TheChek News Flow

"Living in the world we live in, in which Elon runs this company and it is a private business under his control, we are living off his good graces."

SpaceX's Starlink satellite terminals have quickly become an essential tool in the arsenal of soldiers on the front lines in Ukraine. The small dishes allow them unfettered lines of communication, despite Russia's attempts to dismantle their networks.

The development has thrown SpaceX CEO Elon Musk right into the middle of a political crisis — worrying, considering the mercurial CEO's unpredictable and often brash nature.

And that has Pentagon officials quaking in their boots, as journalist Ronan Farrow writes in a sprawling new feature for the New Yorker.

"Even though Musk is not technically a diplomat or statesman, I felt it was important to treat him as such, given the influence he had on this issue," former Under-Secretary of Defense for Policy at the Pentagon Colin Kahl told Farrow.

Musk's stance on the war in Ukraine has changed considerably since SpaceX initially started handing out terminals to soldiers in early 2022.

At first, his commitment to supporting the Ukrainian cause was palpable, seemingly celebrating the delivery of hundreds of terminals.

But over the coming months, and around 20,000 terminals later, Musk became weary of the costs SpaceX was accruing, tweeting that it was "unreasonable" for the company to keep supporting the growing data usage.

His purported meetings with Russian president Vladimir Putin also raised eyebrows, crystalizing Pentagon officials' concerns over Musk's involvement in the war.

At the time, Starlink communications broke down, throwing Ukrainian lines into chaos. According to the New Yorker's sources, the outages were likely the result of SpaceX intentionally cutting the cord via geofencing.

Meanwhile, in a particularly brazen Twitter thread on October 3, Musk suggested that Ukraine should surrender and give up Crimea as a peaceful outcome of the war.

"My inference was that he was getting nervous that Starlink’s involvement was increasingly seen in Russia as enabling the Ukrainian war effort, and was looking for a way to placate Russian concerns," Kahl told Farrow.

Roughly a week after his controversial outpouring on Twitter, Musk denied having ever spoken to Putin about Ukraine, saying he'd only spoken to Putin roughly "18 months ago," and that the "subject matter was space."

Kahl, however, has a strikingly different recollection of events, with Musk reportedly looking at a Starlink activity map on his laptop during a meeting.

"This was, like, three minutes before he said, 'Well, I had this great conversation with Putin,'" Kahl told Farrow. "And we were, like, 'Oh, dear, this is not good.'"

Fortunately, for the front lines in Ukraine, Starlink connectivity has largely been stable with no major interruptions. On October 15, a clearly annoyed Musk announced that SpaceX would "just keep funding Ukraine govt for free" despite SpaceX "still losing money," a promise he has seemingly upheld ever since.

And it's not just Musk's tenuous relationship with Russia. SpaceX's decision to provide Ukrainian soldiers with Starlink has also come to alienate China — which is complicated, considering Musk's close business ties with the country. Tesla has made considerable investments in China and runs a massive Shanghai-based plant.

The very existence of Musk's Starlink constellation has even raised concerns among Chinese officials, particularly when it comes to the possibility of unrestricted connectivity within the country's borders.

Seemingly in response, China has since announced it's working on its own competing satellite network, which could allow it to rival Starlink's.

"The Starlink constellation has finally shown its military colors in the Russia-Ukraine conflict," one Beijing academic familiar with the project, told The Washington Post back in April. "The focus now is to accelerate the development of China’s own constellation... and explore defensive measures against Starlink-type foreign satellites."

In short, Musk's tremendous influence on the international geopolitical stage has clearly rattled world powers, forcing them to contend with a billionaire business magnate the likes of which the world has never seen before.

Yet despite becoming a "hyper-exposed pop-culture figure," the scope and nature of his "power are less widely understood," Farrow wrote.

"Living in the world we live in, in which Elon runs this company and it is a private business under his control, we are living off his good graces," another Pentagon official told Farrow. "That sucks."

Server production hampered by tight stocks of Nvidia processors and other components

Investors are set to assess whether enormous demand for artificial intelligence products can help offset a slump in global sales for computer hardware when Nvidia reports quarterly results on Wednesday.

The US group said in its previous earnings report that demand for its processors for training large language models, such as OpenAI’s ChatGPT, would drive up revenues by nearly two-thirds and help quadruple its earnings per share in the three months to the end of July.

The world’s most valuable chipmaker now plans to at least triple the production of its top H100 AI processor, according to three people close to Nvidia, with shipments of between 1.5mn and 2mn H100s in 2024 representing a massive jump from the 500,000 expected this year.

With AI processors already sold out into 2024, the massive thirst for Nvidia’s chips is hitting the broader market for computing equipment, as big buyers pour investment into AI at the expense of general-purpose servers.

