Artificial Intelligence

The US government has just banned NVIDIA and its H20 AI GPUs from entering China, with the company taking a $5.5 billion hit in Q1 2025 earnings.

NVIDIA has warehouses full of its H20 AI GPU and were preparing mass shipments to China, but President Trump's latest US export restrictions are now making waste to those $5.5 billion worth of AI chips. H20 was designed to offer a precise amount of AI workload performance that skirted around US export controls, and now those chips are sitting around doing nothing and NVIDIA is taking a $5.5B hit this quarter.

The kicker is that NVIDIA custom-tuned the H20 AI GPU after the increasing US export regulations, so this isn't just another card but a tailor-made chip for Chinese companies like Alibaba and TikTok parent company ByteDance, to enjoy enough AI power without hitting those restrictions.

Huawei just unveiled its new Ascend 920 AI chip, the successor to its China-exclusive Ascend 910C -- which is being used in a new "CloudMatrix" AI server -- within hours of new US export restrictions on NVIDIA's H20 AI GPU entering the country.

In a new report from DigiTimes, we're learning that Huawei's new Ascend 920 is a direct competitor to the NVIDIA H20 AI GPU, with a focus on the supply chain being entirely in-house with this new chip. Huawei's new Ascend 920 will be made on SMIC's 6nm process, which is an upgrade over the Ascend 910C which was fabbed on 7nm at SMIC.

We are expecting to see 900 TFLOPs of compute performance, 4TB/sec of memory bandwidth through its HBM, and a chiplet design that is similar to what we've seen with the Ascend 910C. This should all lead into around 30-40% more performance from the new Ascend 920 over 910C, meaning the Ascend 920 is a better AI chip than the now even-more-restricted H20 AI GPU from NVIDIA.

China is now using Huawei's Ascend 910C AI chips in a huge cluster called "CloudMatrix" and it reportedly has more performance than NVIDIA's leading GB200 NVL72 AI server, but CloudMatrix uses far more power.

In a new report from SemiAnalysis, we're learning about Huawei's rack scale architecture and just how the Ascend 910C will power China's new CloudMatrix 384 AI cluster, which has performance that rivals NVIDIA's most powerful AI server in the GB200 NVL72 AI server rack.

Huawei's new CloudMatrix 384 "CM384" AI cluster is powered by 384 Huawei Ascend 910C AI chips connected in an "all-to-all topology" configuration. Huawei is outweighing the architectural flaws of its AI chips by using 5x more of them than NVIDIA uses with GB200 inside of NVL72 servers. This is why the company doesn't care about the costs, performance inefficiency, scalability ratios, and more.

Once again, the more hidden costs of the AI revolution are under the spotlight, and in this case, specifically it's the topic of how much simply being polite to ChatGPT might run to.

What's being talked about here is, for example, saying thank you to the chatbot, an extra 'query' that ChatGPT processes and responds to, and how much that actually costs in terms of the electricity used therein.

That question was posed by a denizen of X, Tomie, and Sam Altman, the CEO of OpenAI, actually responded to it directly as you can see above (hat tip to TechRadar for picking up on this).

NVIDIA has just announced it will be making AI supercomputers in the US, with the new American-made supercomputers to feature NVIDIA Blackwell AI GPUs.

The company announced the news on its official website, saying it has commissioned more than a million squire feet of manufacturing space to both build and test NVIDIA Blackwell AI GPUs in Arizona, and AI supercomputers in Texas. NVIDIA Blackwell AI chips have started production at TSMC's semiconductor facilities in Phoenix, Arizona.

Jensen Huang, founder and CEO of NVIDIA, said: "The engines of the world's AI infrastructure are being built in the United States for the first time. Adding American manufacturing helps us better meet the incredible and growing demand for AI chips and supercomputers, strengthens our supply chain and boosts our resiliency".

Alphabet has committed to spending $75 billion on building out its data center capacity, even with all of the hoopla surrounding US tariffs.

