Advanced Micro Devices stock surged on Wednesday after the company posted solid first-quarter earnings figures and raised its full-year revenue growth forecast on Tuesday.
AMD turned in first-quarter net income of $555 million on revenues of $3.45 billion.
Diluted earnings per share hit $0.45 compared to $0.14 in the same period a year ago.
The company also more than doubled its all-important data center revenue in the quarter, according to comments from its CEO, Dr. Lisa Su. The data center beat comes after Intel said its data-center sales dropped by 20% last week.
“Our business continued to accelerate in the first quarter driven by the best product portfolio in our history, strong execution, and robust market demand,” Dr. Lisa Su said.
“We had outstanding year-over-year revenue growth across all of our businesses and data center revenue more than doubled. Our increased full-year guidance highlights the strong growth we expect across our business based on increasing adoption of our high-performance computing products and expanding customer relationships,” the CEO added.
AMD’s computing and graphics segment revenue hit $2.10 billion in the first quarter, up 46% year-over-year due to Ryzen processor and Radeon graphics sales growth.
Management also revealed a positive outlook for 2021 in the first-quarter earnings release.
AMD now expects revenues to rise 50% from a year ago, implying a full-year revenue figure of $14.64 billion, compared with its previous forecast for a 39% revenue jump.
The CEO also addressed the semiconductor shortage that has been plaguing tech companies for the past few quarters in the earnings report.
“The entire semiconductor supply chain is very, very tight,” Su told analysts on the quarterly conference call. “That being said, we’ve been working very closely with our supply chain partners. We have seen improvements that have led to the improved full-year guide.”
Analysts remain mostly bullish on shares of AMD. The company boasts 29 “buy” ratings, 13 “neutral” ratings, and just two “sell” ratings.
Most recently, Raymond James tagged the firm with an “outperform” rating and a $100 price target citing its “durable technical advantage” over Intel on April 15.
AMD shares traded up 2.56% as of 9:48 a.m. ET on Wednesday.
Bank of America gave three reasons for investors to be bearish about Intel despite the addition of Pat Gelsinger as CEO and a $20 billion move into the semiconductor production business on Friday.
In a note to clients, analysts led by Vivek Arya reiterated their “underperform” rating and $62 price target on shares of Intel.
The analysts said that the Santa Clara, California-based company has been hurt by rising competition from Advanced Micro Devices (AMD) and others. The team expects “muted” earnings growth over the next three years.
Arya and his team highlighted three specific reasons investors might want to consider alternatives to Intel shares moving forward.
Lack of sales growth
The first reason BofA believes Intel could struggle moving forward is that the firm was unable to grow sales in “a year when PC units and cloud Capex are growing 14-15%.”
Intel reported revenues of $19.7 billion for the first quarter of 2021, a 0.7% drop year-over-year, in Pat Gelsinger’s first earnings report as CEO on Thursday.
BofA’s analysts said that much of the quarterly drop in revenues was due to rising competition from AMD as well as a move to “insourcing” from Apple and Amazon.
Falling gross margins
Intel’s gross margins fell to 55.2% in the first quarter of 2021 compared to 60.6% in the same quarter last year. According to BofA’s analysts, that’s the lowest the firm’s margins have been since 2009.
Arya and his team said that they believe gross margins in the second half of 2021 will be even lower as well, at 55%-55.5%, due to rising depreciation expenses and incremental structural headwinds.
A potentially unprofitable move into semiconductor production
Finally, BofA said that Intel’s move into the foundry business (semiconductor production) is likely to be an expensive and unprofitable transition.
The foundry business is known for having lower margins, and BofA says the company’s limited experience and potential conflicts of interest could lead to poor results.
Intel has also been forced to lower buybacks in order to build infrastructure for its future foundry volumes, creating another headwind to EPS.
The BofA analysts concluded by saying they prefer Intel’s competitor AMD for investors in the coming years.
“We continue to prefer Buy-rated Advanced Micro Devices, which should grow 37% this year and can capitalize on INTC’s process technology missteps and foundry distraction, especially as AMD continues to gain customer share with its own consistent execution and solid pipeline,” the BofA team wrote.
Raymond James initiated coverage on Advanced Micro Devices (AMD) with an “outperform” rating and a $100 price target on Thursday.