Foxconn, the world’s largest contract electronics manufacturer by revenues, last week forecast very strong demand for AI servers for years to come, but also warned overall server revenues would fall this year.

Lenovo, the biggest computer maker by units shipped, last week reported an 8 per cent revenue drop for the second quarter, which it attributed to soft server demand from cloud service providers (CSPs) and shortages of AI processors (GPUs).

“[CSPs] are shifting their demand from the traditional computers to the AI servers. But unfortunately, the AI server supply is constrained by the GPU supply,” said Yang Yuanqing, Lenovo chief executive.

Taiwan Semiconductor Manufacturing Company, the world’s largest contract chipmaker by revenues and exclusive producer of Nvidia’s cutting-edge AI processors, predicted last month that demand for AI server chips would grow by almost 50 per cent annually for the next five years. However, it said this was not enough to offset downward pressures from the global tech slump caused by an economic downturn.

In the US, cloud service providers such as Microsoft, Amazon and Google, which account for the lion’s share of the global server market, are switching their focus to building up their AI infrastructure.

“The weak overall economic environment is challenging for the US CSPs,” said Angela Hsiang, vice-president at KGI, a Taipei-based brokerage. “Since in AI servers every component needs to be upgraded, the price is a lot higher. The CSPs are aggressively expanding in AI servers, but that was not on the cards when capital expenditure budgets were drafted, so that expansion is cannibalising other spending.”

Globally, CSP capital expenditure is expected to grow by just 8 per cent this year, down from almost 25 per cent growth in 2022, according to Counterpoint Research, as interest rates rise and businesses cut back.

Industry research firm TrendForce expects global server shipments to decrease by 6 per cent this year and forecasts a return to only modest growth of 2 per cent to 3 per cent in 2024. It points to a decision by Meta Platforms to slash server purchases by more than 10 per cent to channel investment towards AI hardware, and delays in Microsoft upgrades to its general purpose servers to free up funds for AI server expansion.

Besides the Nvidia chip shortages, analysts point to other bottlenecks in the supply chain that are delaying the AI harvest for the hardware sector.

“There is a capacity shortage both in advanced packaging and in high-bandwidth memory (HBM), both of which are limiting production output,” said Brady Wang, a Counterpoint analyst. TSMC plans to double its capacity for CoWoS, an advanced packaging technology needed to make Nvidia’s H100 processor, but warned the bottleneck would not be resolved until at least the end of 2024. The two main suppliers of HBM are South Korea’s SK Hynix and Samsung.

The Chinese market faces an additional hurdle. Although Chinese CSPs such as Baidu and Tencent are allocating as high a proportion of their investment to AI servers as Google and Meta, their spending is held back by Washington’s export controls on Nvidia’s H100. The alternative for Chinese companies is the H800, a less powerful version of the chip that carries a significantly lower price tag.

A sales manager from Inspur Electronic Information Industry, a leading Chinese server provider, said customers were demanding quick delivery, but manufacturers were experiencing delays. “In the second quarter, we delivered Rmb10bn ($1.4bn) of AI servers and took another Rmb30bn of orders . . . the most troublesome thing is Nvidia’s GPU chips — we never know how much we can get,” he said.

But once the global economy improves and the shortages abate, companies in the server supply chain could reap massive benefits, corporate executives and analysts said.

The KGI brokerage predicts that shipments of servers for training AI algorithms will triple next year, while Dell’Oro, a California-based tech research firm, expects the share of AI servers in the overall server market to rise from 7 per cent last year to about 20 per cent in 2027.

Because of the markedly higher cost of AI servers, “these deployments could constitute over 50 per cent of the total expenditure by 2027”, its analyst Baron Fung said in a recent report.

“For the supply chain, it’s just multiples of everything,” KGI’s Hsiang said. With eight GPUs in one AI server, the demand for baseboards, on which the GPU modules sit, is bound to soar compared with general servers, she said. AI servers also need larger racks on which to position the processor modules.

Why China remains hungry for AI chips despite US restrictions

The much higher power consumption of generative AI servers compared with general purpose ones also creates the need for different cooling systems and new specifications for power supplies.

Foxconn could be among the main beneficiaries of the shift because the group offers everything from the various components to final assembly. Its affiliate, Foxconn Industrial Internet, is already the exclusive provider of Nvidia’s GPU module.

For WiWynn, an affiliate of Foxconn competitor Wistron that specialises in servers, AI orders are already accounting for 50 per cent of revenues, more than double the proportion seen last year, according to Goldman Sachs.

Analysts also see a strong upside for providers of components. Taiwanese printed circuit board (PCB) maker Gold Circuit Electronics could see AI servers jump from less than 3 per cent of its revenues this year to as much as 38 per cent, Goldman Sachs said in a report in June — an expectation driven by the sevenfold increase in PCB content in AI servers over general purpose servers.

