TheChek News Flow

Miner says commodity giant’s bid undervalues the company

Canadian miner Teck Resources Ltd. TECK +1.77% on Thursday rejected an updated roughly $23 billion merger offer from Glencore GLNCY 2.17% PLC, saying its own plan to split into two independent companies was in the best interest of its shareholders.

Earlier this month, Glencore detailed a bid to combine with Teck and create two separate companies for their merged metals and coal businesses. Teck rejected that all-share offer, saying it would expose its shareholders to Glencore’s large thermal coal business. Glencore this week tweaked its proposal, offering Teck shareholders the option to take cash instead of shares in the companies’ combined coal operations.

“Glencore has made two opportunistic and unrealistic proposals that would transfer significant value to Glencore at the expense of Teck shareholders,” Teck Chair Sheila Murray said.

In rebuffing Glencore’s updated offer, Teck said Thursday that a merger would limit its ability to possibly seek other buyers, would expose the company to regulatory risk from antitrust authorities and would take too long to complete. It also pointed to environmental, social and governance concerns.

“We believe there are some significant structural flaws that have been contained in the proposal that Glencore has made,” said Jonathan Price, Teck’s chief executive.

Glencore’s tweaked proposal didn’t change the value of the proposed deal, which represented a premium of about 20% on Teck’s closing share price on March 24, and which would be one of the largest mining tie-ups in several years.

London-listed, Switzerland-based Glencore has said a combination with Teck would create a leading player in cobalt and copper, crucial for the transition to less polluting forms of energy.

Teck instead favors its own, existing separation plans. In February, the company said it planned to split into two companies, with one focused on base metals and another on coal. Shareholders are set to vote on that plan on April 26. Mr. Price said Thursday that he expects there to be strong interest in the metals company after a potential separation.

“We think the separation opens up a spectrum of opportunities for those businesses to create value,” he said.

Teck urged Glencore to engage with the company once its planned separation had been completed, and if Glencore itself decides to spin out its thermal coal business, as well as separate its oil operations.

Teck also said Thursday it would tweak its existing separation plans in response to shareholders’ concerns about the length of time it would take to split off its coal and base metals’ units. It said its coal business, to be named Elk Valley Resources Ltd., would now pay royalties to the base-metals unit, to be named Teck Metals Corp., for at least three years, rather than the original proposal of at least 5½ years.

On Thursday, proxy advisory firm Institutional Shareholder Services, recommended that Teck shareholders vote against the miner’s separation plan, because the proposal was “a less compelling outcome” than the status quo.

The new plan allows for a shorter path to full separation, something shareholders said they preferred, Teck said. The company will also cap the coal unit’s annual capital spending at 1.3 billion Canadian dollars, equivalent to $970 million, giving shareholders more certainty about how much in royalties would flow to the metals unit.

Meantime, activist investor Bluebell Capital Partners Ltd. came out against its proposal for Teck—in a letter Wednesday to Glencore’s chair and chief executive—and urged the company to spin off its thermal coal business.

A spokesman for Glencore declined to comment.

Lowest-rated ‘junk’ bonds miss out on rebound from last month’s banking crisis

Investors are shying away from the riskiest US corporate debt as fears of an impending recession fuel a growing divide between the highest- and lowest-rated companies in the $1.4tn high-yield bond market.

Last month’s banking crisis sparked a sell-off in so-called junk bonds of all stripes. But while higher-quality debt has clawed back its losses, investors have been reluctant to re-enter more speculative bets as they worry that an economic downturn could lead to defaults among the most indebted companies.

“Investors do see some type of recession coming,” said Steve Caprio, head of European and US credit strategy at Deutsche Bank. “If there is going to be some US growth and earnings slowdown, they would rather be in the higher-quality securities today.”

Average yields on double-B rated US bonds — the top rung of the non-investment-grade ladder, comprising half the overall junk bond market — have fallen to 6.8 per cent from a peak of 7.5 per cent in mid-March, trading close to levels seen in early February.

By contrast, borrowers with weaker ratings have remained under pressure. An index of triple-C and lower bonds tracked by Ice Data Services currently yields 15.3 per cent — down slightly from a high of 15.6 per cent on March 20, but still well above levels from two months ago.

