Foxconn's Apple era fades as AI servers drive growth in Taiwan tech sector
TAIPEI (Reuters) -Taiwan's Foxconn, which rose to become a global tech manufacturing juggernaut by assembling millions of iPhones, can now say its main business is no longer Apple as it takes advantage of the AI-boom to diversify its income.
Its revenue from making AI servers and other cloud and networking products, including for major customer Nvidia, surpassed smart consumer products such as iPhones for the first time in the second quarter, marking the culmination of a shift that began years ago and has swept through Taiwan's tech industry.
Foxconn's heavy reliance on the smartphone business has long been viewed by investors as a significant risk, as demand growth for new iPhones has gradually weakened since they were first introduced nearly two decades ago, leaving the top iPhone assembler grappling with slowing sales momentum, analysts said.
Wary of the risk, Foxconn Chairman Young Liu has been championing new businesses such as AI servers, electric vehicles and semiconductors since taking the top job in 2019.
While its expansion into EVs and chips has yet to show a meaningful contribution to its topline, Foxconn's success in AI server manufacturing - the company is Nvidia's biggest server maker - is the result of its early bets before the technology was thrust into the limelight with the advent of ChatGPT in late 2022.
Consumer electronics accounted for 35% of Foxconn's total revenue in the second quarter, while cloud and networking business represented 41%. In 2021, consumer electronics represented 54% of its revenue.
The firm's prudent wagers years back helped it cultivate a now-prized relationship with the U.S. AI chip firm and other major AI players, analysts said.
"The company has been in the business for years, meeting higher quality requirements, diversifying assembly and operations across sites, and pursuing vertical integration,' said Ming-Chi Kuo, an analyst at TF International Securities.
Foxconn began producing reference designs for Nvidia's graphics cards around 2002 and started making general-purpose servers for cloud service providers' data centres as early as around 2009. Its AI server business with Nvidia is in many ways the culmination of that history, analysts said.
Foxconn says it is now one of the world's largest suppliers of both general-purpose and AI servers, with a market share of nearly 40% in each.
The company has also shown a willingness to commit investment to a project at an earlier stage than other companies, Kuo said, citing its past investments for Apple and similar moves for Nvidia. 'In long-term partnerships, Foxconn is more willing to take the initiative,' he said.
Foxconn's plan to build factories in Houston, Texas — part of Nvidia's $500 billion U.S. investment plan — and in Mexico to produce AI servers for the U.S. client underscores this strategy, analysts said.
Foxconn now expects its AI server revenue would grow more than 170% in the third quarter year-on-year.
Foxconn and Nvidia declined to comment. Apple did not respond to request for comment.
BROADER SHIFT
The shift at Foxconn mirrors a broader trend in Taiwan's technology sector, where companies once centred on consumer electronics — such as Foxconn with iPhones, and Quanta Computer and Wistron Corp with notebooks — are now investing heavily in AI servers.
Nvidia partner Wistron's revenue for January to July rose 92.7%, while Quanta's grew 65.6% in the same period.
"The monthly sales jump for Taiwan ODMs in the first half of 2025 is evidence of this trend,' said Robert Cheng, head of Asia technology hardware research at BofA Global Research, referring to original design manufacturers like Foxconn that contract manufacture products for their clients.
Their fast transition into AI servers is also the result of Taiwanese tech supply chain working closely with U.S. tech giants on data centre infrastructure work for a decade now, according to Chris Wei, industry consultant at Taiwan's Market Intelligence & Consulting Institute.
He estimates Taiwan accounts for about 80% of global server shipments and more than 90% of AI servers.
Cheng agrees.
"We think this shift toward AI servers, whatever form it takes, is good for Taiwan's tech industry," he said, noting Taiwanese firms' ability to rapidly shift to cater to changing needs from their customers.