The price target represents a potential 19% jump from Thursday’s intraday highs.
In their note to clients, analysts led by Chris Caso cited AMD’s “durable technical advantage” over Intel and growing server business as key reasons for their bullish view.
The analysts said that Intel’s decision to stick with internal manufacturing has cemented AMD’s technology lead through 2024 and that the stock’s recent pullback is a buying opportunity.
“We think the stock’s pullback has been driven by improved sentiment that Intel will solve their manufacturing challenges, which will reverse AMD’s successes. We’re taking the other side of that view,” Caso and his team wrote.
“Now that Intel has committed to internal manufacturing, we think it’s unlikely that Intel ever regains a transistor advantage vs. AMD,” Caso added.
Intel announced last month it would double down on its in-house chip manufacturing business with plans to spend $20 billion on two new Arizona factories. The Santa Clara, California-based firm also plans on opening up its chip foundries to other companies so they can build their own designs.
Raymond James analysts explained how Intel’s move to stick with its 7 nanometer(nm) process for internal manufacturing while AMD is moving to Taiwan Semiconductor’s 5nm process next year-and likely to 3nm by 2024-is a big problem for the firm.
According to the analysts, the decision means AMD will hold a transistor advantage over Intel for at least the next three years.
Caso and his team also discussed cloud market share growth in their note to clients, calling it an important driver for AMD moving forward.
The analysts said the launch of AMD’s ‘Milan’ chip for data centers represents the firm’s first move into the enterprise server market and that a number of server OEMs are launching AMD designs for the first time this year. The team of analysts expects 59% year-over-year growth in the segment.
As far as risks to AMD’s rise, Raymond James said a slowdown in PC sales could hurt revenue growth and that they “believe pandemic PC purchases pulled forward demand for several years.”
However, the investment bank’s analysts noted that AMD’s increasing market share of PC sales and enterprise servers will mitigate much of the demand drawdown.
Finally, Raymond James expects 2022 earnings per share to hit $2.81, 12% ahead of the Street’s consensus estimates. The analysts used a ~36x multiple on their 2022 EPS estimate to reach their $100 price target.
The team said they believe much of the bear case around AMD is due to fears of Intel’s resurgence, but they “don’t expect there to be much to catalyze those fears for a long while.”
The news sent shares of Nvidia surging by as much as 4% on Tuesday, while both Intel and AMD fell by about 4%. Nvidia has historically manufactured premium GPU processors, also known as video cards, with its target customer being PC gamers.
“The result of more than 10,000 engineering years of work, the NVIDIA Grace CPU is designed to address the computing requirements for the world’s most advanced applications, including natural language processing, recommender systems and AI supercomputing,” Nvidia said.
After losing its chip manufacturing edge, Intel has been stuck in a business-model identity crisis. Should it split its manufacturing and design operations? Should it produce chips for other companies? Should it outsource its own production? Intel CEO Pat Gelisinger, just one month in the top job, gave his answer on Tuesday: Yes to all of the above!
Intel will double down on chip manufacturing, pouring $20 billion to build two new fabs in Arizona. The fabs will produce the most advanced chips for Intel, and for outside customers – a shrewd move that allows Intel to benefit from US and European anxieties about dependence on China, and to (eventually) tap into all the new chip buyers that have created today’s shortages.
Perhaps unconsciously, Nadella channeled his predecessor former Microsoft CEO Steve Ballmer. “Creation, creation, creation” he told Bloomberg in an interview last month, describing his view that internet creators will drive growth in the cloud business. Nadella’s phrasing, if not his delivery, was reminiscent of Ballmer’s meme-famous war chant from two decades ago of “Developers, Developers, Developers!”
WeWork’s SPAC Shaq attack. The office space sharing business has only gotten uglier in the 18 months since WeWork scrapped its IPO, with the pandemic turning downtown business centers into ghosttowns. But WeWork will get its public listing after all, thanks to the SPAC boom. The company will merge with BowX Acquisition Corp, a blank-check company that counts basketball legend Shaquille O’Neal as an advisor, in a deal that values WeWork at $9 billion – about one quarter of its $47 billion valuation in 2019.
