25/04/2026

BIZ & FINANCE SATURDAY | APR 25, 2026

14

DeepSeek rolls out long-awaited AI model

Xpeng plans full-scale ‘flying car’

deliveries by 2027 BEIJING: Chinese electric vehicle (EV) maker Xpeng expects to start large scale production of its “flying” cars next year and of its humanoid robots in the fourth quarter of 2026, president Brian Gu told Reuters on Thursday. Gu also said there is “tremendous potential” to increase cooperation with German automaker Volkswagen, which last month began mass production of its first EV model, developed jointly with Xpeng. “There are a lot of areas that we can partner and really provide value to each other,“ Gu said, adding that Xpeng was also open to partnerships with other automakers. “We need to be nimble and willing to partner with different players in different regions.” Xpeng has received more than 7,000 orders for its flying cars, the majority of which are in China, where the company is working on obtaining approval from the country’s aviation authorities. The automaker expects to start making early deliveries of the flying cars late this year, according to Gu. Speaking to Reuters ahead of the Beijing Auto Show, Gu said the company would start robotaxi tests in the southern Chinese city of Guangzhou this year and that 2027 will be a “critical year” for “tests around the world with partners.” He said the company will likely produce hundreds to thousands of robotaxis over the next 12 to 18 months. Gu said its humanoid robots will initially be used as receptionists or in sales to interact with customers. He said that within the next 10 to 20 years, Xpeng’s robot business should be larger than its automotive division as “there will be more use cases for humanoid robots in our lives.” Like many other Chinese automakers, Xpeng has expanded overseas and currently operates in about 60 countries outside China. Gu said that last year, Xpeng generated about 10% of its sales volume and around 15% of its revenue from overseas sales. He added that in the next five to 10 years, “more than 50% of the revenue should come from outside of China.” – Reuters

o New V4 platform cuts compute costs and expands capabilities, marking a turning point for large-scale artificial intelligence deployment BEIJING: Chinese startup DeepSeek released a new artificial intelligence (AI) model with “drastically reduced” costs yesterday, more than a year after it stunned the world with a low-cost reasoning model that matched the capabilities of US rivals. The AI race has intensified the rivalry between China and the US, and the White House on Thursday accused Chinese entities of a massive effort to steal artificial intelligence technology. Hangzhou-based DeepSeek burst onto the scene in January last year with a generative AI chatbot, powered by its R1 reasoning model, that upended assumptions of US dominance in the strategic sector. The new version, DeepSeek-V4, “features an ultra-long context of one million words”, the company said in a statement on social media platform WeChat, hailing it as “world-leading with drastically reduced compute (and) memory costs” in a separate announcement on X. The model’s context length, which determines how much input a model is able to absorb to help it complete tasks, “(achieves) leadership in both domestic and open-source fields across agent capabilities, world knowledge, and reasoning performance”, the WeChat statement said. A “preview version” of the open source model is now available, the company said. Experts say V4’s release marks an “inflection point” in terms of hardware and cost. “This addresses the long-standing issues of slower performance and higher costs associated with long context lengths, marking a genuine inflection point for the industry,“ Zhang Yi, the founder of tech research firm iiMedia, told AFP. “For end users, this will bring widespread, accessible benefits. For instance, if ultra-long context support becomes a standard feature, long-text processing is expected to move beyond high-end research labs and enter mainstream commercial applications,“ he said. The new V4 is released as two versions,

Chinese startup DeepSeek released a new AI model yesterday, more than a year after it stunned the world with a low-cost chatbot that matched the capabilities of US rivals. - AFPPIX

crackdown. At home, DeepSeek’s AI tools have been widely adopted by Chinese municipalities and healthcare institutions as well as the financial sector and other businesses. This has been partly driven by DeepSeek’s decision to make its systems open source, with their inner workings public – in contrast to the proprietary models sold by OpenAI and other Western rivals. But the White House has accused Chinese firms of vying to “steal” American technology, ahead of an expected summit between Donald Trump and Xi Jinping in Beijing next month. “The US has evidence that foreign entities, primarily in China, are running industrial-scale distillation campaigns to steal American AI,“ Trump’s science and technology chief adviser Michael Kratsios said in a post on X. Distillation is a common practice within AI development, often used by companies to create cheaper, smaller versions of their own models. DeepSeek’s announcement yesterday also came as Meta said it planned to cut a tenth of its staff as it looks for productivity gains from the rest of the workforce while investing heavily in artificial intelligence. Reports said Microsoft was also looking to trim its ranks. – AFP

