Taiwan's government strengthens 'silicon shield,' restricts exports of TSMC's most advanced process technologies
Taiwan plans to tighten control over exports of advanced process technologies as well as outbound semiconductor investments, reports Economic Daily. The new legal measures will enforce the 'N-1' technology restriction, essentially barring TSMC from exporting its latest production nodes, and introduce penalties for violations— but there's a major catch for TSMC.
The 'N-1' policy, confirmed by Premier Cho Jung-tai, will apply to TSMC's planned production in the United States. This approach restricts export of the most advanced process technology, allowing only one generation older to be deployed abroad. Before this amendment, Taiwan's regulations did not explicitly require such controls for semiconductor manufacturing processes. These rules are based on Article 22 of the amended Industrial Innovation Act, which is expected to take effect by the end of 2025.
There is a major catch about TSMC's most advanced process technology, though. Today, TSMC has one leading-edge node: N3P manufacturing technology. But by the end of the year, it will start producing chips on its N2 fabrication process, which will become its flagship. Starting from late 2026 and onwards, however, TSMC expects to have two flagship nodes: N2P for client applications that do not need advanced power delivery and A16 with Super Power Rail backside power delivery for HPC applications that consume a lot of power. It remains to be seen which process technology will be considered as 'flagship' by Taiwanese authorities, and therefore export restricted, or if the government will ban exports of both nodes for a year when TSMC introduces successors for N2P and A16, its A14 and A16P nodes.
In addition, the amended law, passed after its third reading in the Taiwan parliament, gives Taiwan's authorities the right to reject or cancel overseas investments if they are found to compromise national security, damage the country's economic development, violate treaty obligations, or result in unresolved major labor disputes. Under the new law, these six conditions are retained, but are now backed by higher-level legislation. The revised Article 22 also includes the possibility of partially or fully rejecting investments or attaching conditions to approval. If a company receives approval but later triggers any of these risks, the central authorities are authorized to demand corrective action or revoke the investment entirely if the issue is serious. The new law elevates existing investment restrictions from sub-regulations to formal legislation and adds legal consequences for non-compliance.
The Ministry of Economic Affairs stated that the law's implementation date will be announced only after the sub-regulations are revised, within six months. This means the earliest enforcement could begin by late 2025. The regulation's rollout comes amid rising geopolitical risks and after TSMC announced plans to increase its investments in its American production capacity from $65 billion over four years to $165 billion over an undisclosed period.
The amendment also introduces penalties that were not present before. Companies that invest abroad without prior approval may face fines ranging from NT$50,000 to NT$1 million ($30,830). If an investment is approved but the company later fails to correct identified violations — such as endangering national security or harming economic development—the authorities can impose repeated fines of NT$500,000 ($15,414) to NT$10 million ($308,286). But given that TSMC plans to invest $165 billion in its U.S. facilities, a $300,000 fine will hardly affect the company's bottom line.Follow Tom's Hardware on Google News to get our up-to-date news, analysis, and reviews in your feeds. Make sure to click the Follow button.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
44 minutes ago
- Yahoo
Prediction: Nvidia Stock Will Soar in the Second Half of 2025, Thanks to This Incredible News From Taiwan Semiconductor Manufacturing
Key Points Artificial intelligence (AI) has supercharged Nvidia's growth, but investors have been leery about the future of AI. Taiwan Semiconductor just reported blockbuster results, confirming that the AI revolution is alive and well. Nvidia has another catalyst that will likely boost its results in the second half. 10 stocks we like better than Nvidia › After a blistering run that spanned more than two years and stock price gains of over 1,000%, Nvidia (NASDAQ: NVDA) investors are taking a more measured approach. Concerns about the future growth prospects of artificial intelligence (AI), uncertainty about sales in China, and fears over the impact of tariffs have all weighed on the chipmaker in recent months. Perhaps the biggest question for investors is whether the AI revolution is sustainable. Many experts believe the widespread adoption of AI is just getting started, but investors have been looking for additional evidence. Taiwan Semiconductor Manufacturing (NYSE: TSM), commonly referred to as TSMC, reported its second-quarter results Thursday morning, providing the clearest sign yet that AI still has room to run. Foundry to the AI stars TSMC is the world's largest chip foundry, controlling an estimated 60% of the contract semiconductor manufacturing market. Perhaps more importantly, it has an estimated 90% of the market for the most advanced chips, those used for high-performance computing and AI -- and Nvidia is one of its biggest customers. As such, TSMC has its finger on the pulse of AI demand, and its recent performance is directly tied to the accelerating adoption of the technology. While TSMC investors had high hopes for the results, they got much more than they bargained for. In the second quarter, the company generated revenue of $30.1 billion, up 39% year over year and 11% quarter over quarter. This resulted in adjusted earnings per share (EPS) that surged 61% to a record high $2.47 (per American depositary receipt unit). To put the results into context with expectations, analysts' consensus estimates were calling for revenue of $30.2 billion and EPS of $2.38. The results were driven higher by TSMC's high-performance computing (HPC) division due to strong demand for AI. The company noted that its most advanced chips -- 7 nanometer and smaller -- represented 74% of total wafer sales during the quarter. If that weren't enough to get investors excited, TSMC raised its 2025 forecast and now expects full-year revenue to climb 30% in