China Open-Source, Compute Arms Race, Reordering Global Trade | BG2 w/ Bill Gurley and Brad Gerstner

The conversation highlights the evolving landscape of global trade, particularly focusing on the US's trade agreements and tariffs. The US has established favorable terms, such as a 0% tariff on goods going to Europe, while Japan commits $550 billion to the US without imposing tariffs. This trade strategy, perceived as high risk but potentially rewarding, has led to unexpected positive outcomes, with multiple deals secured contrary to predictions of trade wars.

The discussion shifts to technology and regulation, emphasizing the US's need to maintain leadership in AI against China's rapid advancements. Concerns are raised about US regulations potentially stifling progress in technology sectors. China is making significant strides in open source AI models, with several high-quality providers emerging, while American models are losing traction. Key factors contributing to China's success include a flexible stance on intellectual property, the collaborative nature of Chinese developers, and the rapid development of competitive models.

Brad and Sunny discuss the evolution of AI model development, noting a shift from compressing vast data to utilizing real-time information retrieval. Sunny credits OpenAI for pioneering reasoning models, which have improved problem-solving capabilities. Bill acknowledges the validation of his arguments regarding the advantages of open source for China, aligning with his predictions about the success of Chinese models.

The historical context of China's open source movement is explored, emerging partly in response to global IP theft accusations. Open source software has gained traction in China, where the belief in sharing ideas for greater prosperity contrasts with traditional IP protection. The role of the Chinese government in promoting open source initiatives is debated, with indications that competitive dynamics among emerging AI companies are driving rapid innovation.

Sunny notes the growing demand for Chinese open source models, which offer comparable quality to American models at a significantly lower cost. A chart analysis illustrates the relationship between intelligence and cost, showing that Chinese open source companies cluster in the high intelligence, low-cost quadrant, providing substantial value in the AI market. Anticipation is high for OpenAI's upcoming open-source model, which could significantly impact the market if it delivers a competitive product.

Recent developments in compute demand are underscored by notable figures like Elon Musk and Sam Altman, who highlight the increasing need for compute resources. The scale of upcoming compute clusters is expected to be much larger than previously anticipated, with a shift towards inference time reasoning and agent-to-agent interaction. Bill mentions the unprecedented scale of current fundraising in the tech industry, particularly for companies like Anthropic, which recently raised $5 billion.

The competitive landscape in AI is rapidly evolving, with the rise of models from China and the impact of open-source developments leading to commoditization of the model layer. This shift is expected to lower intrinsic value in intelligence and operating layers, pushing the focus towards building consumer applications. Despite lower margins, the market is transforming quickly, allowing multiple players to thrive.

The conversation shifts to tariffs and their implications for global trade. Bill seeks Brad's insights on market reactions to tariffs, noting initial nervousness during "Liberation Day" due to unpredictability. Brad observes that the market's interpretation of tariffs has evolved, reaching all-time highs despite previous predictions of negative impacts. While the consensus among economists suggests tariffs will harm consumers, a minority argues that exporting countries like China will absorb the costs to prevent layoffs.

Concerns about inflation persist, particularly as CorePC shows signs of rising. While some consumers may face higher prices due to tariffs, the overall economic situation is better than critics expected, with GDP rebounding to 3%. A key objective of the trade strategy is to bolster domestic production and enhance supply chain resilience, particularly in critical industries. The US-centric supply chain, especially in chips, still relies on small components from other manufacturers, prompting companies to establish onshore operations in response to tariffs.

Brad highlights the complexity of the US-China relationship, which encompasses trade, national security, and the AI race. Ongoing discussions with China are crucial, given the significant trade relationships with the EU and China, alongside a notable goods trade deficit. The potential for a major deal with China is anticipated, addressing tariffs and military cooperation, while current tariffs are expected to remain at or above 15%. There is a recognition that China must pivot towards domestic consumption rather than relying solely on exports.