عربي English Türkçe اردو فارسى
Logo
Latest News
Technology

Alibaba Unveils New AI Chip in Push for Domestic Alternatives

A visitor walks in front of Alibaba booth during the 3rd China International Supply Chain Expo at the China International Exhibition Center, in Beijing, China, Friday, July 18, 2025. (AP)
A visitor walks in front of Alibaba booth during the 3rd China International Supply Chain Expo at the China International Exhibition Center, in Beijing, China, Friday, July 18, 2025. (AP)
TT
TT

Alibaba Unveils New AI Chip in Push for Domestic Alternatives

A visitor walks in front of Alibaba booth during the 3rd China International Supply Chain Expo at the China International Exhibition Center, in Beijing, China, Friday, July 18, 2025. (AP)
A visitor walks in front of Alibaba booth during the 3rd China International Supply Chain Expo at the China International Exhibition Center, in Beijing, China, Friday, July 18, 2025. (AP)

Alibaba Group on Wednesday unveiled a new AI chip, the Zhenwu M890, as the Chinese technology giant intensifies efforts to build domestic alternatives to Nvidia processors amid tightening US export curbs.

The chip, developed by Alibaba's semiconductor design subsidiary T-Head, delivers three times the performance of its predecessor, Zhenwu 810E. It is purpose-built for the emerging wave of AI "agents" — software systems that can carry out complex, multi-step tasks with limited human oversight.

Alibaba said the new processor is well-suited to handle the heavy memory ‌and communication demands ‌of agent workloads, where models must retain long ‌stretches ⁠of context and ⁠coordinate with one another in real time.

The company also outlined a multi-year chip roadmap, saying it would follow the M890 with a successor called the V900 in the third quarter of 2027, and a further chip, the J900, in the third quarter of 2028. The V900 is expected to deliver another roughly threefold performance gain over the M890, Alibaba ⁠said, signaling a sustained cadence of in-house silicon upgrades.

The ‌plan underscores China's growing efforts to ‌produce locally developed AI chips as Washington bans the sale of the most ‌powerful US processors to Chinese customers, and follows a similar announcement ‌by Huawei last year.

Hangzhou-based Alibaba last year pledged to spend more than 380 billion yuan ($53 billion) on cloud and AI infrastructure over three years, its largest-ever commitment to the sector.

The investment reflects a broader bet across China's technology ‌industry that demand for AI computing power will continue to surge as enterprises adopt agent-based applications.

Alibaba unveiled ⁠the chip ⁠at its annual Alibaba Cloud Summit, alongside a new server system, the Panjiu AL128, which packages 128 of the accelerators into a single rack.

The system is available immediately to Chinese enterprise customers through Alibaba Cloud's domestic model platform, known as Bailian.

T-Head said it has shipped more than 560,000 Zhenwu units to date, with over 400 external customers across 20 industries, including automakers and financial services firms, having deployed the chips.

Alibaba also announced Qwen 3.7-Max, the latest version of its flagship large language model, which it said is engineered for advanced coding and long-running agent tasks. The company said the model can operate continuously for up to 35 hours without performance degradation.

Most Viewed

 
 
Technology

EU Accepts Action Plan by Elon Musk’s X to Become More Transparent

Teens pose for a photo while holding smartphones in front of the X logo. (Reuters)
Teens pose for a photo while holding smartphones in front of the X logo. (Reuters)
TT
TT

EU Accepts Action Plan by Elon Musk’s X to Become More Transparent

Teens pose for a photo while holding smartphones in front of the X logo. (Reuters)
Teens pose for a photo while holding smartphones in front of the X logo. (Reuters)

‌The European Union, which fined Elon Musk's social media network X €120 million ($137.2 million) last year, said on Wednesday it had accepted an action plan by X to comply ‌with transparency ‌rules under ‌the ⁠EU's Digital Services Act.

"The ⁠European Commission has accepted X's action plan to comply with transparency obligations and researchers' access to data ⁠under the Digital ‌Services ‌Act," it said in ‌a statement.

