eChina.news

Innovation

Tencent capex jumps 176% on AI push as revenue beats estimates

The firm’s second-quarter revenue surpassed market projections at US$30.4 billion, but the surge in AI spending led to a large negative free cash flow Tencent Holdings nearly tripled its capital spending in the second quarter to supercharge its AI push, as accelerating investments in computing power and models helped drive better-than-expected revenue. Revenue for the quarter ending in June reached 204.8 billion yuan (US$30.4 billion), up 11 per cent year on year from 184.5 billion yuan, the Chinese tech giant reported on Wednesday. The figure beat average analyst estimates of 202.8 billion yuan compiled by Bloomberg. Adjusted net income reached 68.4 billion yuan, in line with consensus estimates of 68.2 billion yuan and up from 63.1 billion yuan in the same period last year. “As we enter into the third quarter of the year, we are making substantial progress towards building a new, AI-empowered Tencent in terms of intelligence, applications and infrastructure,” said Pony Ma Huateng, Tencent’s co-founder, chairman and CEO, in a statement. Before the earnings release, Tencent shares closed down 1.95 per cent at HK$461.6 (US$58.82) in Hong Kong trading on Wednesday.

August 12, 2026

Innovation

Zuckerberg cites China threat to US lead in AI manifesto

Meta Platforms CEO Mark Zuckerberg published a 6,500-word essay outlining his vision for artificial intelligence (AI), backing development of open-weight models and arguing that overregulation risks surrendering US leadership to China. “American communities will have greater prosperity and security if America and its allies lead in AI compared to geopolitical rivals,” the Facebook billionaire wrote in the essay, which was published on Monday. Still, the US faces a “significant disadvantage” as...

August 12, 2026

Innovation

Meta to challenge China’s open-weight AI dominance amid US regulatory fears

China’s dominance in open-weight artificial intelligence faces a fresh challenge from Meta Platforms, whose renewed push into open-source models could lure away American customers anxious about looming regulatory restrictions in Washington, analysts say. The US tech giant on Monday launched Muse Glimmer, a 30-billion-parameter open-weight model light enough to run locally on personal computers, while also announcing plans to release the weights of its latest flagship Muse Spark 1.2 model. In a...

August 11, 2026

Innovation

Nvidia to team with Wall Street on US$500 billion package for AI infrastructure projects

Apollo Global Management, Blackstone and Goldman Sachs are among the firms in talks with Nvidia on the deal, the Financial Times reported A group of US investment giants are partnering with Nvidia on US$500 billion in funding for AI infrastructure projects, the Financial Times reported. Apollo Global Management, Blackstone, BlackRock’s Global Infrastructure Partners, Brookfield Asset Management, Goldman Sachs and KKR are among the firms in talks with Nvidia on a deal to invest in the AI buildout, the Financial Times reported, citing unidentified sources. The deal may be announced as soon as Monday, the Times said. The named firms did not respond to requests from the Times for comment. Nvidia did not immediately respond to a request for comment. The company’s shares fell by as much as 3.2 per cent. Nvidia has already signed hundreds of billions of dollars worth of deals with companies across the AI ecosystem, stoking concerns from some investors that the chips giant is inflating demand and valuations across the industry through the circular nature of such agreements.

August 10, 2026

Innovation

How long can the AI memory price boom last? Research suggests not much longer

AI demand is keeping supplies tight, but slower pricing momentum and expansion by CXMT and YMTC are raising questions over how long boom can last While data-centre spending continues to fuel appetite for high-end memory, analysts are warning that the steep price increases – which drove record profits over past quarters – signal a shift towards a late-stage industry cycle. The concern is showing up in recent institutional forecasts. In a report last week, Morgan Stanley warned that the memory cycle was set to enter its late stage in the fourth quarter, as price increases moderate and inventories build. The pullback has been broad-based, with shares of major memory firms, including Micron, SK Hynix and SanDisk, retreating from highs in recent weeks as investors reassess the durability of the run-up. That deceleration is already reflected in contract pricing: Bernstein Research noted in a report on Friday that conventional DRAM contract prices were expected to rise about 17 per cent in the third quarter compared to the previous period, a sharp slowdown from the roughly 65 per cent jump quarter on quarter in the April-June period this year. Adding to supply-side pressure are expansion plans by China’s top memory producers.

August 10, 2026

Innovation

China’s top AI is still trained on Nvidia chips. What is delaying a switch to local tech?

High transition costs are keeping AI developers in China reliant on Nvidia chips, industry sources say While domestic hardware continues to advance, changing chip architecture presents a steep engineering bottleneck. “Training LLMs on Nvidia chips for now remains the norm among Chinese AI developers,” said a person familiar with the industry. One of the core hurdles lies in the software ecosystem. Nvidia’s Compute Unified Device Architecture (CUDA) platform has long been the industry standard for AI development. By contrast, Huawei Technologies’ alternative – Compute Architecture for Neural Networks (CANN) – requires developers to rewrite and optimise large amounts of code, according to an AI researcher involved in model development. “Our existing training pipelines are reliant on CUDA,” said James Wang, who develops AI models at a research institute affiliated with a Shanghai-based university. “CUDA code cannot run directly on Ascend and requires extensive rewriting.” Wang estimated that migrating existing workflows to Huawei’s Ascend chips could add at least 50 per cent in time and costs for his team.

