TSMC reportedly stops taking orders from Huawei after new U.S. export controls

Taiwanese Semiconductor Manufacturing Co., the world’s largest contract semiconductor maker, has stopped taking new orders from Huawei Technologies, one of its largest customers, according to the Nikkei Asian Review. The report said the decision was made to comply with new United States export controls, announced last Friday, that are meant to make it more difficult for Huawei to obtain chips produced using U.S. technology, including manufacturing equipment.

Orders taken before the ban or already in production will not be affected, if they can ship before September 14. Huawei, the world’s largest telecom equipment maker, is TSMC’s second-biggest customer after Apple. TSMC makes many of the advanced chips used by Huawei, including in its smartphones.

The U.S. Commerce Department released its new orders on Friday, which specifically target Huawei by making it harder for the company to create chips using U.S. software and technology, even in foundries located abroad.

On the same day as the Commerce Department’s announcement, TSMC said that it is opening a new $12 billion advanced chip foundry in Arizona with support from the state and the U.S. federal government. Once opened, the plant will allow more of TSMC’s American clients to fabricate their chips domestically.

TSMC’s announcement came after the Wall Street Journal reported that White House officials were in discussions with TSMC and Intel to build foundries in the U.S. in order to reduce reliance on factories in Asia and the international supply chain.

In an email, a TSMC representative told TechCrunch that the company does not disclose customers’ order details. She added that TSMC complies with laws and applicable regulations, and is “following the U.S. export rule change closely” and “working closely with outside counsels to conduct legal analysis and ensure a comprehensive examination and interpretation of these rules.”

This is the latest restriction the U.S. government has leveled against Huawei citing national security concerns. Along with ZTE, Huawei was identified as a potential threat to security by the House Intelligence Committee in 2012.

 

The two companies have denied the charges, but under the Trump administration, the U.S government’s efforts to stop both from doing business with U.S. companies has intensified. According to the Nikkei Asian Review report, Huawei anticipated the Commerce Department’s new orders and has been building a year’s worth inventory of chips needed for its telecom equipment.

TechCrunch has contacted Huawei for comment.

TSMC to build a $12 billion advanced semiconductor plant in Arizona with U.S. government support

Taiwan Semiconductor Manufacturing Co., the world’s largest contract semiconductor foundry, said today that it plans to build an advanced chip foundry in Arizona with support from the state and the United States federal government.

The announcement follows a Wall Street Journal report earlier this week that White House officials were in talks with TSMC and Intel to build foundries in the U.S., as part of its effort to reduce reliance on chip factories in Asia. Based in Hsinchu, Taiwan, TSMC provides chip components for many of the world’s largest semiconductor companies and its U.S. clients include Apple and Qualcomm.

The plant, scheduled to start production of chips in 2024, will enable TSMC’s American customers to fabricate their semiconductor products domestically. It will use the company’s 5-nanometer technology and is expected to create 1,600 jobs and have the capacity to produce 20,000 wafers a month.

The U.S.-China trade war, national security concerns, geopolitical unrest and the COVID-19 pandemic have all underscored the shortfalls of relying on foundries located abroad and international supply chains.

The U.S. government has reportedly been in talks with TSMC for months, though one sticking point for the company was the high cost of building a new foundry. TSMC chairman Mark Liu told the New York Times in October that the project would require major subsidies because it is more expensive to operate a factory in the U.S. in Taiwan.

In today’s announcement, TSMC said “U.S. adoption of forward-looking investment policies to enable a globally competitive environment for a leading edge semiconductor technology operation in the U.S. will be crucial to the success of this project.”

The company expects to spend about $12 billion between 2021 and 2029 on the project, with construction slated to begin next year.

TSMC already operates a foundry in Camas, Washington, and has design centers in Austin, Texas and San Jose, California.

White House reportedly in talks with Intel, TSMC to build advanced chip foundries in the U.S.

White House officials are talking to Intel and TSMC about building semiconductor foundries in the United States, according to a Wall Street Journal report. U.S. tech companies and the government have been trying to reduce the country’s dependence on chip factories in Asia for years, underscored by national security concerns, the U.S.-China tariff war and now the COVID-19 pandemic, which has disrupted supply chains and logistics around the world.

The WSJ also reported that some U.S. officials have also talked to Samsung Electronics about expanding its existing contract-manufacturing operations in the U.S. to produce more advanced chips.