The growing gulf in the high-yield market comes after the failure of Silicon Valley Bank on March 10 and a subsequent rescue deal for Credit Suisse sent shockwaves through financial markets and fuelled investor concerns over the health of the global economy.

Investors are also demanding a lower premium to buy double-B rated debt than they were in mid-March. The spread, or gap between those bond yields and ultra-safe US bonds, now averages 2.9 percentage points, down from 3.66 percentage points.

The triple-C spread is much wider at 11.41 percentage points — narrowing slightly from a gulf of 11.86 percentage points on March 24, but remaining above levels seen on March 9, the day before SVB collapsed.

Spreads are viewed as an indicator of how likely a company is to default on its obligations to lenders, with investors demanding a higher interest rate when that risk increases.

“This is an environment where you will see continued bifurcation between large, high-quality firms that have diversified business mixes, that have better operational agility, more financing options, and those who are smaller and lack any of those options,” said Lotfi Karoui, chief credit strategist at Goldman Sachs.

A dearth of junk bond issuance since SVB and fellow bank Signature failed has also helped to stabilise prices. The minimal new supply has been in stark contrast to the first two months of the year, when risky debt rallied on evidence of cooling inflation.

That static market “has continued to provide a firm technical support” for participants, said Kelly Burton, high-yield portfolio manager at Barings.

Burton said she would be “pretty selective” about buying any lower-rated debt. “We’re content to be patient, sit on cash and wait for a better day, or a better issuer frankly that’s higher quality in nature — a company that either has a better capital structure in general with lower leverage or better sustainable cash flows.”

U.S. businesses with lower credit ratings or significant leverage are increasingly struggling with steep increases to debt-servicing costs

The prospects of U.S. companies with significant leverage or rated several notches below investment grade have turned bleaker in recent months, credit-rating firms say, and default rates for junk-rated companies could more than double by early next year.

While highly rated companies are proving largely resilient during the postpandemic economic turbulence, businesses with lower credit ratings and floating-rate debt are increasingly struggling with steep increases to debt-servicing costs and a possible recession as the Federal Reserve continues interest-rate hikes. What’s more, still-steep inflation and softer demand are also expected to erode some companies’ profit margins, the ratings firms said.

The higher borrowing costs for risky credit are resulting in more rating downgrades and an acceleration of defaults. Default rates for low-rated U.S. companies will likely hit 5.4% in February 2024, up from 2.5% in February 2023 and higher than the long-term average of 4.7%, ratings firm Moody’s Investors Service said in a report last month. A recession as well as an increase in unemployment and wider credit spreads, or the difference in corporate bonds compared with that of safe Treasurys, could cause defaults to rise further, Moody’s said.

“What triggers a default is becoming more relevant in 2023, because everything is kind of worsening,” said Christina Padgett, head of the leveraged finance practice at Moody’s.

Many defaults are expected to come from companies working through a distressed-debt exchange, a path for companies to lessen financial burdens and preserve cash by exchanging some debt. The exchange allows companies to reduce their debt load, but the reduction is often not enough to lead to positive free operating cash flow and frequently is followed by additional defaults, said Gregg Lemos-Stein, chief analytical officer for corporate ratings at S&P Global Ratings.

Ratings firms classify companies pursuing a distressed-debt exchange as being at risk of a conventional default, with the intention of reviewing their financial prospects after the exchange and potentially raising the rating. “A distressed exchange is an attempt to avoid a formal restructuring process, but those attempts aren’t always going to be successful,” said Bradley Rogoff, head of fixed income, commodities and currency research at Barclays PLC, a U.K. banking group.

Speculative-grade-rated companies such as used-car retailer Carvana Co. and office-sharing firm WeWork Inc. were downgraded by at least two ratings firms in recent weeks while attempting a debt exchange.

Ratings firm S&P Global Ratings on March 23 downgraded Carvana’s issuer credit rating by three notches, to CC, amid the retailer’s proposed offer to swap some of its outstanding senior unsecured debt into $1 billion of new second-lien secured debt, at a substantial discount to face value. Moody’s on March 29 dropped its corporate family rating—a type of long-term rating reflecting the likelihood of default—by three notches, to Ca, in part reflecting Carvana’s weak operating performance and negative free cash flow generation.

