China Telecom, a powerhouse in state-owned telecommunications, has achieved a remarkable feat by training two large language models (LLMs) entirely on domestic chips. This advancement underscores China's strategic push for autonomy in artificial intelligence, spurred by escalating U.S. restrictions on access to sophisticated semiconductors.

According to China Telecom's AI Institute, the TeleChat2-115B model and another unnamed model benefited from training on a vast number of Chinese-produced chips. This development is particularly significant amidst tightened U.S. export policies targeting high-performance processors from Nvidia and other foreign manufacturers. The AI Institute, through a WeChat statement, heralded this success as proof of China's capability to independently train LLMs, marking a new chapter of innovation and technological self-sufficiency.

The scale of these AI models is noteworthy. Among them, the unnamed LLM boasts an impressive one trillion parameters, setting a high benchmark in AI complexity and power. Parameters, fundamental in AI education, determine a model's ability to learn and process information. With more parameters, AI systems become increasingly sophisticated.

Chinese enterprises are making strides to rival global AI leaders based overseas. U.S. sanctions on Nvidia's state-of-the-art AI chips, including the A100 and H100 models, have propelled Chinese industries toward self-reliance. Consequently, firms like China Telecom are developing indigenous processors to reduce their dependence on Western tech. For instance, the TeleChat2-115B, with its 100 billion parameters, competes with international mainstream platforms effectively.

While China Telecom has not disclosed specific chip sources, Huawei's Ascend processors are pivotal in China's AI blueprint. Despite facing U.S. sanctions, Huawei is ramping up its AI initiatives, testing its cutting-edge Ascend 910C chip domestically. Major Chinese server manufacturers and internet companies previously reliant on Nvidia chips are evaluating this new technology. Ascend chips represent China's strategy shift, aiming to diminish its reliance on foreign hardware.

Beyond Huawei, China Telecom is partnering with domestic innovators like Cambricon, a startup dedicated to AI chip manufacturing. This collaboration reflects a larger movement in China's tech industry, fostering a self-sufficient AI ecosystem to mitigate the ramifications of U.S. trade policies.

Progress in homegrown AI technology is steadily reducing China's dependency on imported hardware, such as Nvidia's costly GPUs. As U.S. restrictions challenge Chinese companies' access to the latest Nvidia offerings, a clandestine market for foreign chips has emerged. To avoid legal ambiguities, many choose prior-generation Nvidia models, securing official support while complying with regulations.

This breakthrough symbolizes a shift in China's approach to AI and semiconductors, emphasizing self-reliance and resilience in a competitive global economy, influenced by American protectionism.

As China continues to foster innovation and self-sufficiency in AI technologies, opportunities for creative applications in visual media also expand. AI advancements enable transformative experiences in digital expressions and storytelling, linking seamlessly to the use of new tools that enhance visual narratives.

The Future of AI in Video Content Creation

In today’s fast-paced digital world, video remains the most powerful tool for communication and storytelling. Whether you're a content creator or someone who simply loves sharing life’s moments, producing captivating videos is key to standing out online. However, traditional video production often comes with steep learning curves, high costs, and time-consuming edits.

That’s where the evolution of the AI video generator truly shines.

With platforms like Dreamlux, anyone can turn simple photos or prompts into polished, animated videos in just minutes—no editing skills required. From seamless transitions to expressive effects, these tools streamline content creation while fueling creativity in ways never before possible.

And beyond automation, AI is now enhancing something even more powerful: emotional expression.

Enter the World of AI Smile Generator

Among the growing list of playful and heartwarming AI features, the AI Smile Generator stands out for its simplicity and charm. Using advanced facial recognition and generative models, this tool brings your photos to life—adding a natural, radiant smile to faces with stunning realism.

Whether you're updating a portrait, enhancing a memory, or just curious to see someone’s happier side, the AI Smile Generator transforms still images into lively, expressive visuals. It’s a subtle yet powerful way to inject warmth and positivity into your content—perfect for social media posts, digital gifts, or just brightening someone’s day.

How to Use Dreamlux AI Smile Generator

Creating a photo that beams with joy is easier than ever with Dreamlux. Just follow these steps:

1. Visit the official https://dreamlux.ai and click on "Templates"
2. Select the "Free AI Smile Generator" from the available effects
3. Upload a clear portrait of the person you’d like to see smiling
4. Click "Create", and within seconds, the AI will deliver a vivid, heartwarming version of the image—complete with a lifelike smile

With Dreamlux, even the smallest expression can become a memorable, shareable moment.