Ant Group Claims AI Breakthrough Using Chinese Chips, Challenging Nvidia's Dominance

A Chinese tech giant says it slashed AI costs using only Chinese chips | TechCrunch

Earlier this year, DeepSeek briefly crashed Nvidia's stock because of speculation that its models require far fewer chips. Now, Chinese fintech giant Ant Group, which is backed by Alibaba founder Jack Ma, is claiming a major AI breakthrough. Ant was able to use Chinese chips made by Alibaba and Huawei to create methods that cut AI training costs by 20%, Bloomberg reported, citing sources familiar with the matter. What's more, the Chinese-made chips performed about as well as Nvidia chips during Ant Group's tests, Bloomberg sources claim. If these Chinese chips catch on, it could harm Nvidia's current and highly lucrative status as the most popular AI chip producer. Nvidia chips remain highly-sought after, including in China, where buyers are reportedly getting its latest Blackwell chip despite U.S. export controls. Ant Group and Nvidia didn't immediately respond to requests for comment.

Chinese fintech company uses domestic semiconductors for AI breakthrough

The company still uses Nvidia but now relies mainly on Chinese chips Ant Group, the financial technology giant backed by Alibaba, has announced a major achievement in artificial intelligence (AI) by successfully training a model using domestically produced semiconductors. According to Bloomberg, source said that Ant Group leveraged chips from Chinese tech giants Alibaba and Huawei to train its AI model, reaching performance levels comparable to those obtained with Nvidia's H800 chips. While Ant Group continues to utilize Nvidia's hardware for certain AI development tasks, the company is now relying increasingly on alternatives -- particularly chips from AMD and Chinese manufacturers -- for its latest models. This strategic pivot reflects a broader trend within China's tech industry, driven in part by tightening U.S. sanctions that limit access to Nvidia's most advanced GPUs. This development demonstrates China's growing AI capabilities and suggests that domestic and non-U.S. alternatives to Nvidia's GPUs are becoming viable for large-scale AI training. A key highlight of Ant Group's achievement is a reported 20% reduction in costs compared to using Nvidia hardware. High-performance AI training requires substantial computational power, and Nvidia's GPUs have long been the gold standard in the industry. However, with access to Nvidia's chips increasingly constrained, Chinese firms have ramped up investments in their own semiconductor technologies and diversified their hardware sources. This also raises comparisons to China's DeepSeek AI, which recently outperformed OpenAI's GPT-4 on certain benchmarks. If Ant Group's breakthrough represents a similar leap in AI training efficiency, it could mark another step toward reducing reliance on Western technology. However, questions remain about whether Chinese chips and alternative suppliers like AMD can scale effectively and whether they can match Nvidia's long-term performance and ecosystem support. While specific details about the chips used in Ant Group's AI training remain undisclosed, reports suggest that Alibaba's in-house AI hardware and Huawei's Ascend series chips played crucial roles. If other Chinese firms can replicate these results, it could accelerate China's AI ambitions and lessen the country's dependence on foreign technology. Whether these domestic and alternative AI chips can maintain competitiveness in the long run remains an open question. But this development is a clear indication of China's push toward technological independence.

Jack Ma-Backed Ant Touts AI Breakthrough Built on Chinese Chips

Jack Ma-backed Ant Group Co. used Chinese-made semiconductors to develop techniques for training AI models that would cut costs by 20%, according to people familiar with the matter. Ant used domestic chips, including from affiliate Alibaba Group Holding Ltd. and Huawei Technologies Co., to train models using the so-called Mixture of Experts machine learning approach, the people said. It got results similar to those from Nvidia Corp. chips like the H800, they said, asking not to be named as the information isn't public. Ant is still using Nvidia for AI development but is now relying mostly on alternatives including from Advanced Micro Devices Inc. and Chinese chips for its latest models, one of the people said.

Alibaba-affiliate Ant combines Chinese and U.S. chips to slash AI development costs