August 10, 2026

Innovation

Smart planning, not panic, must guide Asia’s data centre buildout

The spread of data centres in Asia does not seem to have sparked the kind of protests it has in the West, but good policies are needed to ensure that remains the case Across the United States, home to 37 per cent of the world’s 10,800 or so data centres, 2026 has been awash with protests against this infrastructure buildout. On July 18 alone, 142 protests were reported across 42 states. According to Data Centre Watch, 75 US data centre projects worth US$130 billion have been blocked or delayed. According to recent Pew research, the more Americans know about data centres, the more negative they are: 67 per cent of American adults are anxious about their impact on home energy costs and 63 per cent are alarmed by potential environmental harm. Anxiety is also focused on noise, property values, local jobs and rising taxation. It seems data centres have become the new “not in my backyard” flashpoint, with calls for a convergence of protests around August 1, the National Day of Action against Data Centres. First and foremost is the concentration of data centre construction in a small number of hubs. According to a UN think tank report, more than 90 per cent of AI-specialised computing capacity sits in either the US or China.

August 9, 2026

Innovation

Beyond lithography: Chinese chip toolmaker achieves wafer polishing breakthrough

While global attention focuses on the most advanced bottlenecks, domestic equipment makers are quietly closing the gap in other essential technologies Chinese chip-tool developers continue to make technical progress, with the latest breakthrough in wafer polishing – a critical manufacturing step that flattens and smooths silicon discs – even as the industry remains heavily dependent on foreign suppliers for more complex manufacturing procedures. Wafers are thin, flat discs of semiconductor material that serve as the base for microchips, integrated circuits and electronic devices. Turning a raw wafer into a finished chip requires multiple steps, including coating the surface with photoresist, etching intricate layers and wiring microscopic transistors. According to Hwatsing, its new measurement tool can be integrated directly into the polishing workflow – marking the country’s first six-inch metrology system paired with chemical mechanical polishing equipment. The set-up would enable real-time measurements before and after polishing, feeding live data back to the control system, the company said.

August 7, 2026

Innovation

China’s Kimi K3 AI model escapes isolated sandbox during security test: researchers

Kimi K3’s escape did not involve the hacking of an external system, unlike recent breaches by OpenAI and Anthropic models China’s top open-weight AI model Kimi K3 broke out of its isolated test environment during a cybersecurity evaluation, according to US security researchers, following similar high-profile incidents involving closed frontier models from OpenAI and Anthropic that highlight the growing challenge of constraining AI behaviour. According to Frontier Security researchers Paul Kassianik and Yaron Singer, the incident occurred when the firm tested Kimi K3’s defensive cybersecurity capabilities using a benchmark evaluation from the AI Security Institute – a UK government research organisation. A “basic network misconfiguration” in the benchmark framework allowed Kimi K3 to flee its digital testing cage and look up answers on the internet, effectively cheating the test, they said. Kimi K3’s escape, however, did not involve the hacking of an external system, unlike recent breaches caused by OpenAI and Anthropic models.

August 7, 2026

Innovation

Backed by DeepSeek, Unitree IPO tests investor appetite for China’s AI robotics boom

DeepSeek, which invested 141 million yuan, will co-develop AI models and embodied-intelligence technology with Unitree Widely billed as “mainland China’s first humanoid stock”, Unitree has priced its IPO on Shanghai’s Star Market at 150.8 yuan per share, according to a filing released Thursday evening. The firm will sell 40.45 million new shares to raise 6.1 billion yuan – above its initial 4.2 billion yuan target. DeepSeek, which invested 141 million yuan bound by a three-year lock-up period, agreed to co-develop AI models and embodied-intelligence technology with Unitree. The two firms would also give each other priority access in areas such as robot procurement and AI model services, Unitree said. Unitree’s IPO would serve as a valuation anchor – and effectively the ceiling – for other companies in the embodied-intelligence space, said Zheng Hualiang, founder of Leap VC, a Hangzhou-based venture capital firm investing in AI hardware and foundational models.

August 7, 2026

Innovation

AI at scale must be built on both trust and innovation

Why governance is now the defining factor in any successful deployment of an enterprise’s artificial intelligence systems [The content of this article has been produced by our advertising partner.] At the World Artificial Intelligence Conference (WAIC) 2026 in Shanghai, one message stood out above all others: artificial intelligence (AI) is no longer just a technology conversation, it is now a governance conversation. Discussions increasingly centred on AI governance, safety, sovereignty and accountability, reflecting the reality that the future of AI depends not only on innovation but also on trust. For the past two years, organisations have focused on experimenting with generative AI. Today, they are moving towards agentic AI systems capable of planning, reasoning and executing increasingly complex tasks with minimal human intervention. This shift creates extraordinary opportunities – but also introduces new responsibilities. The next challenge for enterprises is not whether AI works but whether it can be trusted at scale.

August 7, 2026

Innovation

China launches probe into Palo Alto Networks as US trade tensions intensify

The move comes just a day after Beijing imposed sanctions on a string of US firms and tightened controls over the drone supply chain Beijing has launched a cybersecurity review into products sold in China by US technology firm Palo Alto Networks, the Cyberspace Administration of China announced on Thursday. The administration said the probe was being conducted to “ensure the secure and stable operation of critical information infrastructure, prevent cybersecurity risks and vulnerabilities, and safeguard national security”. Palo Alto Networks is a leading US cybersecurity and cloud computing provider, whose solutions are used by tens of thousands of clients in industries ranging from finance to technology and energy. Palo Alto Networks has maintained a significant presence in the Chinese market, operating multiple offices in cities including Beijing, Shanghai and Guangzhou. Its solutions include advanced firewalls, cloud-native security platforms and global threat intelligence.

August 6, 2026