Intel, TSMC and Samsung Electronics are able to make chips of 10-nanometers or lower, the fastest and most power-efficient chips currently on the market.

In an April 28 letter obtained by the WSJ, Intel CEO Bob Swan told Defense Department that the company is willing to build a commercial foundry in partnership with the Pentagon “given the uncertainty created by the current geopolitical situation.”

Intel already has U.S. operations that make chips for its own products, but the new factory would serve other companies as well. TSMC, a Taiwanese semiconductor contract manufacturer, would continue making chips for other companies (its customers include Qualcomm, Nvidia and Advanced Micro Devices).

The newspaper reports that TSMC has been in talks with Commerce and Defense department officials and Apple, one of the biggest clients, about building a semiconductor factory in the U.S. The company said it is considering opening an overseas plant, but has finalized a decision yet.

“We are actively evaluating all suitable locations, including the U.S., but there is no concrete plan yet,” a TSMC spokesperson told the WSJ.

Other solutions that have been proposed by U.S. officials and industry groups include government investment in the domestic chip industry to support the high cost of building foundries, tax credits for semiconductor makers to buy and install equipment at U.S factories, and implementing more export restrictions for U.S. companies that ship microchips to buyers in China.

TechCrunch has contacted Intel and TSMC for comment.

Nvidia’s Q4 financials look to brighter skies with strong quarterly revenue growth

Major artificial intelligence and graphics chipmaker Nvidia reported its 2020 Q4 financials today (the company’s fiscal quarter ends on January 26th, 2020). The company announced revenues of $3.11 billion for the quarter, a jump of 41% from the year-ago quarter and a small bump from the third quarter.

Even more importantly, the company’s gross margin improved remarkably year-over-year, moving from 54.7% to 64.9%. The company reported a net income of $950 million for the quarter. After-hours traders jumped into the stock, with Yahoo Finance reporting a roughly 6.32% increase in the company’s share price immediately following the earnings.

That positive news didn’t overcome the full-year fiscal numbers, though, which painted a more complicated picture for the company. Revenue was down slightly for the 2020 fiscal year compared to 2019, and operating expenses, operating income, net income and diluted earnings all headed the wrong way, in some cases by more than 30%.

Nvidia’s struggles in 2019 weren’t unique to the chipmaker, as last year was bruising for the chip industry overall. The industry’s total sales declined the fastest in more than a decade because of a number of factors, including less demand in some parts of the market, oversupply in other parts of the market (driving down prices and thus sales revenue), as well as on-going trade tensions between the U.S., China, South Korea and Japan.

Nvidia itself has had a huge number of ups and down in recent years. Riding the crest of the crypto wave, the company’s stock soared as crypto miners sought the company’s GPUs, which were well-positioned to handle the hashing functions at the core of many proof-of-work crypto protocols. Yet, the crypto winter crushed the stock, which saw a precipitous decline of 50% at the tail end of 2018.

The past year though has seen Nvidia turn something of a corner. It started the year with a share price of around $150, and today closed at nearly $271, a gain of more than 80%. Part of that story — as it is with the rest of the chip industry — is the sense that a whole new set of workflows (and therefore markets) are moving to silicon, including in automotive, high-performance computing (where Nvidia acquired Mellanox for $6.9 billion early last year), Internet of Things and even in 5G.

That excitement on the big corporate side has also shown up in the venture world, as well. Startups like Cerebras, Nuvia, Graphcore and more are targeting these new workflows, putting pressure on Nvidia, Intel, and other incumbents to outperform these upstarts.

Takeaways from Nvidia’s latest quarterly earnings

Nvidia has been on a wild growth ride the past five years. Surfing a wave around AI deep learning and cryptocurrency where its specialized chip architecture is among the highest performing, the company’s share price rose from the low $20s in late 2014 to eventually soar to almost $300 in September 2018. And then crypto winter set in, and within weeks the company’s market cap was sliced nearly in half as crypto miners canceled their orders and inventories at Nvidia started building up a glut of chips.

Since that nadir in late 2018, the company has mostly been on the upswing as it has pushed expansion into a variety of other verticals like automotive, most notably by announcing the purchase of Israeli chip maker Mellanox for $6.9 billion in an all cash deal.