"The approved ‌measures represent an important step in enabling researchers, civil society and the public ‌in general to gain more transparency into ⁠X's ⁠systems, in particular to monitor X's systemic risks and to assess the platform's broader impact on its users and European society as a whole," it added.

Technology

China's DeepSeek to Raise Fresh Capital at $74 Billion Valuation Ahead of Onshore IPO

FILE PHOTO: A DeepSeek AI sign is seen at a building where the Chinese start-up's office is located in Beijing, China, February 19, 2025. REUTERS/Florence Lo/File Photo
FILE PHOTO: A DeepSeek AI sign is seen at a building where the Chinese start-up's office is located in Beijing, China, February 19, 2025. REUTERS/Florence Lo/File Photo
TT
TT

China's DeepSeek to Raise Fresh Capital at $74 Billion Valuation Ahead of Onshore IPO

FILE PHOTO: A DeepSeek AI sign is seen at a building where the Chinese start-up's office is located in Beijing, China, February 19, 2025. REUTERS/Florence Lo/File Photo
FILE PHOTO: A DeepSeek AI sign is seen at a building where the Chinese start-up's office is located in Beijing, China, February 19, 2025. REUTERS/Florence Lo/File Photo

Chinese AI startup DeepSeek is planning to launch a fresh fundraising round at a valuation of about 500 billion yuan ($74 billion) ahead of a potential mainland initial public offering, two people with knowledge of the matter said.

The plan comes just weeks after the Hangzhou-based company, which drew global attention with its low-cost AI models in 2025, raised about $7.4 billion in June a post-money valuation of about 450 billion yuan, the people said.

The back-to-back fundraising plans underscore strong investor appetite for one of China's most closely watched AI companies, but also point to the rising costs of competing in AI, which requires large amounts of computing power, data-center capacity and engineering talent.

DeepSeek is looking to raise as much as 50 billion yuan in the new funding round, according to a third person briefed on the ⁠matter.

It has also started ⁠early deliberations on a potential IPO on Shanghai's Nasdaq-style STAR Market, the three sources and two other people with knowledge of the plan said.

The company has set an internal target to complete an IPO filing this year, one of them told Reuters.

All the people declined to be identified because the information is not public. The fundraising and IPO plans are at early stages and terms and timetable may change, they said.

DeepSeek did not immediately respond to a request for comment.

Bloomberg News first reported on Tuesday ⁠that DeepSeek was preparing for a possible IPO filing, while the Financial Times reported that the company was weighing a fresh fundraising round at a valuation of at least 480 billion yuan.

DeepSeek shook global technology markets last year after releasing models that appeared to rival leading US systems at lower training and operating costs.

Soon after its maiden fundraising round in June, DeepSeek said it planned to double staff across departments, including in areas such as data centers and AI agents, systems capable of performing tasks with limited prompting.

Some of those initiatives will require significant capital expenditure. Reuters reported earlier this month that DeepSeek was looking to develop its own AI inference chip and had discreetly increased hiring of chip-design engineers for the project.

DeepSeek had long stood out in China's AI sector for rejecting outside ⁠funding. Founder Liang Wenfeng ⁠had largely bankrolled the company using his quantitative hedge fund High-Flyer before its recent external financing, sources previously told Reuters.

But the cost of staying at the frontier of AI has risen sharply, forcing a change in strategy.

DeepSeek has in the past year faced stiff competition at home from tech giants including ByteDance and Alibaba, as well as well-funded AI startups such as Z.ai, Moonshot, and MiniMax.

In the June funding round, DeepSeek founder Liang personally committed 20 billion yuan, while Tencent Holdings and battery giant CATL chipped in 10 billion yuan and 5 billion yuan respectively to become the largest external shareholders, Reuters reported at the time.

Other investors include China's national AI fund, gaming developer NetEase and e-commerce giant JD.com , as well as investment firms IDG Capital, Loyal Valley Capital, Monolith Management and Shixiang Capital, according to sources and media reports.