Google CEO Sundar Pichai said that the $75B investment would buy the chips and build the servers needed to increase Alphabet (Google) and its core offerings, including search, as well as continued support in the development of AI services including its Gemini AI model. Pichai said: "The opportunity with AI is as big as it gets" as a surprise appearance was made at Google's annual conference for its cloud computing unit.

When asked about the potential for US tariffs to increase the cost of building out its data centers, Google Cloud infrastructure unit vice president and general manager, Sachin Gupta, said that the cost of importing hardware might increase but customer demand continued to necessitate the increased investment, he told Reuters: "We're all processing what's happening with tariffs".

China has released another video mocking overweight Americans working in factories amid the ongoing trade war between the US and China.

The second AI-generated video circulated on China's social media and has now made its way over to Western social media platforms, where it has garnered millions of views. The first video, titled "Make America Great Again," was a clear nod to President Trump's campaign slogan and depicted overweight people working in factories building clothes and smartphones.

The timing of the first AI-generated video came only days after President Trump announced heavy tariffs on China, and now a second video has begun circulating showing similar overweight Americans working in factories for prominent American companies such as Nike, Apple, NVIDIA, and Tesla. The video shows Americans building Nike shoes and eating fast food with a play on Trump's slogan on the wall, "Make America Strong Again". The video then cuts to Americans working on Tesla vehicles, smartphones, and then NVIDIA chips.

The US government has just reversed its plans to restrict exports of NVIDIA's H20 AI GPU to China, after CEO Jensen Huang had dinner at Mar-a-Lago with President Trump and his administration.

In a new report from NPR that cites two sources, the newly-tweaked US export controls on the H20 AI GPU had been in the works for a few months now, and that the change in plans kicked off after NVIDIA promised the Trump administration that it would be investing around $200 billion into new US investments in AI data centers.

NVIDIA CEO Jensen Huang recently attended a dinner at Mar-a-Lago, with the Trump administration reversing its restrictions on the H20 AI GPU entering China. Not only that, but the Trump administration also had planned to roll out additional restrictions, but they too have been put on hold according to "two sources with knowledge of the plan who were not authorized to speak publicly" adds NPR.

Samsung Ballie is described as a "home AI companion robot" that will arrive in the United States and Korea this summer. The company confirmed this week that Google Cloud's generative AI technology will power the AI side of the cute robot. This will allow owners to communicate with Ballie using natural conversational language for all their companion robot needs.

Suppose you're wondering what Ballie is about outside of rolling around and looking like a cute robot from a Star Wars or Pixar film. In that case, it's essentially a range of generative AI tools and functionality paired with an inbuilt projector. Plus, it can do things like greet guests, adjust smart home things like lighting and cooling, and set reminders for appointments and things. A robot butler that doesn't serve food or do any cleaning.

Samsung's description of how Ballie will use cloud-based AI is vague - though it provides some examples. Yongjae Kim, Executive Vice President of the Visual Display Business at Samsung Electronics, states that Ballie "moves with users, anticipates their needs, and interacts in more dynamic and meaningful ways than ever before," thanks to generative AI.

Amid President Trump's recently announced trade tariffs, an artificial intelligence-generated video of Americans working in factories has gone viral on Chinese social media. The video has now made its way to Western social media platforms, where it has garnered quite a lot of attention.

The video has been copied and reshared on various social media platforms and shows a satirical take on what American people look like working in factories. The video was titled "Make America Great Again," a clear nod to President Trump's campaign slogan. It depicts overweight people working in what appears to be sweatshops, making various items of clothing and smartphones. The AI-generated people in the video also look miserable.

The timing of the AI-generated video mocking Americans comes only days after President Trump announced heavy tariffs on China, along with numerous other countries. Trump's goal of the tariffs is to make better deals with its trade partners and ultimately bring manufacturing back to the United States. It appears China is attempting to show Americans what the future holds for them with this satirical AI generated video.