So with its latest earnings announcement coming after the bell yesterday, the big questions were how it was continuing to navigate chip inventories, and whether its transaction with Mellanox would close. The company ultimately presented a bit of a mixed bag, and Wall Street seems to have barely budged on the stock price as we all wait resolution on some of the key questions facing the company.

Before we dive into the analysis, first the high level numbers for Q3, which ended on October 27: top-line revenues declined slightly to just above $3 billion, from roughly $3.2 billion in the year ago quarter. Gross profits were flat from a year ago, but net income was down 27% to $899 million, mostly due to higher R&D costs and lower income from operations. Earnings per share was $1.47, down from $2.02 a year ago.

Now though, there were some more interesting takeaways from the results beyond the sort of lukewarm numbers emanating off the income statements.

China trade war still affecting Nvidia through Mellanox

The renaissance of silicon will create industry giants

Every time we binge on Netflix or install a new internet-connected doorbell to our home, we’re adding to a tidal wave of data. In just 10 years, bandwidth consumption has increased 100 fold, and it will only grow as we layer on the demands of artificial intelligence, virtual reality, robotics and self-driving cars. According to Intel, a single robo car will generate 4 terabytes of data in 90 minutes of driving. That’s more than 3 billion times the amount of data people use chatting, watching videos and engaging in other internet pastimes over a similar period.

Tech companies have responded by building massive data centers full of servers. But growth in data consumption is outpacing even the most ambitious infrastructure build outs. The bottom line: We’re not going to meet the increasing demand for data processing by relying on the same technology that got us here.

The key to data processing is, of course, semiconductors, the transistor-filled chips that power today’s computing industry. For the last several decades, engineers have been able to squeeze more and more transistors onto smaller and smaller silicon wafers — an Intel chip today now squeezes more than 1 billion transistors on a millimeter-sized piece of silicon.

This trend is commonly known as Moore’s Law, for the Intel co-founder Gordon Moore and his famous 1965 observation that the number of transistors on a chip doubles every year (later revised to every two years), thereby doubling the speed and capability of computers.

This exponential growth of power on ever-smaller chips has reliably driven our technology for the past 50 years or so. But Moore’s Law is coming to an end, due to an even more immutable law: material physics. It simply isn’t possible to squeeze more transistors onto the tiny silicon wafers that make up today’s processors.

Compounding matters, the general-purpose chip architecture in wide use today, known as x86, which has brought us to this point, isn’t optimized for computing applications that are now becoming popular.

That means we need a new computing architecture. Or, more likely, multiple new computer architectures. In fact, I predict that over the next few years we will see a flowering of new silicon architectures and designs that are built and optimized for specialized functions, including data intensity, the performance needs of artificial intelligence and machine learning and the low-power needs of so-called edge computing devices.

The new architects

We’re already seeing the roots of these newly specialized architectures on several fronts. These include Graphic Processing Units from Nvidia, Field Programmable Gate Arrays from Xilinx and Altera (acquired by Intel), smart network interface cards from Mellanox (acquired by Nvidia) and a new category of programmable processor called a Data Processing Unit (DPU) from Fungible, a startup Mayfield invested in.  DPUs are purpose-built to run all data-intensive workloads (networking, security, storage) and Fungible combines it with a full-stack platform for cloud data centers that works alongside the old workhorse CPU.

These and other purpose-designed silicon will become the engines for one or more workload-specific applications — everything from security to smart doorbells to driverless cars to data centers. And there will be new players in the market to drive these innovations and adoptions. In fact, over the next five years, I believe we’ll see entirely new semiconductor leaders emerge as these services grow and their performance becomes more critical.

Let’s start with the computing powerhouses of our increasingly connected age: data centers.

More and more, storage and computing are being done at the edge; that means, closer to where our devices need them. These include things like the facial recognition software in our doorbells or in-cloud gaming that’s rendered on our VR goggles. Edge computing allows these and other processes to happen within 10 milliseconds or less, which makes them more work for end users.

I commend the entrepreneurs who are putting the silicon back into Silicon Valley.

With the current arithmetic computations of x86 CPU architecture, deploying data services at scale, or at larger volumes, can be a challenge. Driverless cars need massive, data-center-level agility and speed. You don’t want a car buffering when a pedestrian is in the crosswalk. As our workload infrastructure — and the needs of things like driverless cars — becomes ever more data-centric (storing, retrieving and moving large data sets across machines), it requires a new kind of microprocessor.

