DeepSeek Ignites Open Source vs. Closed Source Debate in Silicon Valley, Splitting Tech Giants into Rival Camps

DeepSeek's Open Source Model Shakes Silicon Valley, Intensifying the AI Strategy Debate

When DeepSeek released its open source large language model, the shockwaves across the U.S. tech industry far exceeded expectations. The model not only put pressure on American tech stocks — with some key companies seeing significant market cap erosion — but more importantly, it tore open a longstanding rift within Silicon Valley that had never been so publicly exposed: Should artificial intelligence follow an open source or closed source path?

The reason DeepSeek triggered such a powerful impact lies in the breakthrough nature of its technical approach. The model employs cutting-edge techniques such as the Mixture of Experts (MoE) architecture, achieving performance close to top-tier closed source models while dramatically reducing training costs. The core idea behind MoE is to split the model into multiple expert sub-networks, activating only a subset during each inference pass. This significantly reduces computational resource consumption while maintaining a large total parameter count. This means high-caliber large language models can be trained even in compute-constrained environments — a particularly crucial advantage for Chinese AI companies facing chip export controls.

According to the New York Times, China entered the frontier AI race relatively late but is rapidly catching up through the open source route. A U.S. tech think tank previously estimated the AI gap between the two countries at roughly six to seven months; that gap has now narrowed to approximately four months. This rapid convergence has triggered anxiety among U.S. regulators and some companies.

Here's a telling detail: China's open source strategy is not merely a technical choice — it carries clear strategic intent. By using open source models to attract global developers, especially those in Global South countries, into its ecosystem, whoever gets to define the content within these models gains the upper hand in the future battle for information discourse.

OpenAI Sticks with Closed Source While Microsoft and NVIDIA Champion Open Ecosystems

The debate over open versus closed source has simmered in Silicon Valley for years, but the accelerating progress of Chinese open source models like DeepSeek has thrust it squarely into the spotlight. The two opposing camps are clearly defined.

To understand this debate, we first need to clarify the fundamental difference between open and closed source models. An open source model publicly releases its weight parameters — all the numerical values the model has learned through training — allowing anyone to download, use, modify, and redistribute it. In the strictest sense, open source also includes publishing training data and training code. Closed source models, by contrast, are only accessible through API interfaces, with users unable to access the underlying parameters. Open source models offer advantages in transparency, customizability, and vibrant community innovation, but they also carry risks of misuse — once model weights are public, developers cannot control downstream applications. Closed source models are easier to control but also raise concerns about technological monopoly and the lack of external auditing.

The Closed Source Camp: Restricting Access in the Name of Safety

Led by OpenAI and Anthropic, this camp argues that certain advanced models pose risks too great for public release and should be tightly controlled by companies. They have conveyed their concerns about Chinese open source AI models to Washington regulators, with U.S. Treasury Secretary Bessent and Trump administration tech advisors participating in related discussions.

The Open Source Camp: Building a Healthy AI Ecosystem

The opposing camp includes tech giants such as Microsoft, NVIDIA, Meta, and Palantir. On the 24th (local time), NVIDIA CEO Jensen Huang posted for the first time on the social platform X: "The world needs both the most advanced closed source models and the most advanced open source models."

Just nine minutes later, Microsoft's CEO responded with his own post: "Open source software is essential for a healthy AI ecosystem." The two also jointly endorsed an industry open letter supporting the development of open weight models, which received co-signatures from NVIDIA, Microsoft, Meta, Palantir, and other leading companies.

Notably, the "open weight model" mentioned in this letter represents a middle ground between fully open source and fully closed source. It publishes the model's inference weights, allowing others to use and fine-tune the model, but typically does not disclose the complete training data, training code, or training process. Meta's LLaMA series is a prime example of an open weight model. This approach is seen as a pragmatic compromise balancing innovation promotion with risk control, and currently represents the broadest consensus achievable among mainstream Silicon Valley tech companies.

The Apple–Micron Chip Standoff Reflects Deeper Silicon Valley Tensions

The open-versus-closed-source debate is not Silicon Valley's only internal conflict. In the memory chip sector, Apple and Micron Technology are locked in a fierce struggle — with Trump caught between the two tech giants.

