Sam Altman’s description of Chinese AI progress as “remarkable” underscores how the contest for artificial general intelligence has evolved into a full-stack race spanning chips, models, infrastructure, and monetization.
The progress of Chinese technology companies across the artificial intelligence stack is “remarkable,” OpenAI Chief Executive Sam Altman said in an interview with CNBC, offering a candid assessment of a rivalry that now stretches from semiconductor fabrication to large language models and mass deployment.
Altman said the pace of technological advance in “many fields,” including AI, is “amazingly fast.” In some areas, he noted, Chinese firms are “near the frontier,” while in others they lag behind U.S. counterparts. The distinction is significant: it suggests that while American firms still dominate certain layers of the stack — particularly advanced GPU design — Chinese players are closing the gap in applications, model optimization, and system-level integration.
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026).
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab.
The broader context is the accelerating race toward artificial general intelligence (AGI), a theoretical milestone at which AI systems can perform most economically valuable tasks at the human level or beyond. Both the United States and China view leadership in AGI as strategically consequential, not only commercially but geopolitically. The competition is therefore not confined to software breakthroughs; it encompasses chip supply chains, energy capacity, cloud infrastructure, and capital mobilization.
At the hardware layer, China has intensified efforts to build domestic semiconductor capabilities capable of competing with global leaders such as Nvidia. U.S. export controls have restricted the sale of certain advanced AI chips and semiconductor manufacturing equipment to Chinese firms, prompting Beijing to accelerate support for homegrown alternatives. The strategy includes scaling local chip designers and investing heavily in fabrication capacity, even as performance gaps remain at the cutting edge.
The financial markets have responded to the policy push. Shares of Chinese AI-linked companies have rallied on domestic exchanges as investors bet on long-term state backing and expanding internal demand. China’s vast digital economy, combined with strong government coordination, provides an environment where AI systems can be rapidly deployed across e-commerce, logistics, surveillance, finance, and manufacturing.
Altman’s remarks also echo concerns voiced by other U.S. executives. Brad Smith, president of Microsoft, told CNBC that American technology companies should “worry a little bit” about the subsidies Chinese competitors receive from their government in the AI race. That comment highlights a structural asymmetry: while U.S. firms rely largely on private capital and market-driven incentives, Chinese firms often benefit from direct state support, industrial policy alignment, and preferential financing.
The contest, analysts say, is effectively a full-stack competition. At the base lies semiconductor design and fabrication. Above that sit cloud infrastructure providers that assemble compute clusters and manage data center operations. On top are foundational model developers such as OpenAI, and finally, the application layer that integrates AI into enterprise workflows and consumer platforms. Gains in one layer can compound advantages in others.
The strategic landscape is intertwined with OpenAI’s own capital needs. According to data from Dealroom, investors have ploughed around $70 billion into the company. Sources told CNBC that OpenAI is seeking to close a $100 billion fundraising round, potentially one of the largest private raises in technology history. Such capital is necessary to finance model training, infrastructure partnerships, and global expansion.
The economics of advanced AI remain demanding. Training frontier models requires massive clusters of GPUs, extensive electricity consumption, and sophisticated cooling systems. Inference — serving millions of user queries — generates ongoing operational costs. The ability to sustain rapid growth, therefore, hinges on achieving “reasonable unit economics,” as Altman described it.
“We are growing at an extremely fast rate right now,” he said. “I think as long as we can have reasonable unit economics, we should focus on continuing to grow faster and faster, and we’ll get profitable when we think we when we think it makes sense.”
That stance signals that OpenAI is prioritizing scale over immediate profitability. Rapid user adoption can create network effects, attract enterprise customers, and justify infrastructure investments. However, sustained losses would eventually test investor patience, especially given the magnitude of capital deployed.
One emerging revenue lever is advertising within ChatGPT. Altman said OpenAI is still determining the optimal format.
“I think we still have some work to do to figure out the exact ad format that’s going to work best,” he said, noting that plans remain at an early stage.
He cited “Instagram style ads where you discover something new that you might really like and otherwise wouldn’t have known about” as a model he personally favors, adding that OpenAI has “a real opportunity to push in that direction with ads in ChatGPT.”
The advertising concept marks a potential strategic shift. Until now, OpenAI’s primary revenue streams have included subscription tiers such as ChatGPT Plus, enterprise licensing agreements, and API usage by developers. Ads could introduce a consumer monetization layer similar to social media platforms, though integrating commercial messages into conversational AI presents design, trust, and regulatory considerations.
OpenAI plans to test adverts first in the United States before expanding to other markets, Altman said. The approach suggests a phased rollout aimed at refining user experience while minimizing backlash. The success of such experiments may influence how conversational AI platforms balance commercial imperatives with user expectations.
Meanwhile, China’s AI ecosystem continues to expand across applications. Large domestic platforms are embedding generative AI into search, e-commerce, and productivity tools. State policy has also encouraged AI integration in manufacturing and public services. While U.S. firms currently lead in certain frontier model benchmarks, China’s scale advantage in deployment could generate rapid feedback loops that enhance model performance and user adoption.
The geopolitical dimension has added complexity as export controls, supply chain constraints, and regulatory scrutiny have introduced friction into cross-border technology flows. If AI development fragments into parallel ecosystems — one centered on U.S.-allied supply chains and another on China’s domestic stack — interoperability and standards may diverge.
Altman’s acknowledgment of China’s momentum reflects a more nuanced view emerging in Silicon Valley. Rather than dismissing Chinese efforts, U.S. executives are increasingly recognizing a credible, well-funded competitor operating across multiple layers of the AI value chain.
Against this backdrop, OpenAI’s immediate priority remains scaling usage and infrastructure while securing fresh capital. However, the longer-term question of when and how profitability emerges remains open.