Starlink in the Pacific and Baidu’s PR headache

Hello everyone, this is Cissy from Hong Kong.

Nearly three months after DeepSeek stunned Silicon Valley with its low AI inference costs, the start-up’s impact continues to grow in China, where it has become a source of national pride. Its technology is now ubiquitous, integrated into government agencies, cars, hospitals — even household appliances like air conditioners, refrigerators, TVs and vacuums.

There have also been attempts to trademark the start-up’s name. Last month, the National Intellectual Property Administration rejected 63 malicious trademark applications for “DeepSeek”, while the Companies Registry in Hong Kong recently issued a directive requiring at least four companies named “DeepSeek” to change their names. The directors and shareholders of these four “DeepSeek” companies are all from mainland China, and none of the directors are Liang Wenfeng, the founder of the real DeepSeek.

Tencent chair Pony Ma, who sat next to Liang Wenfeng when Chinese President Xi Jinping met with a group of prominent private entrepreneurs including billionaire Jack Ma last month, had high praise for the AI start-up on Wednesday. “By being the first to openly share its deep thinking, it has benefited not only many companies but also the development of state-owned large AI models,” Ma said in his company’s earnings call. “I believe this is a remarkable achievement. That’s why we deeply respect and wholeheartedly embrace the company.”

Ma’s comment is noteworthy for two reasons. One, some tech companies claim their AI models are open source but withhold key components from public access; two, he skilfully acknowledged DeepSeek’s contribution to state-owned models at a time when the government is betting on tech advances to boost the economy.

Indeed, there have even been reports that Beijing is extending its influence over the star start-up by asking staff not to leave China, as well as taking away passports of key employees. DeepSeek has not commented on these reports.

Reaching for the Starlink

As Starlink continues its rapid expansion around the globe, governments across the Pacific are responding with a mix of enthusiasm and caution to Elon Musk’s satellite internet provider. While many welcome the service for its ability to bring reliable, high-speed connections to some of the most remote places on Earth, the arrival of the global tech giant is also raising concerns over regulatory oversight, infrastructure investment risks and data security, write Nikkei Asia’s Shaun Turton and Rurika Imahashi.

Traditional providers, particularly those reliant on undersea fibre-optic cables, now face pressing questions about the sustainability of their business models. With Starlink offering faster deployment and competitive pricing, the pressure is mounting on incumbents to adapt or risk losing market share. At the same time, some governments also remain cautious about Starlink’s influence.

Cynthia Mehboob, a researcher at Australian National University, said Starlink’s spread was not without risks. “It speaks to how more and more power, when it comes to information technology, is now being concentrated into the hands of Big Tech. Implications for data security and privacy would become a big question, going forward, because you’ve got an entity like Musk with his whims, and you’ve got his very close connection to President Trump,” she said.

Alibaba’s AI overhaul

A little more than two years ago, Jack Ma and Alibaba were at a low ebb. The Chinese tech group’s founder had retreated from public view and relocated to Tokyo after falling foul of Beijing. Alibaba’s shares had plummeted, and the planned initial public offering of its fintech arm Ant Group had been scrapped.

Yet today, things look starkly different, write the Financial Times’ Zijing Wu and Eleanor Olcott. Ma has quietly presided over a strategic turnaround that investors are betting puts Alibaba in peak position to capitalise on an expected boom in artificial intelligence demand in China, which has been galvanised by the buzz around DeepSeek.

The FT interviewed more than two dozen current and former Alibaba employees, business partners, competitors and analysts, who described how the company executed its change in strategy. Key to this shift has been Ma’s decision to elevate his co-founder and closest ally Joe Tsai as chair and appoint Eddie Wu to lead the company’s AI-focused transformation.

Wu’s leadership has centralised decision-making, divested non-core assets and prioritised AI development as the key driver for future growth. This has seen Alibaba ramp up its investment in AI, backing start-ups, spending lavishly on chips and hiring researchers by the dozen. Alibaba’s Qwen, its family of large language models, is now considered a market leader in China and Apple has chosen it to run AI functions on iPhones in the country.

More fizz than funding

Bar charts showing Chinese venture capital AI and machine learning deals in number and value

Since the emergence of DeepSeek, Chinese AI has grabbed headlines and been the talk of tech bosses around the globe. However, venture investors in China seem to be in no hurry to ink deals as the country’s venture capital market is still in decline, writes Nikkei Asia’s Yifan Yu.

As of March 11, China had logged 100 deals in the AI and machine learning space, with a total value of $917.8mn. Although the quarter is not yet over, the pace of investment this year is well behind the $2.6bn from 196 deals seen in the first three months of 2024, according to data from PitchBook. Still, China venture investors led bets in the artificial intelligence and machine learning sector in Asia last year, despite an overall decline in the VC market since 2021.

One reason for the muted response from venture investors might be China’s unique AI investment ecosystem. Last year, the Biden administration finalised an investment ban on China that prevents US companies and individuals from investing in Chinese businesses related to AI semiconductors and quantum technologies.

Doxxing drama

One year after its former vice-president Qu Jing struck a public nerve by defending the tech industry’s gruelling work culture, Chinese search giant Baidu is embroiled in another public relations crisis. This time, it was triggered by another vice-president, or rather the vice-president’s teenage daughter, to be precise.

Earlier this week, Baidu vice-president Xie Guangjun’s daughter was caught repeatedly “doxxing” strangers amid a dispute over her favourite K-pop idol. This raised concerns that she may have accessed personal data through her father’s job and sparked calls to boycott Baidu’s cloud services over privacy fears, writes Nikkei Asia’s Cissy Zhou.

Baidu stated on Wednesday that the girl’s actions were unrelated to the company, claiming she obtained the data from a channel on an “overseas platform that collects private information through illegal means”. However, many social media users remain sceptical.

The timing could hardly have been worse for Baidu, which had just launched two new artificial intelligence models on Sunday, including a new reasoning-focused model that it says rivals DeepSeek’s model. The comment section on its post about the release was flooded with angry remarks from users asking if the company was leaking their personal data, forcing Baidu to ban comments at one point.