Technology

Why Qualcomm’s supply issues are ‘in the rear view’ during a global chip shortage

Published

4 years ago

November 5, 2021

admin

In this article

QCOM

Qualcomm CEO Cristiano Amon.

Carlo Allegri | Reuters

Qualcomm stock rose more than 12% on Thursday, one day after it reported September quarter earnings that not only beat what Wall Street expected, but also included bullish guidance for the December quarter.

Part of the reason for the strong guidance is that Qualcomm, a leading semiconductor company, is more optimistic about the global chip shortage than many of its rivals. For example, Apple says chip shortages will cost it more than $6 billion in the December quarter.

Qualcomm CEO Cristiano Amon said on Wednesday that it expected its own supply issues to be materially better by the end of December and the company will have enough supply to meet demand by the second half of next year.

That’s sooner than predictions about the end of global chip shortage from Intel, which predicts that shortages will persist through 2023, and closer to AMD‘s forecast, which says that challenges related to chip supply will persist until the second half of 2022.

“We had been cautious on Qualcomm ahead of supply issues but those are fading into the rear view now,” Goldman Sachs analyst Rod Hall wrote in a note on Wednesday.

Amon said Qualcomm’s ability to increase chip revenue 56% during a global shortage was the result of the company’s moves from earlier this year, and that new capacity from suppliers that was planned months and years ago is starting to come online.

“Supply worked exactly as we planned,” Amon told CNBC on Thursday. “Scale helps, we addressed the issue early … we put capacity plans in place and it’s working exactly as we planned.”

Here’s why Qualcomm was able to navigate the ongoing chip shortage and why it’s optimistic about next year.

Multiple suppliers

Qualcomm’s biggest individual line of business is in systems-on-a-chip, or SoCs, that combine central processing with cellular connectivity, and are the most expensive and most important component in an Android smartphone. Nearly every top-tier Android smartphone uses a Qualcomm Snapdragon chip.

Sales for handset chips grew 56% annually in the September quarter, Amon said.

These chips are made using what is called leading node processes, or the most advanced and capital intensive chip manufacturing techniques. Leading node processes create smaller transistors, which can be packed tightly together, creating faster chips that use less power and therefore more desirable smartphones.

It turns out, Qualcomm is able to manufacture its processors using two different foundries, or chip factories. Currently, Samsung and TSMC are operating the most advanced leading node, called 5-nanometer, so Qualcomm is buying from both.

“We are one of the few companies that have the ability to do multi-sourcing at the leading node, and we have done a lot of that with our roadmap,” Amon said in April.

That’s in comparison to companies like Apple, which relies on one supplier — TSMC — for its own SoCs.

On Wednesday, Amon credited double sourcing as a major reason that it was able to increase chip sales, and said that the company had three different parts on sale that were coming from two sources.

“We act early, we put a lot of things in place, multi-sourcing, capacity expansions, and we said that we expect to see material improvement in our supply towards the end of the calendar year,” Amon said Wednesday on a call with analysts.

Match issues and moving upmarket

However, other executives have said in the past month the main issue isn’t with leading node chips, but instead on the less-advanced but still essential chips, like display or power chips.

Both Intel and AMD’s CEOs have called this a “match set” issue, where PC makers “may have the CPU, but you don’t have the LCD, or you don’t have the Wi-Fi,” as Intel CEO Pat Gelsinger said in an interview last month.

Qualcomm supplies more smartphone makers than PC original equipment manufacturers, but it’s facing the same issues, said Qualcomm CFO Akash Palkhiwala.

“We’re definitely seeing some mismatch of parts in the short-term at some of our customers,” Palkhiwala said. “But you should think of those as really timing issues.”

Qualcomm officials went on to say that when smartphone makers don’t have enough parts, they prioritize more expensive models. Premium phones use Qualcomm’s most advanced processors, which cost more, and the company is able to “allocate” its supply capacity to prioritize more profitable chips.

Unit sales of premium devices with the most advanced Qualcomm chips increased 21% in the September quarter, Qualcomm said.

“We’re focusing really on the premium and high-tier units, and so when our customers have supply mismatch, they actually end up supplying the premium in high-tier devices,” Palkhiwala said, saying that match issues are not “a big factor” for Qualcomm in the short term.

Qualcomm says it still has supply constraints, and that while the company would still “ship more” if it could make more, it sees the global chip shortage going according to its plans.

“We do have constraints really across-the-board and you have to figure out how the demand would have played out if there was supply across the industry,” Palkhiwala said. “But we feel pretty comfortable that the overall supply picture is playing out exactly as we had planned.”

Technology

Microsoft says employees will be expected in office three days a week

Published

4 hours ago

September 9, 2025

admin

Microsoft says employees will be expected in office three days a week

Microsoft CEO Satya Nadella speaks at Axel Springer Neubau in Berlin on Oct. 17, 2023.

Ben Kriemann | Getty Images

Microsoft said on Tuesday that employees will be expected to work in an office three days a week starting next year.

