DeepSeek Momentum Slows Amid Chip Controversy

DeepSeek emerged as a major disruptor in the artificial intelligence sector in early 2025. The Chinese company launched an assistant capable of competing with ChatGPT but at a significantly lower cost. The release caused a massive ripple effect, leading to a loss of over $1 trillion in market value for major American technology firms. This event was dubbed the "momento Sputnik," drawing parallels to the surprise of the 1957 satellite launch.

Despite this initial success, DeepSeek is currently facing significant headwinds. The company has delayed the release of its next-generation model, DeepSeek R2, which was originally expected in May. Furthermore, allegations surfaced in December suggesting the company utilized restricted Nvidia Blackwell chips for training, despite US export bans. While DeepSeek remains a low-cost alternative to Western AI, its growth has slowed, and it has failed to maintain the search dominance it achieved shortly after its launch.

DeepSeek entered the market with significant momentum. The application launched on January 20, and within a week, it occupied the top spot on the Apple App Store rankings. The tool was viewed as a capable rival to ChatGPT, offering similar performance metrics but with a drastically reduced price tag.

According to the company's developers, the total cost to build DeepSeek was only US$ 5.6 million. This stands in stark contrast to the resources deployed by American competitors; OpenAI reportedly invested US$ 5 billion in ChatGPT during 2024. Despite this efficient engineering, DeepSeek has struggled to retain the same level of user interest. Data from Google Trends indicates that while search interest spiked immediately after the launch, the company could not sustain the momentum. Today, it has not reached the same level of interest as ChatGPT.

One of the primary factors contributing to DeepSeek's slowing momentum is the delay of its next major release. The DeepSeek R2 model, the successor to the current assistant, was scheduled for a public release in May. However, the launch was postponed.

The delay was reportedly driven by internal quality control. Liang Wenfeng, the CEO of DeepSeek, was said to be unsatisfied with the model's performance. Additionally, the company faces structural challenges compared to its rivals. DeepSeek operates with a smaller infrastructure, particularly regarding the hardware used to train these complex systems. This limitation may be contributing to the slower development cycle compared to larger American firms.

DeepSeek has also become entangled in the geopolitical battle over advanced technology. In early December, reports alleged that the company was using the Nvidia Blackwell chip for AI training. The United States has banned the export of this specific chip to China due to the ongoing AI race.

Nvidia responded to these allegations by stating they had no proof their chips were being diverted. A company representative stated, "Although this smuggling seems unlikely, we investigate any information we receive." In a related development, US President Donald Trump indicated he would authorize Nvidia to sell the H200 chip to China. The H200 is an older model, roughly 18 months behind the most advanced technology, but sales would be permitted only for "approved clients" under strict national security conditions.

The launch of DeepSeek is widely regarded as a pivotal moment in the AI industry. The event caused American tech giants to lose more than $1 trillion in market value, a shock compared to the 1957 launch of the Soviet satellite Sputnik. However, the question of "which AI is best" remains open. While DeepSeek proved that high-performance models can be built cheaply, its inability to release the R2 on time and the scrutiny over its hardware supply chain have stalled its progress. The company must now navigate strict US regulations and internal performance standards to regain its early 2025 form.