Tools & Platforms

Samsung flags big miss in second-quarter profit, citing U.S. AI chip curbs on China

Published

2 months ago

July 8, 2025

FILE PHOTO: Samsung projected a 56% drop in second-quarter operating profit from a year earlier, far worse than analysts expected.
| Photo Credit: Reuters

Samsung Electronics on Tuesday projected a 56% drop in second-quarter operating profit from a year earlier, far worse than analysts expected as its sales of artificial intelligence chips slowed in the United States and China.

The world’s largest memory chipmaker blamed the profit miss on U.S. restrictions on advanced AI chips for China, but analysts said the earnings slump was also due to delays in supplying chips to key U.S. customer Nvidia.

Samsung said in a statement on Tuesday that its improved high-bandwidth memory (HBM) products were undergoing customer evaluation and proceeding with shipments, without elaborating further.

“The DS Division recorded a quarter-on-quarter decline in profit due to inventory value adjustments and the impact of U.S. restrictions on advanced AI chips for China,” Samsung said in a statement.

Samsung estimated an operating profit of 4.6 trillion won for the April-June period, versus a 6.2 trillion won LSEG SmartEstimate.

That would be its weakest in six quarters, down from 10.4 trillion won in the same period a year earlier and 6.7 trillion won in the preceding quarter.

Revenue would likely fall 0.1% to 74 trillion won from a year earlier, the filing showed.

The earnings miss will fuel investor doubts about Samsung’s fundamentals, although its earnings are expected to recover gradually in the third quarter helped by rising sales of HBM chips to non-Nvidia customers and new phone launches, said Greg Roh, head of research at Hyundai Motor Securities.

“The worse-than-expected profit will be negative for investor sentiment,” he said.

Samsung said earnings in its foundry business also fell, driven by sales restrictions and related inventory value adjustments stemming from U.S. export controls on advanced AI chips for China, as well as continued low utilisation rates.

Last year, the U.S. ordered Taiwan Semiconductor Manufacturing Co to halt shipments of advanced chips to Chinese customers that are often used in AI applications, Reuters reported.

Samsung said it expects the operating loss in its foundry business to narrow in the second half of the year as utilisation improves in line with a gradual recovery in demand.

The company is expected to release detailed results including a breakdown of earnings for each of its businesses in late July.

Published – July 08, 2025 09:34 am IST

Source link

Up Next

Apply Now: $100,000 African AI Startup Training Program

Don't Miss

AI video becomes more convincing, rattling creative industry

The Editors

Click to comment

Tools & Platforms

How AI Will Unlock Small Business Growth

Published

1 hour ago

September 11, 2025

Gelila

Artificial Intelligence

getty

If AI is going to matter at all in our economy, it has to matter for small businesses first.

As a search fund entrepreneur, I’ve met and worked with more than 300 CEOs and founders in cities across the U.S. from New York to Las Vegas, Sunnyvale to Maryland. In every conversation, the same concern surfaces: the AI models in our smartphones are more advanced than the technology stacks running our businesses. While the devices in our pockets update monthly, most business systems remain unchanged for years. And nowhere is this more evident than in small businesses.

This matters because small businesses employ more than 61 million people, nearly half the private workforce. Yet just a fraction of 1% are building the kinds of technology ventures that attract institutional capital.

So this means that the overwhelming majority of small businesses are self-funded and family-run. They’re bootstrapped by owners who pour their own savings into businesses that anchor the communities they care about. These are the businesses still running payroll on technology built a decade ago.

Family Businesses Meet AI Startups

Trusted relationships build businesses.

getty

The other day, I joined a lunch meeting with 80 small business owners who lead multi-generational family firms. This gathering was a masterclass in human networks. However, not once was technology mentioned in the entire meeting.

Every three weeks, all 80 members gather at a private club on Park Avenue for a three-course meal. The purpose was to share business priorities and make referrals and introductions for each member. There were green pens and notepads on the tables embossed with the motto: “Trusted relationships build businesses.” Over lunch, if each of the 80 members received just five new customer introductions, that’s over 400 new channels opened before dessert.

Four hours later, I was downtown at the city’s newest restaurant for a startup forum. Here, every conversation was about AI and technology, from AI-powered roll-ups and cybersecurity to founders turning New York’s vacant warehouses into sushi pop-ups.

Startup Forum in NYC Restaurant

New York, NY

Two Worlds In One City

Across every city I’ve traveled to, these two business communities live side by side but rarely meet. One is led by families built through trusted introductions and intellectual property developed over decades. The other is driven by startups fueled by the race to deploy the newest technology at scale.

What happens when these two worlds connect? Imagine today’s most advanced technology powering small family-owned businesses.

For the past 40 years, one model of entrepreneurship has created more than $10 billion in value by doing exactly this: investing in established small businesses and building them with new technology and leadership. The Search fund model, first launched at Stanford in 1984, was designed to bring innovation into established firms. One of the earliest search funds invested in a 50-person roadside assistance company and built it into Asurion, now a global tech-care enterprise with 23,000 employees and 300 million customers. Another transformed a compliance services firm into RIA-in-a-Box, a leading SaaS platform used by over 2,600 firms nationwide.