擷取數據時發生錯誤
登入存取你的投資組合
擷取數據時發生錯誤
擷取數據時發生錯誤
擷取數據時發生錯誤
擷取數據時發生錯誤
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Chicago Tribune
11 minutes ago
- Chicago Tribune
Bradshaw: Advice for high schoolers entering the AI era
Dear Freshman, You're starting high school with an advantage: you already know you're interested in math and computers. That focus can set you apart — but you need to understand the world you're stepping into. By the time you graduate, artificial intelligence will be doing much of the work people train years for today. Coding simple programs, solving standard math problems, even designing basic websites — AI can already do these things faster, cheaper, and often better than most humans. And it's improving. The newest version of Chat GPT-5 was released last Thursday. It's a doctorate-level expert on any subject. Anyone can create software by typing in simple English language prompts. It's called 'vibe coding.' That's not a reason to give up. It's a reason to aim higher. Your goal isn't just to learn skills — it's to learn how to think, adapt, and work with AI, so you're the one directing the tools, not the one being replaced by them. 1. Build your math foundation — because reasoning is still human territory. Math as a set of procedures is easy for AI. Math as a way of thinking is still a human edge. Algebra and geometry aren't just boxes to check—they're your training ground for logical reasoning, problem decomposition, and spotting errors. AI can solve a problem, but it often can't judge whether the problem makes sense. Aim for mastery, not speed. If you take calculus by senior year, great — but the bigger win is learning to frame problems, question assumptions, and verify solutions, especially when AI hands you an answer. Those skills translate into every field AI will touch — which is all of them. 2. Learn to code — but as a designer, not just a typist. Yes, AI can write code. In fact, it can write decent code with just a short prompt. That means your value isn't in typing every line — it's in knowing what to build, why it matters, and how to guide the AI to produce it. Python is still a great starting point, but think of it as learning to read and write in a new language so you can collaborate with AI fluently. The earlier you understand the structure of programs, the easier it will be to spot AI's mistakes, combine AI-generated components into something original, and add the creativity and judgment that machines still lack. 3. Join competitions and projects — but choose ones AI can't dominate. Math competitions and coding hackathons are still valuable but understand the landscape: AI can already solve many contest style problems. The human advantage now is in creative problem framing, strategy, and interpreting messy, incomplete data. Look for contests or projects that require innovation, interdisciplinary thinking, or human insight — like robotics design, ethical AI challenges, or data projects tied to real world communities. If your school doesn't have a club that takes this approach, start one. Colleges will notice a student who organizes an 'AI + Society' club more than another generic coding group. 4. Make summers your AI-era laboratory. Summer projects matter more than ever — but the projects that will stand out are those that combine AI with something unique to you. Building yet another calculator app won't impress anyone. Using AI to analyze local environmental data and present it to city planners? That's original. By the time you're applying to college, admissions officers will see thousands of AI-assisted projects. The ones that stand out will be those where the student clearly drove the vision, used AI as a partner, and produced something tied to a personal interest or local need. 5. Read widely — especially about how technology reshapes society. The technical history of people like Alan Turing and Grace Hopper is still inspiring. But now you should also study the thinkers wrestling with AI's impact — economists, ethicists, historians of technology. Understand not only how to build a tool, but how that tool changes jobs, politics, and even human relationships. Books like 'Life 3.0' by Max Tegmark or 'Prediction Machines' by Agrawal, Gans, and Goldfarb will give you a broader view of AI's role in the economy you're heading into. 6. Communication is no longer optional, it's survival. Ironically, as AI gets better at writing and speaking, human communication skills are becoming more valuable. In your high school years, practice turning complex, AI-assisted work into clear, persuasive presentations. Lead a meeting, explain your process, write a compelling project report. These skills will help you manage AIdriven teams later on. 7. Treat curiosity as your competitive advantage. AI is trained on the past. Your edge is seeing possibilities that aren't in its data yet. That's why you should follow your curiosity beyond the obvious — physics, economics, art, and philosophy. Many of the breakthroughs in AI itself come from unexpected intersections of disciplines. When something sparks your interest, chase it down — talk to experts, experiment, connect it back to your math and computer skills. The more unique the mix of your knowledge is, the harder you are to replace. No better place to ask these questions than as a student on the high school newspaper. 8. Find mentors who are already living in the AI-augmented world. Seek out people who use AI in their work today — engineers, doctors, entrepreneurs. Ask not just how they use the tools, but how those tools are changing the nature of their jobs. Learn what tasks have been automated, what new opportunities have opened up, and where the human role is shifting. Here's the truth: by the time you finish high school, AI will be far better at many of the skills schools still test you on. But the people who thrive won't be the ones competing with AI on speed or memory — they'll be the ones orchestrating it, combining its output with human insight, creativity, and values. Your mission over the next four years is to train yourself to think critically, work adaptively, and use AI as a force multiplier for your own ideas. That's not just how you protect your future, it's how you lead it.