Another area that requires new processing architectures is artificial intelligence, both in training AI and running inference (the process AI uses to infer things about data, like a smart doorbell recognizing the difference between an in-law and an intruder). Graphic Processing Units (GPUs), which were originally developed to handle gaming, have proven faster and more efficient at AI training and inference than traditional CPUs.

But in order to process AI workloads (both training and inference), for image classification, object detection, facial recognition and driverless cars, we will need specialized AI processors. The math needed to run these algorithms requires vector processing and floating-point computations at dramatically higher performance than general purpose CPUs provide.

Several startups are working on AI-specific chips, including SambaNova, Graphcore and Habana Labs. These companies have built new AI-specific chips for machine intelligence. They lower the cost of accelerating AI applications and dramatically increase performance. Conveniently, they also provide a software platform for use with their hardware. Of course, the big AI players like Google (with its custom Tensor Processing Unit chips) and Amazon (which has created an AI chip for its Echo smart speaker) are also creating their own architectures.

Finally, we have our proliferation of connected gadgets, also known as the Internet of Things (IoT). Many of our personal and home tools (such as thermostats, smoke detectors, toothbrushes and toasters) operate on ultra-low power.

The ARM processor, which is a family of CPUs, will be tasked for these roles. That’s because gadgets do not require computing complexity or a lot of power. The ARM architecture is perfectly designed for them. It’s made to handle smaller number of computing instructions, can operate at higher speeds (churning through many millions of instructions per second) and do it at a fraction of the power required for performing complex instructions. I even predict that ARM-based server microprocessors will finally become a reality in cloud data centers.

So with all the new work being done in silicon, we seem to be finally getting back to our original roots. I commend the entrepreneurs who are putting the silicon back into Silicon Valley. And I predict they will create new semiconductor giants.

Why chipmaker Broadcom is spending big bucks for aging enterprise software companies

Last year Broadcom, a chipmaker, raised eyebrows when it acquired CA Technologies, an enterprise software company with a broad portfolio of products, including a sizable mainframe software tools business. It paid close to $19 billion for the privilege.

Then last week, the company opened up its wallet again and forked over $10.7 billion for Symantec’s enterprise security business. That’s almost $30 billion for two aging enterprise software companies. There has to be some sound strategy behind these purchases, right? Maybe.

Here’s the thing about older software companies. They may not out-innovate the competition anymore, but what they have going for them is a backlog of licensing revenue that appears to have value.

Samsung posts 55.6% drop in second-quarter profit as it copes with weak demand and a trade dispute

As it forecast earlier this month, Samsung reported a steep drop in its second-quarter earnings due to lower market demand for chips and smartphones. The company said its second-quarter operating profit fell 55.6% year-over-year to 6.6 trillion won (about $5.6 billion), on consolidated revenue of 56.13 trillion won, slightly above the guidance it issued three weeks ago.

Last quarter, Samsung also reported that its operating profit had dropped by more than half. The same issues that hit its earnings during the first quarter of this year have continued, including lower memory prices as major datacenter customers adjust their inventory, meaning they are currently buying less chips (the weak market also impacted competing semiconductor maker SK Hynix’s quarterly earnings).

Samsung reported that its chip business saw second-quarter operating profit drop 71% year-over-year to 3.4 trillion won, on consolidated revenue of 16.09 trillion won. In the second half of the year, the company expects to continue dealing with market uncertainty, but says demand for chips will increase “on strong seasonality and adoption of higher-density products.”

Meanwhile, Samsung’s mobile business reported a 42% drop in operating profit from a year ago to 1.56 trillion won, on 25.86 trillion won in consolidated revenue. The company said its smartphone shipments increased quarter-over-quarter thanks to strong sales of its budget Galaxy A series. But sales of flagship models fell, due to “weak sales momentum for the Galaxy S10 and stagnant demand for premium products.”

Samsung expects the mobile market to remain lackluster, but it will continue adding to both its flagship and mass-market lineups. It is expected to unveil the Note 10 next month and a new release date for the delayed Galaxy Fold, along with new A series models in the second half of the year.

“The company will promptly respond to the changing business environment, and step up efforts to secure profitability by enhancing efficiency across development, manufacturing and marketing operations,” Samsung said in its earnings release.