Universal Music artists, songwriters to explore royalties and content-creation tools in age of generative AI

YouTube has recruited Frank Sinatra’s estate and other Universal Music Group UMG 0.09%increase; green up pointing triangle acts to influence how artists are paid when their work is used by artificial-intelligence tools.

Universal artists including country singer Rosanne Cash, composer Max Richter and rapper Yo Gotti will work with Google’s YouTube on an effort to explore the future of music royalties and creative expression as generative AI tools make it easier than ever for internet users to manipulate art.

The partnership is an attempt by Universal to ensure its artists are compensated appropriately as the technology develops, after the music industry found itself at a disadvantage during the rise of streaming and other new services.

Creative professionals in fields ranging from film and music to book publishing have begun advocating for compensation from AI-generated content providers such as Alphabet’s GOOG 0.59%increase; green up pointing triangle Google, setting up often-tense discussions between their representatives and tech companies.

Universal and YouTube plan to collaborate on product development, principles guiding the use of music in AI programs and new ways to pay artists whose work is used in AI-generated content. Artists in the YouTube program will get early access to AI products under development and provide feedback to the video service.

“This is about having the option to design how their music is actually used,” said Universal CEO Lucian Grainge. “Artists have never had that before, to this extent, leaning into a new technology.”

The Financial Times reported earlier that Universal and YouTube were discussing ways to collaborate on generative AI products.

OpenAI’s ChatGPT bot became wildly popular when it was made available to the public last year, giving consumers access to systems that can produce passages of text in response to human prompting. Companies such as Google and Facebook owner Meta Platforms have developed other generative AI tools that help users create sounds and images from brief written descriptions.

In February, the day YouTube promoted Neal Mohan to CEO, Grainge called him, asking to talk about his approach to generative AI. They spent the following months developing guidelines for how the video service would work with artists on content and products using generative AI.

As part of principles released Monday, YouTube said it would embrace AI responsibly together with music partners, include appropriate protections when designing products, and invest in systems to fight misuse of the technology.

Any agreement between Universal and YouTube, the largest online video-hosting provider, has the potential to create standards for the rest of the music industry and other creative fields dealing with the spread of generative AI products.

YouTube has a mixed history working with music labels, which for years accused the video service of unfairly profiting from artists when their music was included in user-generated content. The Google unit said last year it paid the industry more than $6 billion from advertising and subscription fees during a recent 12-month period.

Universal in April lodged a copyright-infringement complaint with YouTube and other streaming platforms for hosting a song that replicated Drake and The Weeknd’s vocals.

The incident highlighted how AI-powered tools have made it easier for consumers to create credible songs and videos drawn from artists’ work. A YouTube spokesman said the company removed the video in question after receiving a valid takedown notice.

Universal said the training of generative AI using its artists’ music on streaming platforms was a breach of its agreements with the services and a violation of copyright law. Tech companies have a “legal and ethical responsibility to prevent the use of their services in ways that harm artists,” it said at the time.

Private-equity firm Roark Capital poised to clinch deal with $9.6 billion bid

Roark Capital is nearing a deal to buy the Subway sandwich-shop chain for about $9.6 billion.

After a long, heated auction, a deal for the closely held company could be finalized this week, people familiar with the matter said. Roark has been battling it out with a group of rival private-equity firms including TDR and Sycamore, and in recent days pulled ahead.

It is still possible the other group could come back with a higher offer and prevail.

Milford, Conn.-based Subway, known for its foot-long sandwiches and quick-service restaurants, has been owned by its two founding families for more than five decades. The Wall Street Journal reported in January that the company was exploring a sale. In February, Subway announced it had retained advisers to help run the process.

Subway is the eighth-largest U.S. restaurant chain, with $9.8 billion in domestic sales across 20,810 locations last year, according to market-research firm Technomic. It has around 37,000 restaurants globally.

Subway became one of the world’s largest restaurant chains by aggressively building new locations, but stumbled in the last decade. The chain’s global sales peaked at $18 billion in 2012, Technomic said. Stores closed and franchisees exited the system, while new sandwich-shop rivals expanded.

Late co-founder Fred DeLuca ran the chain for decades before he was diagnosed with leukemia, and his sister Suzanne Greco took the reins of the company. (She retired in 2018.) Peter Buck—who lent DeLuca $1,000 to open a sandwich shop in Bridgeport, Conn., in 1965 and helped co-found the chain—died in 2021. The DeLuca and Buck families still have interests in Subway.

John Chidsey, the chain’s first CEO from outside of Subway’s founding families, has worked to turn around the chain since taking the helm in 2019.