What once took years and significant capital investment can now be done in months. Today, enterprise-grade tools once reserved for Fortune 500 corporations are within reach of nearly every business. The playbooks that small businesses have relied on for decades to build multi-generational, value-based businesses can, when paired with AI, scale impact in weeks instead of years.

Across core functions, generative AI is cutting work times by more than 60%. If you can sketch an idea on a napkin, it can be built in hours, not weeks. Supply chains, compliance, document processing and technical workflows are already showing double-digit productivity improvements. In some cases, technical tasks have been reduced by as much as 70%.

Search funds are one proven path to bringing technology into legacy businesses. Others are emerging as well, including AI consulting firms, AI studios and AI-powered roll-up strategies, each with their own strategies to rebuild established firms with the most advanced technology available today.

The tools are here, the cost has never been lower, and the door is wide open—for now.

If AI is going to matter at all, it has to matter for small businesses first. Once it’s put to work, it will power growth across Main Street and fuel an economy that directly supports half the workforce.

Source link

Tools & Platforms

Israel’s next strategic bet is Deep Tech, not just AI

Published

1 hour ago

September 11, 2025

The Editors

Israel’s next strategic bet is Deep Tech, not just AI | The Jerusalem Post

Deep Tech creates technologies that form the backbone of national defense and industrial sovereignty

Semiconductor chips are seen on a circuit board of a computer in this illustration picture taken February 25, 2022.

(photo credit: REUTERS/FLORENCE LO/ILLUSTRATION)

ByMERAV DAVIDOVITS

SEPTEMBER 11, 2025 08:09

Updated: SEPTEMBER 11, 2025 08:15

Source link

Tools & Platforms

Can Generative AI transform healthcare?

Published

2 hours ago

September 11, 2025

The Editors

Generative AI may be the fastest-adopted technology in history, but in healthcare it is still largely seen through the narrow lens of chatbots. That perception, argued experts at WHX Tech-EHS Summit in Dubai during a panel discussion, risks blinding policymakers and providers to its deeper potential, and to the barriers that stand in its way.

“It’s like saying the internet is only e-mail,” said Christian Hein, former Novartis vice-president for digital transformation. “Chatbots are just the front-end. The real power lies in what sits beneath, which is an information engine capable of synthesising scientific literature, drafting clinical trial protocols, automating reimbursement coding and extracting unstructured data from medical records.”

The discussion, moderated by health AI consultant Sigrid Berge van Rooijen, opened with a question: Is generative AI destined to remain a glorified customer-service tool?

For Tatyana Kanzaveli, founder of Open Health Network, the danger lies in merely bolting new technologies onto outdated systems. “We cannot just deploy GenAI to augment old business processes,” she said. “Imagine agentive AI predicting when MRI equipment is about to fail, ordering the part, scheduling the engineer and coordinating the fix automatically. Or a digital twin monitoring your health data, arranging prescriptions, transport and care without you lifting a finger. That is the world we should be building.”

Bharat Gera, who has spent 25 years working on digital health transformation, echoed the need for caution but also saw promise in simple tools such as summarisation. “Doctors spend huge amounts of time reading patient histories. Summarisation is a powerful use case, here and now,” he said. But he warned against overloading clinicians with alarms and unvalidated signals: “Healthcare is fundamentally human. If we forget that, technology will make things worse.”

Regulation, risk and responsibility

If technology is racing ahead, regulation is struggling to keep pace. Amil Khanzada, CEO of Virufy, highlighted how laws differ dramatically across jurisdictions. “In Dubai, anonymised medical data cannot be sent overseas. In Pakistan, there isn’t even a privacy law yet. Patients have the right to delete their data, but what happens once that data has already trained a model? Do you retrain it from scratch?”

Consent forms, he added, are another minefield. “You can try to use generative AI to summarise them, but you still need human validation. And patients often sign without reading. The legal and ethical risks are enormous.”

Kanzaveli pointed to the dangers of misplaced trust. “Generative AI is persuasive. You trust it. But in healthcare, a wrong answer can mean a missed diagnosis, or worse. We spent longer building the risk-management framework for a virtual psychologist than we did building the engine itself. That is our responsibility.”

The human factor

Perhaps the most sobering intervention came from Anne Forsyth, Vice-Chair of Digital Health Canada and IT lead at Toronto’s Women’s College Hospital. She recounted how a cancer diagnosis was delayed for a year because test results were stuck in a hospital IT interface. “What do you tell the patient? Tech will never be perfect. We must always plan for failure,” she said. “If you are building GenAI tools for hospitals, think about what happens when they fail, and what supports clinicians will have.”

Hein agreed. “Technology is easy. Change is the hard part. The real work is persuading people that you are there to augment, not replace them. Without that, AI will never scale.”

A future too important to ignore

Despite their differing emphases, the panellists agreed that generative AI is already reshaping healthcare and that ignoring it is not an option. “There is no industry that can remain competitive without deploying these technologies,” Kanzaveli said. “The only question is how responsibly we do it.”

As Berge van Rooijen concluded, the challenge is not whether generative AI is more than a chatbot. It clearly is. The question is how to harness its promise without repeating the mistakes of past digital health revolutions, and without losing sight of the people at the heart of the system.

Source link