DeepSeek-V4-Pro and DeepSeek-V4-Flash, with the latter being “a more efficient and economical choice” because it has smaller parameters. V4-Pro has 1.6 trillion parameters while the V4-Flash has 284 billion parameters, which refine models’ decision-making ability. The model has also been “optimised” for popular AI Agent products such as Claude Code, OpenClaw, OpenCode and CodeBuddy, the DeepSeek statement said. “In world knowledge benchmarks, DeepSeek-V4-Pro significantly leads other open source models and is only slightly outperformed by the top-tier closed-source model (Google’s) Gemini-Pro-3.1,“ the statement added. Last year’s so-called “DeepSeek shock” sparked a sell-off of AI-related shares and a reckoning on business strategy in what was also described as a “Sputnik moment” for the industry. The chatbot performed at a similar level to ChatGPT and other top American offerings, but the company said it had taken significantly less computing power to develop. However, its sudden popularity raised questions over data privacy and censorship, with the chatbot often refusing to answer questions on sensitive topics such as the 1989 Tiananmen “Your clients are not Chinese. This product is not available in China,“ therefore setting up shop in Singapore increases the chance of selling to the US, Goh said of Topview, which has received over US$8 million (RM32 million) in Kamet investment since 2024. US President Donald Trump brought Sino-US tech rivalry to the forefront during his first term with talk of security risks, leaving tech firms to steer through tit-for-tat measures that intensified in his second term with the spread of AI. Complicating matters, tech firms have had to contend with Trump’s overhaul of H-1B visas for highly skilled workers, shaking firms that routinely brought or sent workers to the US. Such developments have bolstered Singapore’s ambition to be the most AI-powered

Singapore seen as neutral hub in Sino-US artificial intelligence rivalry SINGAPORE: Singapore is transforming from East-West gateway to neutral ground for the AI sector, with Chinese startups hoping to operate beyond government reach while US firms seek foreign talent without the headache of stricter visa regulations. establish AI video business Topview in Singapore in anticipation of international clients’ wariness of Chinese government oversight. economy, with initiatives like a visa for AI talent and tax breaks for registering intellectual property. Its “ecosystem enablers” have attracted investment from firms of different sizes and geographies, said an Economic Development Board spokesman. developer Addepar and note-taking device maker Plaud AI, with legal platform Harvey AI joining in June.

US AI developer Anthropic, whose US$30 billion fundraising was led by Singapore sovereign wealth fund GIC, plans to open a Singapore office, three people familiar with the matter said, joining heavyweights OpenAI, Superintelligence Labs from Meta and DeepMind from Alphabet’s Google. Anthropic declined to comment. While Singapore intensifies efforts to attract AI talent, the superpowers are working to protect technology. The US has prevented Nvidia from selling market-leading AI chips to China and has blocked access to chipmaking equipment. China imposed a travel ban on founders of AI startup Manus, the Financial Times reported, after Manus shifted to Singapore from China last year and was bought by US tech giant Meta. – Reuters

The city-state, a longtime darling for its business-friendly stance and bilingual population, is increasingly being seen as a place to keep both China and the US at bay, rather than acting as a bridge, as the superpowers vie for technological superiority through avenues such as export and talent control. Setting up in Singapore“gives a lot of comfort” to international clients that a startup’s intellectual property is on the island and therefore not subject to controls from either China or the US, said Kerry Goh, CEO of Kamet Capital. Goh advised two former executives of Chinese tech leader Alibaba seeking investment to

They have also attracted companies from China hoping to distance themselves from political preconceptions by appearing more Singaporean, and from the US seeking engineers without visa hurdles, industry executives and analysts said. “Singapore is increasingly becoming a neutral hub for AI companies from both the US and China,“ said Brad Gastwirth, global head of research at Circular Technology. AI companies in Singapore with ties to – or parents in – China or the US include automation platform Workato, wealth management tool