Competition regulation

US investigates DeepSeek for potential AI chip violations

The US Commerce Department is investigating whether DeepSeek, the Chinese AI company that recently launched a high-performing assistant, has been using US chips in violation of export restrictions. These chips are prohibited from being shipped to China, raising concerns about DeepSeek’s rapid rise in the AI sector. Within days of launching, its app became the most downloaded on Apple’s App Store, contributing to a significant drop in US tech stocks, which lost around $1 trillion in value.

The US has imposed strict limits on the export of advanced AI chips to China, particularly those made by Nvidia. These restrictions aim to prevent China from accessing the most sophisticated AI processors. However, reports suggest that AI chip smuggling from countries like Malaysia, Singapore, and the UAE may be circumventing these measures. DeepSeek has admitted to using Nvidia’s H800 chips, which were legally purchased in 2023, but it is unclear whether it has used other restricted components.

The controversy deepened when Anthropic’s CEO Dario Amodei commented that DeepSeek’s AI chip fleet likely includes both legal and smuggled chips, some of which were shipped before restrictions were fully enforced. While DeepSeek has claimed to use only the less powerful H20 chips, which are still permitted to be sold to China, the investigation continues whether these practices undermine US efforts to limit China’s access to cutting-edge AI technologies.

DeepSeek’s use of Nvidia chips raises security concerns in Washington

Pressure is mounting on the Trump administration to impose stricter export controls on Nvidia’s AI chips following concerns over China’s DeepSeek. A bipartisan call from lawmakers John Moolenaar and Raja Krishnamoorthi urges officials to assess the national security risks linked to the firm’s reliance on advanced US technology.

A letter addressed to National Security Advisor Michael Waltz highlights that DeepSeek’s latest AI model has extensively used Nvidia’s H20 chips, which remain outside current export restrictions. The request forms part of a wider review into US export policies in response to strategic threats.

Concerns in Washington are growing over China’s rapid AI advancements, with the US House of Representatives warning against using DeepSeek’s technology. Officials fear AI could bolster cyber threats or facilitate bioweapons development, leading to increasing efforts to limit China’s access to cutting-edge chips.

Nvidia has stated that its products fully comply with US regulations and expressed willingness to work with authorities. Reports indicate that Trump’s administration is already considering new restrictions on H20 chip exports to China as part of its evolving AI security strategy.

French privacy watchdog to investigate DeepSeek AI over data concerns

France‘s data protection authority, the CNIL, announced it will question DeepSeek to better understand how the Chinese company’s AI system operates and assess potential privacy risks for users. The move comes as European regulators intensify scrutiny of AI, following concerns raised by Italy and Ireland over DeepSeek’s handling of personal data.

DeepSeek recently gained international attention after revealing its latest AI model, DeepSeek-V3, was trained using less than $6 million worth of Nvidia H800 computing power. European authorities have been particularly vigilant about data protection, with the EU’s General Data Protection Regulation (GDPR) setting stringent standards for privacy. Under GDPR, violations can result in fines of up to 4% of a company’s global turnover.

The CNIL’s investigation follows a broader European push to regulate AI technology, with new rules imposing strict transparency obligations on high-risk AI systems. Other countries, such as Italy and Ireland, have also launched inquiries into DeepSeek, reflecting growing concerns over AI’s implications for data privacy across the continent.

Chinese startup DeepSeek challenges AI spending norms

Chinese startup DeepSeek has claimed to build an advanced AI model at a fraction of the cost and time required by US tech giants, raising questions about the industry’s spending. The company stated it developed its V3 AI model in just two months for under $6 million, using Nvidia’s less-advanced H800 chips. An app powered by this model became the most downloaded app on the US iPhone store on Monday.

Founded in 2023, DeepSeek claims its AI models rival or outperform top competitors despite significantly lower costs, challenging the notion that scaling AI demands massive resources. This comes as US tech firms, part of the “Magnificent Seven,” have collectively added $10 trillion in market value since the AI boom sparked by ChatGPT in late 2022.