Employees that work near Microsoft’s headquarters in Redmond, Washington, or the Puget Sound area, will be required to be in the office three days a week, starting in February. After that, the policy will extend to other U.S. locations and then to international offices.

The company is sending emails to employees who live within 50 miles of a Microsoft office around Puget Sound about the shift, the company said.

“As we build the AI products that will define this era, we need the kind of energy and momentum that comes from smart people working side by side, solving challenging problems together,” Amy Coleman, Microsoft’s human resources chief, said in a memo posted on the company’s website.

Microsoft previously had a policy, dating back to the Covid-19 pandemic, that allowed most employees to work from home half the time without manager approval.

Microsoft has held several rounds of layoffs this year, but Coleman wrote that “this update is not about reducing headcount,” and instead is “about working together in a way that enables us to meet our customers’ needs.”

In its latest earnings report in July, Microsoft reported better-than-expected results, briefly lifting the company’s market cap past $4 trillion.

Don’t miss these insights from CNBC PRO

Technology

Abu Dhabi launches low-cost AI reasoning model in challenge to OpenAI, DeepSeek

Published

5 hours ago

September 9, 2025

admin

Abu Dhabi launches low-cost AI reasoning model in challenge to OpenAI, DeepSeek

Omer Taha Cetin | Anadolu | Getty Images

A new challenger in the global artificial intelligence race has entered the ring.

The Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), an AI-focused research university established by the United Arab Emirates, announced on Tuesday the release of a new, low-cost reasoning model to rival OpenAI and DeepSeek.

It comes after DeepSeek, a Chinese AI lab, earlier this year shocked the world with the release of a reasoning model called R1 which it said could outperform OpenAI but with far less training costs.

At just 32 billion parameters, MBZUAI’s model, dubbed K2 Think, is much smaller than competing systems from OpenAI and DeepSeek. It was built on top of Alibaba’s open-source Qwen 2.5 model and is run and tested on hardware provided by AI chipmaker Cerebas.

For context, DeepSeek’s R1 has a total of 671 billion parameters, which is essentially another term for the variables that an AI language model learns to understand and generate language. OpenAI doesn’t disclose the parameter counts of its AI models.

K2 Think was developed in partnership with G42, the buzzy UAE-based AI firm backed by U.S. tech giant Microsoft. The researchers behind it say it delivers performance on par with the flagship reasoning models of OpenAI and DeepSeek — despite being a fraction of the size.

They cited the benchmarks AIME24, AIME25, HMMT25 and OMNI-Math-HARD, which relate to math, coding benchmark LiveCodeBenchv5 and science benchmark GPQA-Diamond.

How did they do it?

Hector Liu, director of MBZUAI’s Institute of Foundation Models, told CNBC the team behind K2 Think were able to achieve such high levels of performance by using a number of methods.

They include long chain-of-thought (CoT) supervised fine-tuning — a method of step-by-step reasoning — as well as so-called test-time scaling, which is a technique for improving performance by allocating extra computing resources during “inferencing” — or, applying learned knowledge to data it’s never seen before.

“What was special about our model is we treat it more like a system than just a model,” Liu told CNBC. “So, unlike a regular open-source model where we can just release the model, we actually deploy the model and see how we can improve the model over time.”

There's a water crisis looming. Big Tech and AI could make it worse

“If you ask me which one of the single steps is the most important, it’s very hard to say. It’s more like a system method work where all these methods combined delivered the final result,” he added.

Why does it matter?

There are two countries on the world stage that stand out as the forerunners in the AI race: the U.S. and China.

America’s tech giants and startups like OpenAI led the early momentum with so-called foundation models, which aim to fulfill a wide range of tasks by relying on vast amounts of training data. However, DeepSeek’s breakthrough with R1 earlier this year reinforced China’s position as a formidable AI player in its own right.

More recently, the UAE has sought to position itself as a global leader in AI in a bid to enhance its geopolitical influence and diversify its economy beyond crude oil dependency.

The region can point to its AI development firm G42 as an example of how it’s gaining ground in the space. However, it faces fierce competition from neighboring Saudi Arabia, which is looking to develop full-stack AI capabilities via Humain, a company launched under the Public Investment Fund in May.

Focus on scientific breakthroughs

While K2 Think demonstrates performance on par with OpenAI, the system’s developers say the aim is not to build a chatbot like ChatGPT. Richard Morton, managing director for MBZUAI’s Institute of Foundation Models, explains the model is intended to serve specific uses in fields like math and science.

“The fact is that the fundamental reasoning of the human brain is the cornerstone of all the thinking process,” Morton told CNBC.

“With this particular application, instead of taking 1,000, 2,000 human beings five years to think through a particular question, or go through a particular set of clinical trials or something like that, this vastly condenses that period.”

It could also expand the reach of advanced AI technologies in regions that don’t have access to the kind of capital and infrastructure U.S. firms possess.

“What we’re discovering is that you can do a lot more with less,” Morton said.