U.S. dollars. CEO C.C. Wei left no doubt about what drove the blockbuster results, which he said were "mainly due to continued robust AI and HPC-related demand." He went on to say that "recent developments are also positive to AI's long-term demand outlook." He pointed to increasing AI model usage and adoption, saying the company sees "AI demand continuing to be strong." What this means for Nvidia There's no denying that Nvidia has been one of the primary beneficiaries of the broad adoption of AI. In the past, the company was primarily known as a gaming company, supplying the graphics processing units (GPUs) that create lifelike images in video games -- but those days are long gone. In Nvidia's fiscal 2026 first quarter, revenue of $44 billion jumped 69% year over year, but drilling down into the results tells a story. Data center revenue, including AI chips, soared 69% to a record $39 billion. At the same time, gaming and PC revenue jumped 42% to $3.8 billion. Put another way, Nvidia's data center sales accounted for 89% of the company's considerable revenue. That's why investors are watching so closely for any sign of weakening demand for AI, as it's so central to Nvidia's success over the past few years. Looking ahead, Nvidia is guiding for second-quarter revenue of roughly $45 billion, which would represent year-over-year growth of 50%. It's worth noting that Nvidia incurred a $4.5 billion charge in the first quarter, the result of export controls on its H20 chips destined for China. Just this week, CEO Jensen Huang revealed in a blog post that Nvidia has applied for a license to resume sales of these chips and had received assurances from the Trump administration that its application will be approved. Analysts estimate that the additional sales could be worth as much as $15 billion in the second half of the year. The available evidence suggests the AI revolution is alive and well. That, combined with the resumption of chip sales to China, suggests Nvidia has a long runway for growth ahead. And since it's selling for just 29 times next year's expected earnings, it's still attractively priced. And that's why I predict Nvidia stock will likely soar for the rest of 2025. Do the experts think Nvidia is a buy right now? The Motley Fool's expert analyst team, drawing on years of investing experience and deep analysis of thousands of stocks, leverages our proprietary Moneyball AI investing database to uncover top opportunities. They've just revealed their to buy now — did Nvidia make the list? When our Stock Advisor analyst team has a stock recommendation, it can pay to listen. After all, Stock Advisor's total average return is up 1,048% vs. just 180% for the S&P — that is beating the market by 867.59%!* Imagine if you were a Stock Advisor member when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you'd have $652,133!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $1,056,790!* The 10 stocks that made the cut could produce monster returns in the coming years. Don't miss out on the latest top 10 list, available when you join Stock Advisor. See the 10 stocks » *Stock Advisor returns as of July 15, 2025 Danny Vena has positions in Nvidia. The Motley Fool has positions in and recommends Nvidia and Taiwan Semiconductor Manufacturing. The Motley Fool has a disclosure policy. Prediction: Nvidia Stock Will Soar in the Second Half of 2025, Thanks to This Incredible News From Taiwan Semiconductor Manufacturing was originally published by The Motley Fool
Yahoo
8 hours ago
- Yahoo
Google Gemini just got an AI video feature that beats ChatGPT — here's how to use it
When you buy through links on our articles, Future and its syndication partners may earn a commission. Although we generally enjoy watching videos, they can be rather time-consuming. They can also be a waste of time, particularly if they contain content that you ultimately find you don't need or won't enjoy. To make life easier, Google Gemini has introduced a new video analysis feature that works with uploaded footage. Use the Android, iOS or the web, and you can ask it to describe what can be seen. Not only that, you can ask follow up questions. It works with videos of up to five minutes in length and you can dive as deep as you want, perhaps asking it to point you to a specific section or piece of information. The feature gives Gemini an advantage over ChatGPT which cannot currently work with video uploads 1. What you need to know about Gemini's video feature (Image: © Future) The feature works with videos of up to five minutes in length and you can dive as deep as you want, perhaps asking it to point you to a specific section or piece of information. It may be able to tell you where the video was taken too. Uploaded footage can be played within Gemini if you decide you would like to watch it after all. The feature is available for free and paid users and it works in 2.5 Flash and 2.5 Pro — giving Gemini an advantage over ChatGPT which cannot, at present, work with video uploads. You should update the app to the latest version (unless you're using it on the web) and bear in mind that you have to upload pre-shot footage since you can't record videos within Gemini yet. We hope this comes soon because it would be great for quickly generating video descriptions on social media. Gemini can still summarize YouTube videos — a long-standing feature. If you have a video longer than five minutes, uploading it to YouTube and using the link in Gemini will get you around the upload restriction. 2. Select a video (Image: © Future) Launch the Gemini app and tap + in the prompt box. Next, select Gallery and choose a video to upload. Follow this by selecting Add. Alternatively, you can go to the website and drag a video into the prompt box. 3. Ask a question (Image: © Future) Now, simply ask Gemini a question. It can be a good idea to start with "What can you see in the video?" to get a comprehensive overview. When you get a response, have a read and ask any follow up questions you may have. The AI can handle detailed queries about specific timestamps, locations, or elements within the footage. Now you've learned how to use Gemini's video feature, why not take a look at our other useful guides? If you're using the AI on mobile, check out 5 smart ways to use Gemini Live with your phone right now and Gemini Live is free for everyone — here's how to share your screen and camera. And now summer's well underway and we're off out exploring the great outdoors, don't miss how to identify poison ivy using Gemini Live.