Industry analysts, however, remain sceptical of DeepSeek’s claims. Bernstein‘s Stacy Rasgon questioned whether the startup’s advancements could outpace those from established global AI labs, suggesting the company’s narrative might be overstated. The developments have drawn attention as investors scrutinise AI spending ahead of earnings reports from key industry players.

Italy blocks DeepSeek chatbot over privacy concerns

Italy’s data protection authority, the Garante, has ordered the Chinese AI startup DeepSeek to block its chatbot in the country, citing insufficient responses to queries about its privacy policy. The watchdog had requested detailed information on data collection practices, sources, purposes, and storage, particularly concerning whether user data is stored in China. DeepSeek’s failure to adequately address these concerns prompted the Garante to impose an immediate ban and launch an investigation.

DeepSeek had removed its AI assistant from Italian app stores earlier this week but claimed it was not subject to local regulation. Agostino Ghiglia, a member of the Garante’s board, stated that the company’s stance worsened its position. Italian users who had already downloaded the app still reported access to the chatbot, while the web version remains operational. The Garante emphasised that European citizens must have clear consent and data protection guarantees, especially regarding servers located in China.

The Garante’s action highlights growing scrutiny of AI platforms in Europe, with data regulators in Ireland and France also questioning DeepSeek’s privacy practices. Italy‘s proactive approach has drawn attention; the country temporarily banned ChatGPT in 2023 over similar concerns. DeepSeek has positioned its AI as a cost-effective alternative to US models, surpassing ChatGPT as the top-rated app on Apple’s US App Store. However, its refusal to cooperate with European regulators may jeopardise its expansion.

Government of Taiwan weighs response to potential US chip tariffs

Taiwan’s government is assessing whether to assist its industry in response to possible US tariffs on semiconductors, Premier Cho Jung-tai confirmed. Authorities are closely monitoring recent developments after former President Donald Trump pledged to impose tariffs to encourage domestic production in the US.

Taiwan, home to leading chipmaker TSMC, plays a crucial role in global technology supply chains. Cho emphasised Taiwan’s strategic position in the industry and pledged continued efforts to strengthen external cooperation and maintain technological leadership. Economy Minister Kuo Jyh-huei downplayed concerns, suggesting any tariff impact would be minimal due to Taiwan’s technological edge.

During Trump’s previous administration, TSMC committed to a $12 billion factory in Arizona, later expanding the investment to $65 billion. The latest tariff proposal follows his recent directive for US agencies to investigate trade deficits and alleged currency manipulation, which could present further challenges for Taiwan.

Taiwan’s trade surplus with the US surged 83% last year, reaching a record $111.4 billion, driven by high-tech exports, particularly semiconductors. TSMC has declined to comment on the potential tariffs, while the government continues to evaluate its response.

Italy suspends DeepSeek AI app amid data protection concerns

The Chinese AI app DeepSeek was removed from Apple and Google app stores in Italy on Wednesday, following a request by the country’s data protection authority for information on its handling of personal data. Italy’s Garante regulator gave DeepSeek 20 days to clarify what data it collects, its sources, purposes, and whether it is stored in China. Concerns over safeguarding underage users, potential bias, and risks of electoral interference were also highlighted by Garante chief Pasquale Stanzione.

The app, which recently surpassed ChatGPT in downloads from Apple’s App Store, remains functional for Italian users who had already installed it. It is also still available in other European Union countries and the UK. Ireland‘s Data Protection Commission has also sought details about DeepSeek’s data processing practices for Irish users, while Germany‘s government has voiced concerns about potential AI-driven election interference ahead of its February vote.

Italy’s Garante is known for its proactive stance on AI regulations, having temporarily banned ChatGPT in 2023 over alleged breaches of EU privacy laws. DeepSeek, which touts itself as a cost-efficient alternative to U.S. AI services, has faced mounting scrutiny as it gains popularity. Meanwhile, Irish regulators noted that DeepSeek has not designated Ireland as its EU headquarters, complicating oversight under EU data protection rules.