Technology

China’s Unitree plans $7 billion IPO valuation, Reuters reports, as humanoid robot race heats up

Published

9 hours ago

September 9, 2025

admin

China’s Unitree plans billion IPO valuation, Reuters reports, as humanoid robot race heats up

Humanoid robot from Unitree Robotics after a boxing match during the World Smart Industry Expo 2025 at Chongqing International Expo Center in Chongqing, China on September 7, 2025.

China News Service | China News Service | Getty Images

Unitree Robotics, one of China’s hottest technology startups, is planning an initial public offering that could value the company at up to 50 billion yuan ($7 billion), and help establish itself as a global leader in humanoid robots.

So-called humanoid robots are artificial intelligence-powered machines designed to resemble humans in appearance and movement, with applications in the industrial and service sectors.

Zhejiang-based Unitree has established itself as a leader in China’s humanoid robot space, and its listing plans could make it one of the first companies specializing in the technology to go public.

The company’s fresh valuation target, first reported by Reuters, citing two people with knowledge of the plans, would mark a sharp jump from its latest fundraising round reported on in June. At the time, the company had attracted major backers such as Geely, Alibaba and Tencent.

Unitree, in a post on its X account on Aug. 27, outlined its plans to IPO, saying that it was actively advancing listing preparations and was expecting to submit the application documents in the fourth quarter of the year.

It remains unclear how much Unitree is seeking to raise in the IPO. The company recently told local Chinese media that it’s been profitable since 2020 and now has revenues exceeding 1 billion Chinese yuan ($140.35 million).

Unitree did not respond to CNBC’s request for comment.

An offering of this size would be one of the largest Chinese tech listings in recent years. The mainland stock market has been gradually reviving following years of tightened regulatory scrutiny and volatility.

Unitree’s listing plan also comes as Beijing steps up efforts to support its local champions in artificial intelligence-related industries. Its founder, Wang Xingxing, was reportedly among a group of tech leaders who attended a rare meeting with Chinese President Xi Jinping earlier this year.

In 2023, China’s Ministry of Industry and Information Technology issued guidelines for humanoid robots, calling for “production at scale” by 2025.

Competition heats up

Unitree is part of a wave of Chinese humanoid robot companies, including Agibot, also known as Zhiyuan Robotics and Galbot, a Beijing-based robotics start-up backed by the Hong Kong government.

These companies have been rushing to get their robots deployed in factories across China. EV makers like BYD and Geely have already reportedly deployed some of Unitree’s humanoid robots at their production lines.

Meanwhile, Chinese humanoid robots have taken center stage in recent publicized events such as the World Robot Conference and World Humanoid Robot Games.

As these companies rush to get their robots deployed in factories across China, an IPO could help Unitree establish itself as China’s leading firm in humanoid robots, according to Lian Jye Su, chief analyst at independent analyst and consultancy firm Omdia.

“Unitree is one of the world’s leading vendors in mobile robots and it will likely be a top player in the humanoid robotics sector,” he said.

According to estimates from Omdia, a total of 15,000 units are expected to be shipped this year, with Unitree’s share second only to its domestic competitor Agibot.

Competition is also heating up internationally. The U.S. has seen its own burgeoning humanoid robot players, such as Boston Dynamics and Figure AI, emerge. However, Tesla’s Optimus appears to be leading the pack in commercial readiness, with CEO Elon Musk previously stating plans to produce about 5,000 units this year.

Nvidia-powered robot maker Galbot hedges against U.S. trade risks with a diversified supply-chain

However, analysts have previously told CNBC that China has established an early lead in the humanoid robotics space in terms of commercial products and pricing.

A research note from Morgan Stanley last month said that the Unitree G1 was likely the most used humanoid robot globally, given its low $16,000 starting price point. Tesla’s Optimus Gen2 humanoid robot is expected to cost at least $20,000.

Meanwhile, Unitree recently unveiled a new humanoid, the Unitree R1, with a starting price of $5,900. According to Morgan Stanley, while these cheaper humanoids may not be the most advanced, they will be valuable for Unitree to collect critical data needed to train its next generation of robot models.

Still, while China may have an early lead in the commercial success of humanoid robots, analysts note that the U.S. has strengths in the broader AI robotics environment.

The U.S. has strong chipset makers like Nvidia and Intel, hyperscalers such as Google and Meta, and robotics software vendors such as Physical Intelligence and Skild AI, which give it an “equally, if not more robust” overall humanoid robot ecosystem, said Omdia analyst Su.

For example, Chinese humanoid robot makers – including Unitree Robotics – have become early adopters of Nvidia’s humanoid robot technologies. That includes Nvidia’s recently released Jetson AGX Thor platform, which enables their machines to have real-time, intelligent interactions with people.

Merrill Lynch analysts estimated in a recent research note that global humanoid robot shipments will reach 18,000 units in 2025 from 2,500 units the year prior. It also estimates a global robot population of 3 billion by 2060.