It’s not just market demand that’s impacting Samsung’s earnings. Along with other tech companies, Samsung is steeling itself for the long-term impact of a trade dispute between Japan and South Korea. Last month, Japan announced that it is placing export restrictions on some materials used in chips and smartphones. Samsung said it still has stores of those materials, but it is also looking for alternatives since it is unclear how long the dispute between the two countries may last (and it could last for a long time).

Japan will restrict the export of some materials used in smartphones and chips to South Korea

Japan’s trade ministry said today that it will restrict the export of some tech materials to South Korea, including polyimides used in flexible displays made by companies like Samsung Electronics. The new rules come as the two countries argue over compensation for South Koreans forced to work in Japanese factories during World War II.

The list of restricted supplies, expected to go into effect on July 4, includes polyimides used in smartphone and flexible organic LED displays, and etching gas and resist used to make semiconductors. That means Japanese suppliers who wish to sell those materials to South Korean tech companies such as Samsung, LG and SK Hynix will need to submit each contract for approval.

Japan’s government may also remove South Korea from its list of countries that have fewer restrictions on trading technology that might have national security implications, reports Nikkei Asian Review.

Earlier this year, South Korea’s Supreme Court ruled several Japanese companies, including Nippon Steel & Sumitomo Metal Corp. and Mitsubishi Heavy Industries, that had used forced labor during World War II must pay compensation and began seizing assets for liquidation. But Japan’s government claims the issue was settled in 1965 as part of a treaty that restored basic diplomatic relations between the two countries and is asking South Korea to put the matter before an international arbitration panel instead.

Devices built with Intel’s Ice Lake and Project Athena specifications will be available in time for the holidays

Even before Computex officially launched today, AMD and Qualcomm threw down the gauntlet at Intel with a new chip and a 5G PC, respectively. Today Intel responded in kind during its keynote presentation in Taipei, introducing new processors and laptops, in addition to unveiling Ice Lake, its 10th generation Intel Core chips.

Now shipping to OEMs, the 10-nm processors will increase speeds for AI computing tasks and graphics and boost wireless speeds up to three times, Intel says. Built on Intel’s Sunny Cove architecture and Gen11 graphics engine, the series includes chips with up to 4 cores and 8 threads, up to 4.1 max turbo frequency and up to 1.1GHz graphics frequency. Gen11 will enable faster graphics in laptops, 4K HDR in a billion colors and games with up to two times faster frames per second, Intel claims. With Thunderbolt 3 and Intel Wi-Fi 6 (Gig+) inside, the company says the chips will also enable up to three times faster wireless speeds. Devices powered by Ice Lake are expected to be available for purchase by the holidays.

The company also unveiled Intel’s new class of laptops, Project Athena. Laptops built to Athena 1.0 specifications wake from sleep in less than a second, claim battery life of 9 or more hours under real-life conditions based on Intel’s testing conditions (with default settings, display brightness set to 250nits and continuous Internet connection with apps like Office 365 and Google Chrome running in the background) or 16 or more hours in local video playback mode. They are built with Thunderbolt 3, Intel Wi-Fi 6 (Gig+) and OpenVINO and scheduled to be available in time for this holiday season.

Lenovo’s senior vice president of consumer devices Johnson Jia, who helped launch Qualcomm’s first Snapdragon-powered 5G laptop yesterday, returned to the stage with Intel to showcase the the ultra-lightweight (1.2kg) Yoga S940 laptop, built on Project Athena, scheduled to go on sale in time for (you guessed it) the holidays.

Yesterday, AMD revealed the 12-core Ryzen 9 3900X, retailing for just half of Intel Core i9 9920X’s $1,100 starting price. Intel recaptured some thunder with its Intel Core i9-9900KS processor. Part of its 9th-generation chip series, the eight-core Core i9-9900KS is aimed at gamers who want to play and livestream at the same time. Like Intel’s other 9th-generation chips, it features mobile 5Ghz, and can run all eight cores at 5GHz all the time. Pricing has not been disclosed, but Intel announced that it will also be available by the holidays.

For gamers, Intel showed off its 9th-generation Intel Core-powered laptops Alienware M15 and M17, which boost mobile Ghz, a 8-core, 16-thread processor and faster frame rates and reaction times. The two laptops are expected to begin selling on June 11 at a starting price of $1,500.

Intel also announced that the Intel Performance Maximizer will be available for free download next month. The software makes overclocking more accessible by testing every core in a 9th-generation desktop processors and bringing it to maximum frequency.