Alan Hoffman, Carvana’s head of corporate affairs, said in a statement that the company is successfully executing on its operating plan and expects to continue making progress on its path to profitability.

S&P on March 21 also downgraded WeWork’s issuer credit rating by three notches, to CC, after it reached a deal to cut its debt by roughly $1.5 billion and extend some maturities. Ratings firm Fitch Ratings on March 27 downgraded WeWork’s long-term issuer default rating by two notches, to C. The company’s sustained negative Ebitda, or earnings before interest, taxes, depreciation and amortization, and free cash flow are key limiting factors that restrict its operating and financial profile, Fitch said. WeWork declined to comment.

For Carvana and WeWork, S&P said it expects to lower the rating to D, or default, if they complete their exchange offers, calling their capital structures unsustainable.

Distressed-debt exchanges are usually a lever that companies pull to avoid filing for chapter 11 bankruptcy protection, which is particularly costly and generally forces equity and ownership to change hands.

Companies approaching a default, for example those incapable of generating cash flow at the rate at which they would have to refinance, have limited options for avoiding a default-related credit rating. A distressed publicly traded company probably will have trouble issuing equity and would need to focus on slashing costs and demonstrating it can generate sustainable cash flow, potentially paving the way for a refinancing, Mr. Lemos-Stein said.

S&P so far this year through the end of March has issued 85 U.S. corporate credit downgrades—for both high- and low-rated companies—up from 48 during the prior-year period. There were 41 S&P downgrades in March, the most in a single month since May 2020. The total number of downgrades in the first quarter exceeded that of upgrades, which stood at 55 at the end of March, down from 80 during the prior-year period.

Meanwhile, U.S. leveraged-loan default volumes totaled $12.6 billion this year through March, up from $4.3 billion during the prior-year period, according to Fitch Ratings. Defaults totaled $26.6 billion in 2022, up from $9.7 billion a year earlier.

The rise in defaults marks a normalization after a year and a half with almost no defaults, as companies’ efforts to preserve margins and profitability came at the expense of draining liquidity positions they strengthened in response to the pandemic, said Lotfi Karoui, the chief credit strategist at Goldman Sachs Group Inc.

“I’ve been surprised by how fast companies have burned that massive amount of excess liquidity that they had built in 2020 and 2021,” Mr. Karoui said.

Many companies haven’t yet faced the full impact of higher debt-servicing costs because they haven’t needed to refinance their debt, S&P’s Mr. Lemos-Stein said. About $696.5 billion in U.S. corporate debt is coming due in 2023 and roughly $982.7 billion due in 2024, 15% and 25% of which comprise speculative-grade debt respectively, according to S&P.

“Some of the pain is still to come once those companies have to refinance and that’s why we see some companies waiting even into the final year,” he said.

The field of image generation moves quickly. Though the diffusion models used by popular tools like Midjourney and Stable Diffusion may seem like the best we’ve got, the next thing is always coming — and OpenAI might have hit on it with “consistency models,” which can already do simple tasks an order of magnitude faster than the likes of DALL-E.

The paper was put online as a preprint last month, and was not accompanied by the understated fanfare OpenAI reserves for its major releases. That’s no surprise: This is definitely just a research paper, and it’s very technical. But the results of this early and experimental technique are interesting enough to note.

Consistency models aren’t particularly easy to explain, but make more sense in contrast to diffusion models.

In diffusion, a model learns how to gradually subtract noise from a starting image made entirely of noise, moving it closer step by step to the target prompt. This approach has enabled today’s most impressive AI imagery, but fundamentally it relies on performing anywhere from 10 to thousands of steps to get good results. That means it’s expensive to operate and also slow enough that real-time applications are impractical.

The goal with consistency models was to make something that got decent results in a single computation step, or at most two. To do this, the model is trained, like a diffusion model, to observe the image destruction process, but learns to take an image at any level of obscuration (i.e. with a little information missing or a lot) and generate a complete source image in just one step.

But I hasten to add that this is only the most hand-wavy description of what’s happening. It’s this kind of paper:

A representative excerpt from the consistency paper. Image Credits: OpenAI

The resulting imagery is not mind-blowing — many of the images can hardly even be called good. But what matters is that they were generated in a single step rather than a hundred or a thousand. Furthermore, the consistency model generalizes to diverse tasks like colorizing, upscaling, sketch interpretation, infilling and so on, also with a single step (though frequently improved by a second).