Better than Zoom. Coronavirus vaccines have put the end of lockdowns in sight. But for those who’ve decided they actually enjoy staying indoors, the cloisered life won’t have to mean taking a vow of celibacy according to London-based Raspberry Dream Labs. The company is developing a virtual reality set up that delivers sounds, visuals and scents, as well as haptic pulses that provide a sense of being touched. Pandemic or not, the company believes the future of intimacy is remote.
“I don’t think we can expect that any platform will find every instance of harmful content. I think we should hold the platforms to be responsible for building generally effective systems of moderating this content.”
– Facebook CEO Mark Zuckerberg at Thursday’s Congressional hearing responding to a question about whether he should personally be held liable for damages caused by misinformation on Facebook.
Taiwan Semiconductor’s stock slipped on Wednesday after Intel announced plans to spend $20 billion on two new chip factories in Arizona.
Intel’s new CEO Pat Gelsinger laid down plans to directly compete with Taiwan Semiconductor and its South Korean rival Samsung in the foundry business creating chips for companies around the world.
“Intel is back. The old Intel is the new Intel,” Gelsinger said in a March 23 virtual presentation. “We’re going to be leaders in the market and we’re going to satisfy the new foundry customers, because the world needs more semiconductors and we’re going to step into that gap in a powerful and meaningful way.”
Intel’s Arizona factories will create over 3,000 permanent high-tech jobs, 3,000 construction jobs, and roughly 15,000 local long-term jobs for the region.
Taiwan Semiconductor shares continued their more than month-long fall on Wednesday after the Intel news broke. The stock is down roughly 20% from February 16 record highs.
In a note to institutional investors on Wednesday, Wedbush’s Brad Gastwirth said he believes TSMC’s fall on the Intel news is overdone.
The Chief Technology Strategist said he sees Intel’s new foundry business as an expensive move that could have an “elongated” production timeline, not producing any revenues for Intel until 2022 or 2023.
Taiwanese Economy Minster Wang Mei-hua added a similar opinion when leaving parliament on Wednesday, saying Intel’s $20 billion investment wouldn’t be a challenge to Taiwan Semiconductor and other Taiwanese chip makers, Reuters reported.
Taiwan Semiconductor has enough of its own problems without Intel’s move into the foundry business.
A water shortage has threatened production at TSMC for over a month now, and Bloomberg reported on Wednesday that the water supply to chip makers has been cut entirely as a drought continues to lower water reserves in the island nation.
Taiwan Semiconductor traded down 4.13% as of 9:58 a.m. ET on Wednesday.
A CPU, or central processing unit, is the most important part of your computer. It’s found on the motherboard and is responsible for executing every command you or an app makes. It opens apps, loads data, shows images – the CPU is involved with nearly everything on your computer.
Your desktop computer or laptop isn’t the only device with a CPU chip. Your phone has one, your video game console has one, your smartwatch has one. If you’ve bought a car made in the last ten years, it probably has a CPU for its dashboard screen.
The CPU is a critical part of any modern device. Here’s what you should know about it.
How CPUs work, and the different types
The CPU is the foundation of your computer, which is why it’s generally the first thing listed in any computer ad. Without a CPU, your computer won’t turn on, much less be usable.
Whenever you try to open a program or file, or type something, data is sent to the CPU. The CPU then decodes the data, and decides whether the command can be done. If it can, it’ll be done. If it can’t, you’ll probably see the program you’re using crash, stop responding, or give you an error message.
Since the CPU is handling data from every part of the computer at once, it’s easy to overload a CPU by flooding it with commands. This is why your computer goes slower when you have too many programs open – you’re constantly sending new data for the CPU to deal with.
And if your CPU is forced to deal with too much at once, you could crash your computer.
There are two major types of CPUs: x86 chips, and ARM chips.
x86 CPU chips
Among Windows-based PCs, x86 chips are far more popular, and are usually made by either Intel or AMD. Most Intel CPUs are a part of the Intel Core family, and have names like the Intel Core i3, Intel Core i5, and Intel Core i7. On the AMD side you have Ryzen CPUs, like the Ryzen 5 and Ryzen 9.
As a general rule, the higher the number in the CPU’s name, the faster it will perform.
Although some of these CPUs are faster than others, they’re all largely interchangeable because they all use a “x86 instruction set” – that’s where the name comes from. In short, they all speak the same language, and can all understand and perform commands that the rest of the computer sends.