BEIJING -- Alibaba-affiliate Ant Group is using both Chinese and U.S.-made semiconductors for building more efficient artificial intelligence models, according to a source familiar with the matter. The combination of chips not only reduces the time and cost of training AI models, but also limits reliance on a single supplier such as Nvidia, the source said, noting the industry trend of tapping multiple networks, known as mixture of experts -- a technique that allows models to be trained with much less compute. The company earlier this month said in a paper it was able to use lower-cost hardware to effectively train its own MoE models, reducing computing costs by 20%. Ant operates Alipay, one of the two major apps for mobile payments in China. Jack Ma founded the company and its affiliate, Alibaba. Bloomberg reported Monday, citing sources, that Ant has used chips from Alibaba and Huawei for training AI models. Ant also used Nvidia chips but now relies more on alternatives from Advanced Micro Devices and Chinese chips, according to the Bloomberg report. Ant did declined CNBC's request for comment. The company on Monday announced "major upgrades" to its AI solutions for healthcare, which it said were being used by seven major hospitals and healthcare institutions in Beijing, Shanghai, Hangzhou and Ningbo. The healthcare AI model is built on DeepSeek's R1 and V3 models, Alibaba's Qwen and Ant's own BaiLing. Ant's healthcare-specific model is able to answer questions about medical topics, and can also help improve patient services, according to the company statement. The U.S. has sought to restrict China's AI development by limiting Chinese businesses' access to the most advanced semiconductors used for training models. Nvidia can still sell its lower-end chips to China.

Jack Ma-backed Ant touts AI breakthrough built on Chinese chips

Ant Group has used Chinese-made semiconductors, including chips from Alibaba and Huawei, to develop cost-effective AI training techniques, reducing costs by 20%. Its AI models, based on the Mixture of Experts approach, are competing with US firms like Nvidia. Ant aims to scale AI with lower-cost, efficient hardware for sectors like healthcare and finance.Jack Ma-backed used Chinese-made semiconductors to develop techniques for training AI models that would cut costs by 20%, according to people familiar with the matter. Ant used domestic chips, including from affiliate Alibaba and Huawei Technologies, to train models using the so-called Mixture of Experts machine learning approach, the people said. It got results similar to those from Nvidia Corp. chips like the H800, they said, asking not to be named as the information isn't public. Ant is still using Nvidia for AI development but is now relying mostly on alternatives including from Advanced Micro Devices Inc. and Chinese chips for its latest models, one of the people said. The models mark Ant's entry into a race between Chinese and US companies that's accelerated since DeepSeek demonstrated how capable models can be trained for far less than the billions invested by OpenAI and Alphabet Inc.'s Google. It underscores how Chinese companies are trying to use local alternatives to the most advanced Nvidia semiconductors. While not the most advanced, the H800 is a relatively powerful processor and currently barred by the US from China. The company published a research paper this month that claimed its models at times outperformed Meta Platforms Inc. in certain benchmarks, which Bloomberg News hasn't independently verified. But if they work as advertised, Ant's platforms could mark another step forward for Chinese artificial intelligence development by slashing the cost of inferencing or supporting AI services. As companies pour significant money into AI, MoE models have emerged as a popular option, gaining recognition for their use by Google and Hangzhou startup DeepSeek, among others. That technique divides tasks into smaller sets of data, very much like having a team of specialists who each focus on a segment of a job, making the process more efficient. Ant declined to comment in an emailed statement. However, the training of MoE models typically relies on high-performing chips like the graphics processing units Nvidia sells. The cost has to date been prohibitive for many small firms and limited broader adoption. Ant has been working on ways to train LLMs more efficiently and eliminate that constraint. Its paper title makes that clear, as the company sets the goal to scale a model "without premium GPUs." That goes against the grain of Nvidia. Chief Executive Officer Jensen Huang has argued that computation demand will grow even with the advent of more efficient models like DeepSeek's R1, positing that companies will need better chips to generate more revenue, not cheaper ones to cut costs. He's stuck to a strategy of building big GPUs with more processing cores, transistors and increased memory capacity. What Bloomberg Intelligence says Ant Group's paper highlights the rising innovation and accelerating pace of technological progress in China's AI sector. The firm's claim, if confirmed, highlights China is well on the way to becoming self-sufficient in AI as the country turns to lower-cost, computationally efficient models, to work around the export controls on Nvidia chips. Ant said it cost about 6.35 million yuan ($880,000) to train 1 trillion tokens using high-performance hardware, but its optimised approach would cut that down to 5.1 million yuan using lower-specification hardware. Tokens are the units of information that a model ingests in order to learn about the world and deliver useful responses to user queries. The company plans to leverage the recent breakthrough in the large language models it has developed, Ling-Plus and Ling-Lite, for industrial AI solutions including health care and finance, the people said. Ant bought Chinese online platform Haodf.com this year to beef up its artificial intelligence services in health-care. It also has an AI "life assistant" app called Zhixiaobao and a financial advisory AI service Maxiaocai. On English-language understanding, Ant said in its paper that the Ling-Lite model did better in a key benchmark compared with one of Meta's Llama models. Both Ling-Lite and Ling-Plus models outperformed DeepSeek's equivalents on Chinese-language benchmarks. "If you find one point of attack to beat the world's best kung fu master, you can still say you beat them, which is why real-world application is important," said Robin Yu, chief technology officer of Beijing-based AI solution provider Shengshang Tech. Ant has made the Ling models open source. Ling-Lite contains 16.8 billion parameters, which are the adjustable settings that work like knobs and dials to direct the model's performance. Ling-Plus has 290 billion parameters, which is considered relatively large in the realm of language models. For comparison, experts estimate that ChatGPT's GPT-4.5 has 1.8 trillion parameters, according to the MIT Technology Review. DeepSeek-R1 has 671 billion. Ant faced challenges in some areas of the training, including stability. Even small changes in the hardware or the model's structure led to problems, including jumps in the models' error rate, it said in the paper.