Whether the image is mostly noise or mostly data, consistency models go straight to a final result. Image Credits: OpenAI

This matters, first, because the pattern in machine learning research is generally that someone establishes a technique, someone else finds a way to make it work better, then others tune it over time while adding computation to produce drastically better results than you started with. That’s more or less how we ended up with both modern diffusion models and ChatGPT. This is a self-limiting process because practically you can only dedicate so much computation to a given task.

What happens next, though, is a new, more efficient technique that can do what the previous model did, way worse at first but also way more efficiently. Consistency models demonstrate this, though it is still early enough that they can’t be directly compared to diffusion ones.

But it matters at another level because it indicates how OpenAI, easily the most influential AI research outfit in the world right now, is actively looking past diffusion at the next-generation use cases.

Yes, if you want to do 1,500 iterations over a minute or two using a cluster of GPUs, you can get stunning results from diffusion models. But what if you want to run an image generator on someone’s phone without draining their battery, or provide ultra-quick results in, say, a live chat interface? Diffusion is simply the wrong tool for the job, and OpenAI’s researchers are actively searching for the right one — including Ilya Sutskever, a well known name in the field, not to downplay the contributions of the other authors, Yang Song, Prafulla Dhariwal and Mark Chen.

Whether consistency models are the next big step for OpenAI or just another arrow in its quiver — the future is almost certainly both multimodal and multi-model — will depend on how the research plays out. I’ve asked for more details and will update this post if I hear back from the researchers.

Alexis Bidinot, cofondateur de l'entreprise Qiara, qui propose des alarmes connectées, vient de lever 14 millions d'euros auprès de Xavier Niel et d’Iliad. Son ambition: reproduire le coup du trublion des télécoms. Il est l'invité du Club Entrepreneurs Challenges-Grant Thornton..

Alexis Bidinot - Nous avons lancé Qiara fin 2020 car nous voulions bousculer le quasi-monopole de la télésurveillance: les principaux acteurs proposent un abonnement de 50 à 60 euros par mois ; il n’y a souvent pas de prix indiqués clairement sur Internet ; et les contrats courent sur une durée de trente-six à quarante-huit mois. Une habitude du monde des télécoms d’il y a vingt ans! Aujourd’hui, 10% des foyers en France sont équipés d’alarme. Selon différentes études, le potentiel du marché peut aller jusqu’à 44%.

La démarche rappelle celle de Free...

Effectivement, et ce n’est pas un hasard. Qiara a été cofondé par des ingénieurs qui faisaient auparavant les Freebox pour 7 millions de foyers. Xavier Niel et Iliad sont associés au projet. Je l’ai rencontré en 2017, quand je travaillais à la direction de Free. J’ai dit à Xavier: "Il faut créer une nouvelle structure, et reproduire le coup des télécoms sur un nouveau gros marché qui est en train d’exploser."

Concrètement, que proposez-vous ?

Qiara, c’est un système d’alarme connectée de nouvelle génération avec télésurveillance. Nous proposons un pack de départ à 99 euros, vendu à prix coûtant. Notre objectif: que ma grand-mère puisse le monter en cinq minutes! Il comprend cinq composants: une caméra, une sirène, un détecteur d’ouverture de porte, un détecteur de mouvement, ainsi qu’un clavier pour taper le code.

Sauf qu’il faut souscrire également un abonnement...

Effectivement, cela fonctionne par abonnement, sans engagement. Le premier est vendu moins de 10 euros par mois, le second comprend la télésurveillance pour 20 euros. Chez certains de nos concurrents, un pack de départ comme celui-là vaut plus de 1.000 euros et l’installation plusieurs centaines d’euros.

Comment avez-vous développé le produit?

Nous avons mis deux ans à mettre au point le matériel des alarmes: le design des cartes électroniques et de nombreux prototypes. Nous avons lancé l’offre début mars. Les composants viennent à 90% de France. Tout est ensuite assemblé près de Laval (Mayenne).

Pourquoi levez-vous des fonds?