ARM CPU chips
Mobile devices, new MacBooks, and some Windows computers use ARM chips. These are less powerful and take less energy, making them perfect for smaller devices. They run a bit slower, but not enough that the average user will notice.
Apple has invested heavily into ARM chips over the past few years. Their most recent set of CPU chips, the M1s, were specially designed to run new MacBooks.
What it looks like inside your CPU
Modern CPUs typically feature billions of transistors etched into the silicon wafer that comprises the chip. The more transistors, the more powerful your CPU.
The average transistor count has jumped dramatically over the years; the Intel 80386 processor, an important CPU in the 1980s, had 275,000 transistors. Just a few years later, the 80486 had 1.2 million transistors. By 2016, the Intel Core i7 processor found in many mainstream PCs had 3.2 billion transistors. The Apple M1, found in the latest MacBook Pro, has 16 billion transistors.
The transistor count is far from the only factor that determines the speed and performance of modern CPUs. Clock speed is also a critical consideration.
The clock speed of the CPU – measured in gigahertz (GHz), or a billion cycles per second – is a measure of how many instructions the CPU can perform in a particular period of time.
You also have to consider how many cores your CPU has. Most modern CPUs are multi-core, which means they can divide incoming data and decode all the parts at once. It’s like having six workers in an office compared to just one – much more efficient. Programs which are optimized for multi-threaded and multi-core CPUs can run dramatically faster.
Another component that affects the performance of a CPU is the cache. Most CPUs have multiple caches, such as L1, L2, and L3. Each cache has a specific purpose. The L3 cache, for example, is faster than the computer’s main memory and is used to feed data and instructions to the cores without bothering the other caches.
Many CPUs also contain a graphics chipset, which allows it to control the computer’s display. High-performance computers might have a separate graphics card or chipset, in which case the graphics data is handed off to that separate card.
If your CPU isn’t going as fast as you’d like, you can try overclocking it. This is a process where you force the CPU to run faster than it’s designed to go. It can give you better performance, but can harm your CPU if you overuse it.
Despite all this, the best way to find a good CPU is pretty simple: Just look for higher numbers. The newer your CPU’s model number is, the more cores it has, the faster its clock speed, the better it is.
Technologists are now exploring the power of quantum computers that are 100 million times faster than any classical computer that will, in theory, be able to solve computation problems deemed impossible today. The appeal of quantum computers is the promise of helping to quickly answer questions so difficult that it would take decades for today’s computers to solve.
“The differences between quantum computers and classical computers are even more vast than those between classical computers and pen and paper,” Peter Chapman, CEO of quantum startup IonQ, told Insider. “Because quantum computers process information differently, they are expected to be able to address humanity’s greatest challenges.”
Regular computers use bits to store information that only has two states: zero or one. Quantum computers, however, allow subatomic particles to exist in more than one state simultaneously so that they can exist as either a zero, a one, or both at the same time.
Quantum bits, called “qubits,” can thus handle a much vaster amount of information much faster than a normal computer.
Quantum computers are not meant to replace typical computers. In practice, they will be separate instruments used to solve complex, data-heavy problems, particularly those that make use of machine learning, where the system can make predictions and improve over time.
Experts expect quantum computing to help us understand biology and evolution, cure cancer, and even take steps to reverse climate change. The quantum computing market is projected to reach $64.98 billion by 2030 from just $507.1 million in 2019.
Additionally, AT&T partnered with the California Institute of Technology to form the Alliance for Quantum Technologies (AQT) with the goal of bringing “industry, government, and academia together to speed quantum technology development and emerging practical applications.” Meanwhile, quantum-focused startups D-Wave and IonQ have raised $199.69 million and $192 million respectively, per PitchBook.
One of the major goals companies are currently striving for is so-called quantum supremacy, when a quantum computer performs a calculation that no classical computer can perform in a reasonable amount of time. In October 2019, Google claimed it reached quantum supremacy, though this claim was disputed.
Some experts, like Intel’s director of quantum hardware, Jim Clarke, think that quantum supremacy is even besides the point: The real goal should be “quantum practicality” he told IEEE, referring to the point when quantum computers can actually do something life-changing and unique. Yes, quantum computers have begun completing basic tasks, but researchers are still slowly inching towards that threshold of quantum computers being able to do anything game-changing.
Experts say there’s plenty of work ahead.