Google appeals EU’s record antitrust fine

Google has appealed to the EU’s top court to overturn a record 4.3-billion-euro antitrust fine imposed seven years ago, arguing that the penalty punished the company for its innovation. The fine was originally levied by the European Commission, which accused Google of using its Android operating system to suppress competition by forcing manufacturers to pre-install Google Search, Chrome, and the Google Play store on devices. While the fine was later reduced to 4.1 billion euros by a lower court, Google maintains that its actions fostered competition, not hindered it.

During Tuesday’s hearing, Google lawyer Alfonso Lamadrid stated that the Commission failed to meet its legal obligations and relied on errors in law. Lamadrid defended Google’s agreements with phone manufacturers, insisting they were not anti-competitive, but rather beneficial to the market. The case centres on whether the European Commission acted appropriately in its investigation and decision to reshape markets through such penalties.

The judges of the Luxembourg-based Court of Justice of the European Union will make a final ruling in the coming months, with no further opportunity for appeal. In addition to this case, Google remains under scrutiny by EU regulators for its advertising business, with another major decision expected later this year.

OpenAI faces legal action from Indian news companies

Several prominent Indian media outlets, including those owned by billionaires Gautam Adani and Mukesh Ambani, are taking legal action against OpenAI. These outlets, such as NDTV and Network18, along with organisations like the Indian Express and Hindustan Times, have filed to join an ongoing lawsuit against OpenAI in a New Delhi court. They allege that OpenAI has been improperly scraping their copyrighted content to train its AI model, ChatGPT, without permission or payment.

The legal claim, which is being led by the Digital News Publishers Association (DNPA), argues that OpenAI’s practices pose a significant threat to the copyrights of its members. The publishers claim that OpenAI’s actions amount to ‘wilful scraping’ and the use of their work for commercial gain, especially as the company generates revenue through ads linked to AI-generated content. This lawsuit highlights broader concerns in the media industry about the influence of large tech companies on content distribution and monetisation.

The legal proceedings are part of a larger global trend, with authors, musicians, and news organisations worldwide suing AI firms for using their works without compensation. In the US, the New York Times has filed a similar lawsuit against OpenAI and its major backer, Microsoft. This new case in India adds significant pressure to OpenAI, which has denied the allegations, arguing that its AI systems rely on publicly available data and that deleting such data could violate US law.

The Indian plaintiffs argue that OpenAI’s failure to strike content-sharing deals with local publishers, while it has done so with international media outlets, undermines the business of Indian news companies. The publishers warn that OpenAI’s practices could weaken the media landscape and negatively impact democracy, calling for greater protection of intellectual property in the age of AI.

Ads to launch on Threads platform

Meta has begun testing advertisements on its Threads platform in the US and Japan, targeting a small group of users with image ads in their home feeds. The trial comes as the platform surpasses 300 million monthly active users. Businesses will have the opportunity to extend their existing Meta campaigns to Threads, with the company closely monitoring the tests before a wider rollout.

Advertisers will also benefit from a new inventory filter powered by AI, enabling control over the type of content their ads appear alongside. Analysts suggest that while Threads is still a minor player in Meta’s overall revenue strategy, growing uncertainty around TikTok has led brands to explore alternative platforms.

Launched in July 2023 as a competitor to X, formerly known as Twitter, Threads continues to attract users following X’s controversial changes under Elon Musk. Meta’s plans to expand its AI infrastructure with a $65 billion investment this year further highlight its ambitions to remain competitive with tech giants such as OpenAI and Google.

While Threads is not expected to contribute significantly to Meta’s revenue by 2025, its integration into Meta’s broader ad ecosystem demonstrates the company’s efforts to capitalise on the platform’s growing popularity.