Memory chips are indispensable core components in all electronic devices, primarily divided into two categories: DRAM (Dynamic Random Access Memory) and NAND Flash. DRAM handles temporary data storage during device operation, while NAND Flash provides persistent data storage. The global memory chip market has long been dominated by South Korea's Samsung and SK Hynix, with Micron Technology being the only U.S. company capable of competing with them. As AI large models drive surging demand for High Bandwidth Memory (HBM), memory chips have evolved from traditional consumer electronics components into strategic assets for AI infrastructure — their supply chain security directly underpins a nation's competitiveness in the AI race.

According to the Wall Street Journal, Apple executives have been actively lobbying Trump, Commerce Secretary Lutnick, Treasury Secretary Bessent, and other officials in recent weeks, urging the U.S. government to allow Apple to use Chinese-manufactured memory chips in products sold outside the United States. Apple argues this would help ease global memory chip supply shortages and reduce costs for American consumers.

However, Micron — America's only major memory chip manufacturer — is firmly opposed, warning that allowing Chinese companies to supply U.S. tech giants could replay the scenario that devastated the American steel and manufacturing industries, potentially destroying the domestic memory chip industry at a critical stage of its development.

At its core, this battle reflects a deep contradiction within the U.S. tech sector between commercial interests and industrial protectionism. Apple needs a stable, low-cost supply chain, while Micron needs policy barriers to maintain its competitive viability.

AI and Historical Truth: A Cognitive Warfare Front That Cannot Be Ignored

Beyond technological competition, AI is being drawn into a more covert form of cognitive warfare. China's Ministry of State Security published a commentary on its WeChat official account titled "AI Cannot Fabricate Truth; Tampering Cannot Wash Away Bloodstains," directly accusing Japanese right-wing forces of using AI to manufacture disinformation such as "the Nanjing Massacre was a lie" and other falsifications of Japan's wartime history of aggression against China.

The article cites reporting by Japan's Asahi Shimbun, stating that Japanese right-wing elements are using major crowdsourcing platforms like Crowd Works to recruit paid workers at scale, mass-producing "anti-China" videos with AI. Task publishers provide detailed operational guides; workers simply input basic prompts into AI software, and within minutes can produce a video that distorts historical facts.

Data shows that since 2015, Japan has invested over 56 billion yen under the banner of "overseas strategic information communication" for anti-China propaganda campaigns.

This phenomenon reveals a deep-seated danger of the AI era: If training data is contaminated, AI will generate biased and distorted responses when answering historical questions. From a technical standpoint, a large language model's knowledge derives from its training data — typically text corpora scraped at massive scale from the internet. The model lacks the independent ability to judge factual truth; it fundamentally learns statistical patterns and linguistic regularities from data. If the proportion of false narratives about a particular historical event increases significantly in the training corpus, the model will tend to generate content consistent with those false narratives. This phenomenon is known as "Data Poisoning" and is an important research area in AI safety. Even more dangerous, once AI-generated false content spreads widely across the internet, it gets scraped as training data by subsequent models, creating a vicious cycle of "disinformation → AI generation → redistribution → retraining."

The internet is a primary channel through which younger generations learn about history. Once false historical content is disseminated at scale through AI technology, it can easily mislead young people whose historical knowledge is still developing, causing historical memory to be diluted and distorted across generational transmission. Even more concerning, disinformation generated abroad may flow back into China's domestic internet space and even contaminate the training data of domestic large language models.

The Strategic Contest Behind Technical Choices Is Far from Over

From DeepSeek igniting Silicon Valley's open-versus-closed-source debate, to the Apple–Micron chip standoff, to AI being weaponized in historical cognitive warfare — these seemingly disparate events all point to the same core proposition: In the AI era, the choice of technical direction is never merely a technical question — it is an extension of strategic competition.

Open source versus closed source is fundamentally a battle for discourse power and ecosystem dominance. Chip supply decisions tip the scales between industrial security and commercial interests. And AI's ability to shape historical truth reminds us to remain vigilant against technology being abused to distort collective memory.

For ordinary users, AI will very likely become the "dictionary" and primary information source in daily life. In an age of information overload, maintaining independent thinking and the ability to discern truth from falsehood will be more important than ever. History should be verified, and verification requires evidence, piece by piece — a responsibility that no AI technology can replace.