“Quantum computing is easily five to 10 years out before it can actually deliver any sort of meaningful value,” VP analyst at Gartner Chirag Dekate told Insider. That’s in part because there are still so many problems to be solved around the physics of quantum computing, like stabilizing the qubits in a system.
Still, Forrester principal analyst Brian Hopkins told Insider that because there will be an exponential curve of capabilities once quantum takes off, the time for investment is now.
“That’s why smart companies are investing today,” Hopkins said. “They know when we’re going to be able to do something useful, how it’s going to impact their industry, and when we might get to that kind of point in the curve where things really take off.”
CEOs like Alphabet’s Sundar Pichai and Microsoft’s Satya Nadella are among the top 100 most overpaid CEOs, according to a new report from As You Sow.
It’s no secret that CEOs of S&P 500 companies make good money. However, As You Sow’s list doesn’t rank by the size of a CEO’s salary. Instead, the corporate responsibility non-profit uses different metrics to identify whether or not a CEO is being overpaid.
To do this, the study took three main factors into account: the amount of extra dollars a CEO receives based on past company performance and pay, the number of shareholders who voted against a CEO’s pay package, and the ratio comparing the executive’s compensation to the company’s median employee pay. The latter was weighed less heavily.
Coincidentally, the highest salary on the list happens to belong to the most overpaid CEO: Alphabet’s Sundar Pichai, who receives a pay of $280,621,552, according to the report. To compare, the median pay of Alphabet workers sits at $258,708, which is a CEO to worker pay ratio of 1,085 to one.
Pichai is being paid an excess of $266,698,263, according to As You Sow.
To compare, the median employee pay at Facebook is $247,883. This amounts to a CEO to worker pay ratio of 94 to one, lower than both Microsoft and Alphabet’s.
However, the list wasn’t just dominated by tech leaders. Bob Iger, the former CEO of the Walt Disney Company, Lachlan Murdoch of Fox Corporation, and Miguel Patricio of the Kraft Heinz Company were all listed among the top 30 most overpaid CEOs.
And according to the study, companies that have consistently graced the list are performing worse than those that have never been mentioned. As You Sow has published this report annually since 2015, and nine CEOs have made the list every year, amounting to a total pay of $2 billion. However, these nine businesses have seen a lower annualized shareholder return compared to S&P 500 companies that have never made the overpaid CEO list.
These nine companies include: Discovery, Walt Disney, Comcast, AT&T, Goldman Sachs, IBM, McKesson, Ralph Lauren, and Regeneron.
This consistent overpaying of CEOs can signal several concerns, specifically “poor accountability, weak governance, and lack of concern for shareholder interests,” the study notes.
However, this overcompensation issue may soon be changing as more shareholders are beginning to vote against these hefty CEO paychecks, according to Rosanna Landis Weaver, the report’s author.
“We might be going into a spring where we see higher votes against pay, particularly at companies that try to insulate their executive compensation from the effects of the COVID-19 pandemic,” Weaver told Insider.
These were the top 30 most overpaid CEOs, according to As You Sow’s new report:
Shares in Intel fell as much as 6% in early trading Friday, after the company said its corporate website was hacked, pushing the chipmaker to release its fourth-quarter earnings earlier than planned.
George Davis, Intel’s chief financial offer, told the Financial Times a hacker gained unauthorized access to sensitive data tied to its earnings report that was set to be published after the market close on Thursday. But upon finding out about the attack, the chipmaker released its results six minutes before the market close.
“An infographic was hacked off of our PR newsroom site,” Davis told the newspaper. “We put our earnings out as soon as we were aware.” Without providing further details, he said the breach was caused by an unlawful action that didn’t involve any unintentional disclosure by Intel.
An Intel spokesperson told Insider the company is investigating reports that non-authorized access may have been obtained to one graphic from its earnings report.
Intel’s fourth-quarter results exceeded investor expectations and beat the company’s own forecast on the back of strong PC sales. The chipmaker saw quarterly revenue fall 1% year-on-year to $20 billion, but still beat the $17.49 billion estimate of analysts polled by Refinitiv. Net income for the quarter came in at $1.52 per share, compared to $1.10 expected.
Intel’s shares closed up almost 7% at $62.46 on Thursday, but erased gains after the reported hacker’s access to information.