Subway has steadily closed U.S. locations, but is expanding abroad. It has signed 15 deals with international franchisees since 2021 and aims to open around 9,000 restaurants through the agreements. Subway said its deals include one to open nearly 4,000 locations in China in the next 20 years.

Subway has also updated its U.S. menus to try to appeal to new customers. The chain said in July that it installed deli slicers in all of its U.S. restaurants to offer freshly sliced meats for the first time. The company said it spent $80 million to provide the slicers to franchisees.

Roark, based in Atlanta, is no stranger to restaurant and food investing. Its portfolio has included pretzel purveyor Auntie Anne’s, sandwich chains Arby’s and Jimmy John’s, as well as ice cream brands Baskin-Robbins and Carvel.

DAX:

MDAX:

SDAX:

As the antibiotic pipeline empties, a fresh approach to studying bacteria has emerged

Just as most of the energy and matter in the cosmos is invisible, most of the world’s bacterial species go unseen because they cannot be conventionally grown in the lab.

Now scientists are finding ways to sift through this so-called bacterial “dark matter”. On Tuesday, an international team announced in the journal Cell that they had identified a potential new antibiotic lurking unnoticed in the sandy soil of North Carolina. The compound, called clovibactin, employs an unusual method of killing bacteria that makes it tough for targets to develop resistance. While clinical trials in humans are several years away, the finding is a glimmer of hope in an increasingly drug-resistant world.

According to The Lancet, at least 1.2mn people died as a direct result of drug-resistant bacterial infections in 2019, more than from HIV or malaria. The phenomenon of antimicrobial resistance (AMR) — in which infections become untreatable as pathogens evolve resistance to drugs — is not just a public health challenge but a drain on the economy. A landmark 2016 UK review predicted that, by 2050, “superbugs” would kill 10mn a year and cumulatively cut world gross domestic product by $100tn.

The postwar years were the heyday of antibiotic research, yielding such compounds from soil bacteria as tetracyclines, a class of broad-spectrum antibiotics, from the 1940s, and vancomycin in the 1950s. But, with only about 1 per cent of bacterial species culturable in the lab, progress began stalling in the 1980s.

That stasis gave microbes the evolutionary upper hand. As Markus Weingarth, an antibiotics researcher at Utrecht University and co-author on the Cell paper, explains, battling drug resistance depends on finding new medicines that work in different ways: “Most antibiotics are derived from natural products and there may be really novel molecules waiting for us in the other 99 per cent, this bacterial dark matter.”

These understudied species are demanding; they require special nutrients, for example, or the presence of other micro-organisms to thrive. NovoBiotic Pharmaceuticals, founded by professors Kim Lewis and Slava Epstein from Northeastern University in Boston, has been tapping into that pool by collecting micro-organisms from soil samples, mimicking their natural environment in specially designed chambers, and so cultivating “domesticated” variants capable of growing in the lab. These variants are then screened for bug-killing properties, by being placed on plates with the bacteria Staphylococcus aureus (MRSA, a strain resistant to the antibiotic methicillin, is one of the most-feared superbugs).

The clovibactin plate showed a tell-tale “zone of inhibition”, where the staph bacteria had died off. Subsequent studies showed it could also clear several different bacterial infections in mice. Critically, Weingarth, together with colleagues at the University of Bonn, found it displayed an unusual modus operandi: latching on to three different components used to build the bacterial cell wall and essentially forming a deadly cage (its name derives from klouvi, meaning cage in Greek).

That three-pronged attack, Weingarth explains, makes it tough for a bacterium to evolve resistance. Clovibactin also targets the immutable parts of those wall components, additionally lowering the chances of resistance and potentially providing a drug with a long shelf-life. According to the company, the compound is effective in the lab against MRSA, bacterial pneumonia and vancomycin-resistant enterococcus; it is now being tested against other diseases, including anthrax and tuberculosis.

Success is by no means assured, but it is at least another candidate in a relatively empty pipeline. NovoBiotic has also previously identified teixobactin and darobactin from soil. The latter shows promise against “gram-negative” bacteria such as E. coli and salmonella, which have an additional protective membrane. MDR-GNB, or multi-drug resistant gram-negative bacteria, has become a grimly familiar acronym in hospitals.

The real AMR challenge, though, is perhaps not the science but the lack of market incentive. It takes at least a decade and perhaps a billion dollars to bring a new antibiotic to market — which must then be used only sparingly. The NHS is experimenting with delinking payments and volume; the Pasteur bill, currently with the US Congress, is floating a similar subscription-based model.

It would be wonderful to think that, with all those potential superbug slayers under our feet, someone, somewhere is going to hit AMR pay dirt. But that does presuppose governments being prepared to pay for the dirt.