Yahoo
8 hours ago
- Yahoo
Forget Apple Intelligence — Nothing Phone's killer AI feature just got better
When you buy through links on our articles, Future and its syndication partners may earn a commission. London-based phone maker Nothing has taken a gamble with its latest device, the Nothing Phone 3. By pricing it at $799 and proclaiming it a true flagship, it's going toe-to-toe with established players like the Google Pixel 9 and Samsung Galaxy S25 — also both $799. Aside from the divisively eye-catching design (I like it, but others may disagree) the standout feature of Nothing's handset is the "Essential Space". You can tell the company thinks the feature is important, because it gets its own hardware button. In theory, Essential Space is a centralized repository of information that can then be indexed and analyzed by AI. What's the Essential Space? Well, it's yet another implementation of AI. Before you groan and stop reading; hear me out. In theory, Essential Space is a centralized repository of information that can then be indexed and analyzed by AI. At the moment, it's bare-bones — screenshots and recordings — but there's a lot of potential there to make it a killer feature. As long as Nothing doesn't start charging for it. For instance, with the Nothing Phone 3, Nothing added a new Flip to Record feature which lets you start a voice note from anywhere (using that hardware button called the "Essential Key") and store it in the Essential Space, where it's transcribed and summarized. Now, Nothing has just announced a major upgrade that includes three new functionalities: Google Calendar integration, Editable Memory, and Essential Recorder. Let's pick through them one at a time. Google Calendar integration is pretty self-explanatory; any task or event created in the Essential space can be automatically synced over to your Google Calendar without you needing to manually do anything. If the AI gets it wrong (if it mis-hears for example) you can still go in and edit the event yourself. That's the "Editable Memory" part — and while I've not seen these features in action myself, I'd hope the AI learns from your edits to make less mistakes in future. Finally, the "Essential Recorder" option takes that Flip to Record feature and takes it to the next level. The AI will turn any recording into an image, PDF or Markdown, that can then be shared anywhere, like a group message or email chain. If you're one of the people that uses a Nothing phone — like my colleague Anthony Spadafora — and you've got auto-update enabled, you won't need to do anything. The Essential Space will update automatically via the Google Play Store. According to Nothing, these updates are live now on Nothing Phone (3), and will come to Phone (3a) and Phone (3a) Pro "soon". AI used the right way AI on phones is something of a tricky sell right now. How much value consumers are actually getting for it isn't fully realised yet. But for my money, Nothing is taking a different approach than just shoe-horning in a bunch of AI overviews around the OS. It's giving users a centralised point of focus for their AI experience with hardware to back it up. It's far from fully realised yet and might not be enough of a reason to tempt people away from a Pixel 9 or Galaxy S25, but I still value the concept. Speaking of the S25, Samsung recently confirmed it would be making its "default" Galaxy AI features free to all users indefinitely while Google's Gemini just got a feature even ChatGPT can't match. The AI arms race on phone features is certainly getting interesting right now. Follow Tom's Guide on Google News to get our up-to-date news, how-tos, and reviews in your feeds. Make sure to click the Follow button. More from Tom's Guide Nothing Phone 3 Review: A first flagship that mostly sticks the landing Google Gemini just got an AI video feature that beats ChatGPT — here's how to use it Samsung's 'default' Galaxy AI features will remain free indefinitely — here's what you need to know