Tom's Guide
11 minutes ago
- Tom's Guide
Mac Mini with M5 and M5 Pro just tipped to launch this year — here's what we know
A new report from AppleInsider claims Apple is working on an upgraded Mac mini that would feature an M5 or M5 Pro chipset. The latest leak is a follow-up to a July rumor that revealed the entire forthcoming Mac lineup through 2026. The roadmap revealed a Mac mini codenamed J837s, set to release next year. However, this new leak suggests the upgraded tiny computer might launch before the end of 2025, likely in October when Apple typically debuts its next-generation Macs. The currently available M4 Pro Mac mini had the codename J773s, with the M4 Mac Mini dubbed J773g, which lends credence to the assumption that the J837s is the M5 Pro Mac Mini. The Mac mini M4 introduced a huge redesign with plenty of ports on both the front and back of the device, more memory and a smaller footprint compared to the 2023 M2 Mac mini. It's our pick for the best mini PC, especially if you prefer Apple's OS over Windows. Coupled with the M4 chipset, it's a powerhouse, even with its controversial power button placement. That glaring flaw is fixable with some fun and clever solutions. According to AppleInsiders, the M5 mini likely won't get any design changes at the level of the M4 Mac Mini. Instead, all of the upgrades will be internal, though beyond the new chip, we're not sure what other upgrades the device is slated to receive. As for the M5 chip, it's supposedly being manufactured using TSMC's 3nm process and is meant to 'enhance artificial intelligence performance.' We do know that any new Macs will launch with macOS 26 Tahoe. Tahoe adds new Apple Intelligence features, including a better Image Playground, better Writing Tools and access to an improved Genmoji. Get instant access to breaking news, the hottest reviews, great deals and helpful tips. Apple's 'liquid glass' design is coming to Macs, bringing the cross-platform crystalline liquidity to Macs, meaning more transparent menu bars and customizable menus. We've tested Tahoe in beta, and it may be our favorite macOS update in years. The M5 and M5 Pro Mac mini are unlikely to be the only new Macs Apple launches this year. Based on the previous roadmap, we should also see a MacBook Pro M5 and M5 Max, a new Mac Pro. 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.
Fast Company
11 minutes ago
- Fast Company
Philips CEO Jeff DiLullo on how AI is changing healthcare today
AI is quietly reshaping the efficiency, power, and potential of U.S. healthcare, even as government health policy and spending drastically shift. Philips, the legacy electronics manufacturer turned medtech provider, is leading the AI healthcare revolution, streamlining and accelerating the workflow of patient care. Philips North America CEO Jeff DiLullo shares how technology can have the biggest impact on health outcomes today—from radiology scans to cancer diagnoses, and what it takes for leaders in any industry to rethink the way we work to best meet the moment. This is an abridged transcript of an interview from Rapid Response, hosted by the former editor-in-chief of Fast Company Bob Safian. From the team behind the Masters of Scale podcast, Rapid Response features candid conversations with today's top business leaders navigating real-time challenges. Subscribe to Rapid Response wherever you get your podcasts to ensure you never miss an episode. AI seems to be changing everything. There's a lot of talk about it, but in some businesses, I feel like the conversation about it is ahead of the actual implementation or the impact, and I'm curious how true that might be in medtech. How is AI impacting things now, today, versus what you think it can do in the future? If you remember, we released the Future of Health Index. One of the things that we realized is that AI, in some of these compartments I'm talking about, is quite mature. FDA cleared, very safe for clinical use. Other areas, it's more experimental. But the trust factor of the use of that AI is actually quite nascent. It's the biggest barrier right now to larger scale deployment. Yeah. That health index that you mentioned, the 2025 Future Health Index, I mean, there was this sort of trust gap in it, right? That something like 60, 65% of clinicians trust AI, but only about a third of patients or certainly older patients do. How do you bridge that gap? Is it Philips's job to bridge that gap? Whose job is it? So I have the benefit of having two Gen Zs and a millennial, they are digitally fluid. They don't worry at all about the AI models that are coming on the other side of this