
The combination of China’s protectionist subsidies and U.S. companies’ massive cloud investments sets the stage for increased trade tensions. As AI becomes more central to economic competitiveness, we should expect retaliatory measures targeting cloud services and potentially export controls on AI-specific infrastructure components. The clean separation between “domestic” and “foreign” chips in China’s subsidy program suggests both sides are preparing for a fragmented global AI ecosystem. This balkanization will force multinational companies to maintain duplicate infrastructure and navigate increasingly complex regulatory environments.
Strategic Investment Implications
For investors, the key insight is that we’ve moved from the experimentation phase to the scaling phase of AI adoption. The companies winning now aren’t necessarily those with the best technology, but those with the capital and infrastructure to deploy at global scale. This favors established cloud providers and well-funded incumbents over pure-play AI startups. However, the subsidy dynamics create opportunities in companies positioned to benefit from government support programs in both the U.S. and China. The next 12-18 months will separate strategic investments from speculative bets as the true costs of AI scale become apparent.