The participation of the state-backed AI fund highlighted DeepSeek's strategic importance to Beijing's efforts to build domestic AI champions and reduce reliance on foreign technology.

Technology

Generative AI’s Power Sparks Fears of Dumbing Humans Down

An AI sign is seen at the World Artificial Intelligence Conference in Shanghai, China July 6, 2023. (Reuters)
An AI sign is seen at the World Artificial Intelligence Conference in Shanghai, China July 6, 2023. (Reuters)
TT
TT

Generative AI’s Power Sparks Fears of Dumbing Humans Down

An AI sign is seen at the World Artificial Intelligence Conference in Shanghai, China July 6, 2023. (Reuters)
An AI sign is seen at the World Artificial Intelligence Conference in Shanghai, China July 6, 2023. (Reuters)

Generative AI chatbots capable of writing emails and computer code, translating, organizing a trip or coming up with gift ideas are now readily available, prompting some to ask whether human brainpower could suffer for lack of use.

A simple natural-language prompt is usually enough to draw a useable response from a service like ChatGPT or Claude, with the effects making themselves felt in schools and universities, workplaces from offices to courtrooms and our personal lives.

Recent scientific studies suggest there could be harmful consequences to farming out cognitive tasks to AI.

They highlight memory, decision-making and critical thinking as particularly at risk.

One American-British study of 1,222 people, still under peer review, found that using AI tools to solve arithmetic or reading comprehension exercises improved participants' performance in the short term, but in the long run diminished their results and their willingness to keep trying when the tools were unavailable.

"These findings are particularly concerning because persistence is foundational to skill acquisition and is one of the strongest predictors of long-term learning," the authors wrote.

AI's ability to conjure up speedy responses to all kinds of questions "removes learning opportunities" from users, said Carnegie Mellon University doctoral student Grace Liu, the article's main author.

"What makes AI particularly concerning is that it's not a tool designated for one specific kind of activity. It's something that can be used across pretty much any intellectual, reasoning, and cognitive activity."

The technology's adaptability to different kinds of problems sets it apart from previous waves of computerized aids.

Electronic calculators, for instance, may have helped users solve equations, but left the method and reasoning process in human hands.

- Saving energy -

One 2025 MIT study went viral for its finding that students using generative AI to write essays displayed less critical thinking capability.

Other research has pointed the same way, highlighting what has come to be called "cognitive offloading" -- or even "cognitive surrender".

"Human beings have a strong tendency to save energy," said Johann Chevalere, a researcher in social and cognitive psychology at France's publicly-funded CNRS institution.

"In daily life, we often use strategies that get us to the heart of the matter quicker, without necessarily taking the time to study in depth the information we need to process, as this can be cognitively costly," he added.

Generative AI use could strengthen this tendency, Chevalere said.

"If there are activities you never do, the brain -- which works by saving energy -- won't go to the trouble of maintaining connections that aren't being used."

- Encouraging reflection -

Under pressure from critics, generative AI developers have begun building so-called "Socratic" functions into their models, which for now remain mostly aimed at students.

In this mode, chatbots do not simply provide the answer, instead offering hints and asking questions to stimulate users' thinking.

Examples include the "study mode" built into OpenAI's ChatGPT, or "guided learning" in Google's Gemini.

US software giant Microsoft told AFP it had built warnings about the risk of mistakes into its Copilot models.

The AI also reminds users to check the information it provides, just one of several measures designed to keep them actively and critically engaged with its responses.

"The risk of excessive cognitive offloading is real, especially if AI is used to automate tasks that are also valuable for developing skills," Microsoft said, adding that users have to be trained to use the tools correctly.

For now, there is a lack of large-scale, long-term studies to judge the true impact of the new technology on human brains, researchers agreed.

Until they are available, "it's up to us to use AI in a smart way," Chevalere said.

"We'll adapt to this technological revolution just as we have to the previous ones."