Jensen Huang Embraces Chinese Open-Source AI While Jim Cramer Waves the Red Flag
Nvidia's CEO calls models like Kimi K3 'excellent' and says good open-source AI should be used regardless of origin—putting him at odds with a longtime bull.
What matters
- Jensen Huang told Axios that Chinese open-source models like Kimi K3 are 'excellent' and should be used regardless of origin.
- Huang dismissed the idea that strong Chinese models threaten U.S. labs like OpenAI and Anthropic.
- Jim Cramer, a longtime Nvidia bull, takes a protectionist stance and views Chinese AI as inherently compromised.
- The split mirrors a broader industry divide between open-access evangelists and geopolitical hawks.
- Cramer previously praised Huang's Computex keynote and Nvidia's trillion-dollar AI infrastructure roadmap.
What happened
The release of a powerful open-source AI model from China has reignited a familiar debate in the tech industry: should good AI be embraced regardless of its country of origin, or should foreign models—particularly Chinese ones—be treated as suspect?
Nvidia CEO Jensen Huang landed firmly in the first camp. In an interview with Axios, Huang said plainly that Chinese models like the recently released Kimi K3 are "excellent" and that "open-source models that are excellent should be used." He also dismissed the notion that strong models from China pose a threat to U.S. labs such as OpenAI and Anthropic, reportedly arguing there is no scenario where Chinese AI success comes at the expense of American innovation.
That puts Huang on the opposite side of CNBC's Jim Cramer—a notable split given Cramer's long-standing enthusiasm for Nvidia as a stock. Cramer, host of Mad Money, has been a vocal China hardliner, refusing to accept that anything coming from the country could be free of compromise. The Gizmodo report characterizes Cramer as someone who has "never once been deterred by being both loud and wrong," and notes the irony of his position given how frequently he has urged viewers to buy Nvidia shares.
The two men are not strangers. In a March 2025 CNBC interview, Cramer enthusiastically interviewed Huang, calling him "the leader" in artificial intelligence and marveling at Nvidia's trillion-dollar infrastructure roadmap. More recently, in June 2026, Cramer praised Huang's Computex keynote, saying it revealed that the AI infrastructure boom is creating winners beyond Nvidia and that Huang's argument that "compute is revenue" was a "tide-turner" for skeptical investors.
Why it matters
This isn't just a personality clash. The disagreement reflects a deeper fault line in how the U.S. tech industry and investment community should respond to the globalization of AI—particularly the rise of capable open-source models from China.
Huang's position is pragmatic: if a model is good and open, developers and enterprises should be free to use it. That stance aligns with Nvidia's broader interest in driving AI adoption broadly, since more AI usage—on any platform—ultimately increases demand for compute infrastructure, much of which runs on Nvidia hardware.
Cramer's protectionist stance, by contrast, reflects a geopolitical anxiety shared by many U.S. policymakers: that Chinese AI models could serve as vectors for surveillance, data exfiltration, or strategic influence. The tension between these two views will likely shape everything from enterprise procurement decisions to future export-control policy.
For developers and companies building AI products, the question is immediate: should you integrate a Chinese open-source model like Kimi K3 if it performs well, or steer clear on security and compliance grounds?
What to watch
- Whether U.S. enterprises begin adopting Chinese open-source models like Kimi K3 at scale, or whether compliance and security concerns keep them away.
- Any policy response from Washington, including potential restrictions on the use of foreign open-source AI models in government or regulated industries.
- Whether Cramer's skepticism shifts if Chinese models continue to close the performance gap with Western frontier models.
- Nvidia's own positioning: Huang's openness to Chinese models is consistent with Nvidia's hardware-first business model, but it could draw political scrutiny if U.S.-China tech tensions escalate.
What to do next
Developers
Evaluate Kimi K3 or similar Chinese open-source models against your current stack using standardized benchmarks, while documenting provenance and licensing terms.
Huang's endorsement suggests these models are technically competitive, but developers need to weigh performance against compliance and security considerations.
Founders
Establish a clear internal policy on whether your company will use foreign open-source AI models, and communicate it to investors and customers.
The Huang-Cramer split signals that this is becoming a board-level question, not just a technical one.
PMs
Map which product features could benefit from cheaper open-source models and assess the regulatory risk of using Chinese-origin AI in your target markets.
Open-source models may reduce costs, but geopolitical risk could affect go-to-market in regulated or government-adjacent segments.
Investors
Monitor whether enterprise adoption of Chinese open-source models accelerates or stalls, and track any U.S. policy responses that could restrict their use.
Huang's openness and Cramer's skepticism represent two investable theses: broad AI adoption vs. geopolitical containment.
Operators
Run a security and data-handling audit on any Chinese open-source models under evaluation before integrating them into production workflows.
Even if a model performs well, Cramer's concerns reflect real enterprise compliance requirements around foreign-sourced software.
Testing notes
Caveats
- This story concerns an industry debate and executive commentary rather than a testable product or model release. While Kimi K3 is referenced as a released model, the sources do not provide sufficient technical detail (access URL, API, or benchmark scores) to construct reliable testing instructions.