Nous venons de lever 14 millions d’euros auprès de Xavier Niel et d’Iliad afin d’investir dans le marketing et d’augmenter la production. L’objectif: équiper 500.000 foyers de systèmes Qiara à l’horizon 2025.

Pourquoi avoir choisi de construire les composants en interne?

Le matériel, c’est beaucoup de galères et de complexité. La plupart des acteurs confient cette tâche à d’autres. Nous avons développé plusieurs prototypes qui finalement ne fonctionnaient pas. Sauf que fabriquer ses propres produits permet aussi de créer des Tesla et des iPhones, et c’est l’ambition que nous avons avec Qiara.

Votre rêve de croissance?

Réparer une injustice, comme c’était le cas dans les télécoms: casser un monopole en proposant plus de concurrence à un meilleur prix. Nous voulons créer un géant européen de la maison connectée et de la sécurité.

Beijing makes rare backtrack on no-fly zone after drills focused on improving invasion capabilities

The last time Taiwan’s president met a US House Speaker, China launched unprecedented military drills, simulating a complete attack on the island from missile bombardment to amphibious invasion.

But after President Tsai Ing-wen met Speaker Kevin McCarthy in California last week, China’s military response was more narrowly defined. The People’s Liberation Army mainly practised preventing Washington and its allies from coming to Taipei’s rescue if Beijing did attack the country.

“This time, there was a big focus on anti-access and area denial,” said a senior Taiwanese government official, referring to a strategy of blocking US forces from entering and operating in airspace and waters close to China.

“There were a lot of simulated attacks on aerial and sea targets, focused on keeping forces out that would be arriving from outside the island chain.”

China’s latest three-day manoeuvres around Taiwan, launched on Saturday to punish the country for Tsai’s US trip, demonstrated Beijing’s desire to refine tactics crucial for an annexation while limiting the impact of its campaign.

“The level of provocation and intensity was much lower this time than last year,” said a senior Japanese government official. “The Communist party [needs] to demonstrate their determination to uphold what they call territorial integrity, but Tsai’s US visit was less serious than [then-Speaker Nancy] Pelosi’s Taiwan visit” in August.

Beijing briefly set off alarm in Taipei on Tuesday when it told officials it would impose a no-fly zone next week north of the island, in some of the world’s busiest airspace. But a day later, it cut the planned three-day closure to just 27 minutes, a reversal that suggested some restraint.

During last year’s drills, which lasted a week, the PLA went through its complete multi-stage war plan all the way to an invasion.

“This time, the exercises looked more like a campaign to exhaust our military’s fighting power so the PLA can gain air and sea control,” said Chieh Chung, an expert on the Chinese military at the National Policy Foundation, a Taipei think-tank. “The exercise did not follow all three stages — firepower campaign, blockade and invasion campaign — like they did last August.”

Government officials and military experts interpreted the changed patterns as a sign that China was trying to make the drills less disruptive while honing specific skills the PLA needs to conquer Taiwan, which Beijing claims as part of its territory and has threatened to take by force.

Last year, Beijing had declared several areas around Taiwan off limits for ships and aircraft and fired missiles over the island. Some of the missiles landed in Japan’s exclusive economic zone, an area beyond territorial waters extending 200 nautical miles off the coast.

Taiwanese intelligence officials said China’s leadership had refrained from imposing a ring of missile-landing zones around Taiwan this time for fear of threatening supply chains and jeopardising a still-fragile economic recovery.

“Last August, sea transport volumes were still only at 60 per cent of normal levels, so the impact of a simulated blockade would have been much more significant now” than last year, said one official. “They definitely have their own considerations regarding the economic fallout.”

Despite the narrower focus, analysts said the exercise showcased the PLA’s growing capabilities. It set records for daily air incursions, with 91 aircraft operating around Taiwan on Monday, according to the Taiwanese defence ministry.

Even after Beijing declared the manoeuvres complete on Monday, Taiwan’s defence ministry said 35 Chinese military aircraft and eight naval vessels continued to operate near its borders in the 24 hours to Wednesday morning.

“This shows that the PLA’s overall fleet availability, their pilots’ capabilities and their ability to control large forces in the Taiwan Strait and in the airspace off the south-west of Taiwan have all increased significantly,” Chieh said.

The PLA Navy also conducted operations east of Taiwan, which included take-off and landing drills with fighters and helicopters from the Shandong, its newest aircraft carrier in service.