because they're used to it and they understand it. Older patients, not so much. The magic is always the healthcare practitioner that's directly interfacing with the customers or the patients. If they believe what they're doing, if they know it's credible, if they're using it to augment their analysis or their diagnostics, not replacing it, I think ultimately we'll see an uplift. It's our job to provide valid FDA-cleared, very good diagnostic capability leveraging AI. But if our doctors and nurses believe what we're doing and they see the value in increasing their time with patients and also a little de-stressing, we think it's going to really pick up in a parabolic way in the next few years, at least in health. I can understand and see how AI can quickly help some of the back office functionality in healthcare, but you're talking about for practitioners, right? How does that practically work today? So I'm going to give you, let's talk radiology. It's the biggest field right now, diagnostic, right? The earlier the diagnostic, the better the outcome most likely. And when I think of a radiologist, I have to wait a month and a half. I'm in a pretty nice part of Vanderbilt University area, like a lot of health tech around me in Nashville, but I've got to wait over a month to get a scan. So in radiology, we start with the box or the design, right? I have an MRI that is highly efficient. I can move it around, I can put it on a truck. But today, I can get a scan done in half or even a third of the time. The AI built into the system software makes it much faster. Just a few months ago, I had a scan that took only 20 minutes—whereas a couple of years ago, the same scan would have taken about 45 minutes. The smart speed that we have on the system actually compresses the scanning time. It doesn't fill in the blanks, it removes the noise. You actually get a better scan in a shorter time. If you're a radiologist having to do 12 or 15 studies a day, but you can do 20 studies a day, I get more patients through, I drive more reimbursement, it's better for the hospital, it's better for patient care. Then I take it into workflow, and today I can pinpoint things that are happening in that digital image and send it to a radiologist and say, 'You should look here,' in just very simple speak. It's very complicated stuff, but the AI is already mainstream today where we can actually pinpoint areas for radiologists to look at and make a determination. I can digitize the whole process today with digital pathology. And I can have a finding where somebody's waiting, do I have cancer or not? I can do this in hours now because it's all digital. And that kind of workflow and orchestration is a game changer. And the issue of AI hallucinations, which show up with some of the generative AI things, does that apply to healthcare? Are there different kinds of safeguards? Because I guess there's a human who's checking. There's so many things today, like smart speed I just talked about, being able to run that radiology workflow to compress the time of diagnostics, run the tumor boards in hours, on-demand meetings like you and I would on Zoom or teams, all of that is happening today, but not happening at the pace it could. My point is, go do that right now. Every health system, go do that. As you start to unpack these more generative AI models, I think there's real reason to be cautious and make sure we have the right controls and the governance on them, but not experimenting in them also is not an option. We kind of have to. But we see leading institutions, MGB, Stanford, Mount Sinai in New York, we see them really working with population health data to really try to train models on very specific and even broad use cases. There's so much to do right now. In other words, you don't have to go all the way out to the silver bullet of, we're going to live forever or we're going to solve every health problem. You can make the system we have right now more efficient and more effective today. Bob, when you first drove a car, was the first thing you did to go to the Autobahn? Probably not. There's so much to do in the neighborhood. There's so much to do in my town that I can really get good at what we're doing and drive productivity at scale. You need to have the innovation and the creativity to get us to the next place, but 80% of it we can do today. That is just game-changing in terms of how we deliver today, and that's what we think is really the next opportunity here for healthcare. And I think that'll happen with what's mature in AI and virtual capabilities in the next few years because the need is so great.