Jack-Ma's Ant Group Manages To Develop AI Models Using Huawei Chips With 20% Lower Cost, Performance On-Par With Meta's Llama Models

NVIDIA's dominance over the Chinese AI market might be in jeopardy, as the Jack Ma-backed Ant Group is said to have developed AI models using Chinese chips at a bargain. It seems like China's AI industry is witnessing what we should call a "technological revolution" credit to the US export restrictions and how domestic companies like Huawei have emerged to rescue the local markets. One of NVIDIA's biggest revenue markets, China, is seeing massive competition with the emergence of in-house AI hardware, and now, according to Bloomberg, these homegrown chips seem to be doing their job in putting NVIDIA's monopoly to an end. Ant Group, a FinTech-focused company, has been said to have developed models superior to the likes of Meta, using chips from Huawei and others. Diving a bit into the report, it is claimed that Ant Group managed to develop "Ling-Lite and Ling-Plus" AI models at 20% lower cost than the industry standard; this achievement isn't attributed to Chinese AI chips alone, but it is said that Ant Group has developed AI training techniques which are superior as well. Ant Group is said to utilize chips from Huawei, along with their own in-house solutions, and for training, the firm has used the "Mixture of Experts" technique, which enhances the efficiency and scalability of large models. Ant Group is claimed to have achieved performance from its hardware stack on par with NVIDIA's H800 AI GPUs, and the company's developed models are said to outperform Meta in certain benchmarks. However, the data is from the firm's internal testing, so we cannot be sure for now. With Ant's approach, the firm managed to train 1 trillion tokens at 5.1 million yuan, which is around 20% lower than the cost achieved using other mainstream methods. While Ant Group hasn't announced whether their models will be available for public preview, it does show that things are evolving interestingly for China, and that the nation is seemingly on its way to breaking the "West ideology" that AI training requires billions of dollars.

Ant Group Turns to Alibaba and Huawei Chips as Viable Nvidia Alternatives for AI Training

Ant Group, the Jack Ma-backed Chinese financial giant, has used domestic semiconductors from Alibaba and Huawei to cut AI development costs by 20%. The breakthrough marks a significant step forward for the Chinese firm, which is looking to continue the momentum of China's domestic AI wins following DeepSeek's success. Ant Group's Domestic AI Breakthrough Ant Group is using the newly acquired chips to develop techniques for training AI models, such as the Mixture of Experts (MoE) approach, Fortune reported , citing people familiar with the matter. MoE divides AI work among smaller "expert" models instead of one big AI model trying to do everything; each smaller model is designed to handle a specific type of input or task. The domestic chips provided the same results as Nvidia's H800 chips, which are the usual semiconductors the company relies on. Companies like Nvidia have created modified versions of their advanced AI chips specifically for the Chinese market. Nvidia's A800 and H800 chips are designed with reduced capabilities to get around U.S. export controls. Although Ant Group continues to use Nvidia for AI development, it is increasingly moving its focus to find Chinese alternatives, a source told Fortune. AI Alternatives The MoE technique has risen in popularity, gaining recognition in China through its use in DeepSeek. The method works by assigning specific tasks to smaller, specialized models rather than relying on a single large model. Only the most relevant "experts" are activated for each task, making the system more efficient and reducing computing demands. However, powering MoE still requires high-performance chips, a technology that companies like Nvidia typically provide. This has led to a stifling of some AI prowess in China, as U.S. restrictions have limited the exports of its most powerful computing equipment. Is a Cheaper AI Future Possible? Following in the footsteps of DeepSeek, Ant Group has made it its mission to scale high-performing AI models "without premium GPUs," Fortune reported. Despite DeepSeek rattling global investors with its AI model, which it reportedly developed for a fraction of the price, many executives have remained steadfast in stating that the industry will require substantially more funding. In February, Robin Li, CEO of Chinese technology giant Baidu, said more money was needed to create competitive AI models. "The investment in cloud infrastructure is still very much required. In order to come up with models that are smarter than everyone else, you have to use more compute," Li stated .