According to officials briefed on the drills, the carrier, which was accompanied by submarines, mirrored movements by the US aircraft carrier Nimitz to the north-east.

In wartime, both carrier and submarine operations would be crucial to targeting incoming US forces and destroying Taiwan’s air force, which has most of its hardened shelters on the east coast, and its navy, which under Taipei’s plans would seek to survive by sailing out into the western Pacific.

The PLA also practised so-called joint operations, the integration of naval, air and missile forces and units from regional commands.

“Following the military reforms [Chinese leader] Xi Jinping undertook in 2015, they now want to demonstrate that those are coming to fruition,” said Alexander Neill, an expert on the Chinese military at the Pacific Forum.

But the drills also highlighted the PLA’s shortcomings in carrying out — and defending — a successful invasion, experts said.

Su Tzu-yun, a research fellow at the Institute for National Defense and Security Research, a think-tank backed by Taiwan’s defence ministry, said the aircraft carrier operations showed that the PLA Navy was executing flight sorties at a much slower rate than the US Navy.

Experts also observed weaknesses in electronic warfare — operations to disrupt enemy communications and control systems — which China “had a huge problem with during the August exercise”, according to Kitsch Liao, assistant director at the Atlantic Council’s Global China Hub in Washington. He added that the PLA appeared to have made electronic countermeasures a focus in the latest drill.

Moreover, perhaps China’s biggest obstacle to an invasion of Taiwan — getting troops across the strait — remains unproven, said the Pacific Forum’s Neill.

“None of what we have seen in the drill indicates that the PLA is currently able of mounting an amphibious operation and hold the island.”

Rules to be eased on start-up financing as global volatility spooks investors

Easier regulations are on the horizon, Vietnam has told technology start-ups and investors, in contrast to the red tape that has grown amid the country’s widening crackdown on corruption.

Nguyen Duc Long, acting director of the country’s National Innovation Center (NIC), said in an interview that Vietnam would revise Decree 38, which legally defines start-ups and venture capital and lets the government invest in start-ups. But he told Nikkei Asia it would be “more difficult” to change other rules, such as those on cross-border transactions in Vietnam, which strictly controls money sent abroad, including for stock listings.

The NIC, under the Ministry of Planning and Investment, has had multiple meetings with businesses, including in October, when it listened on how to improve Decree 38.

“That will be a priority,” Long told Nikkei.

Suggestions included raising the cap on investors in a venture capital fund, letting them invest in more industries, and allowing for more borrowing options, the NIC said.

Vietnam has sought to offer reassurance amid market volatility, both inside and outside the country. Internationally, headwinds have ranged from bank panics to tech lay-offs after tougher financing conditions pushed companies to cut costs.

“Investors globally are spooked,” said Golden Gate Ventures founding partner Vinnie Lauria at a Forbes tech summit in Ho Chi Minh City at the end of March. “They’re acting irrationally. They’re acting like teenagers. So that means fundraising is always going to be more difficult in this environment.”

Long was also a panellist at the summit, where he struck a congenial chord with entrepreneurs.

“Vietnam has been very open to dialogue between the government, the policymakers . . . the investment funds and the start-up community to identify the pain points of the market,” he said, adding that on the regulatory outlook, “I think we have narrowed down what we need to do.”

But regulations are just one issue. Execution is another. Businesses across a host of sectors are complaining of bureaucratic paralysis due to the ongoing campaign against graft. Permit approvals, for example, have slowed to a trickle as officials fear making mistakes. In this way, the anti-graft campaign is “heightening the risk aversion endemic in the country’s lower to upper tiers of government”, said To Minh Son, political research assistant at Nanyang Technological University in Singapore.

Decree 38 allows VCs to use gold or even land-use rights to finance their funds, though other capital constraints remain. With little clarity on stock options or convertible loans, for example, few start-ups choose to pursue initial public offerings in Vietnam. This limits investors’ ability to exit or sell their investments, as does the difficulty of transferring cash overseas, said Hong-jin Kim, managing director of South Korea-based Stic Investments.

“Because of some kinds of regulations, because of some kinds of size [of start-ups] . . . sometimes it’s quite hard to do the IPO,” he said.