TheChek News Flow

RÉCIT - La maire de la capitale avait fait part durant la campagne des municipales en 2020 de sa volonté de «rendre le centre de Paris piéton». L’élue socialiste a fini par imposer son projet, non sans mal.

«Le centre de Paris devient une zone à trafic limité (ZTL)». Ce lundi 4 novembre, Anne Hidalgo n’a pas manqué de se féliciter sur ses réseaux sociaux. Après quatre années de dur labeur, la maire de la capitale tient enfin sa promesse de campagne de 2020 en limitant l’accès des véhicules au cœur de la capitale... Avec deux ans de retard, et au prix d’importantes concessions. Symbole de sa volonté politique de réduire la place de la voiture en ville, l’instauration de la ZTL lui a donné du fil à retordre. Et a provoqué, à l’échelle locale, des débats animés, ainsi qu’un bras de fer avec l’État et plusieurs maires d’arrondissements.

Les toutes premières traces de la ZTL apparaissent dans le programme de campagne d’Anne Hidalgo lors des municipales de 2020. Celle qui brigue un second mandat a un atout : la loi sur le nouveau statut de Paris, décidée sous François Hollande. Loin d’être anecdotique, cette réforme lui donne de nouveaux pouvoirs, notamment dans le domaine de la circulation. Cela lui permet d’imposer ses projets, comme celui, de rendre piéton le centre de la capitale. Une folie, aux yeux des oppositions, éberluées par cette nouvelle lubie anti-voitures.

Le projet émerge finalement un an plus tard, au printemps 2021. Son premier adjoint Emmanuel Grégoire et son adjoint chargé des mobilités David Belliard présentent alors officiellement le projet de ZTL et promettent un lancement au début de l’année suivante. Le centre de Paris «subit un trafic de transit particulièrement important» qu’il conviendrait de réguler, prêche l’élu écologiste, qui promeut le réseau de transports en commun dense au cœur de la ville. Une position qui n’a pas évolué depuis. «Il existe depuis déjà maintenant 120 ans un métro qui dessert la zone. Peut-être est-ce l’occasion de l’essayer ?», lance-t-il encore aujourd’hui, narquois, aux détracteurs du projet.

D’emblée, toutefois, l’ampleur du projet de la mairie fait grincer des dents. Son périmètre initial, qui comprend Paris Centre et toute la partie de la rive gauche située au nord du boulevard Saint-Germain, y compris l’Île de la Cité et l’Île Saint-Louis, pose problème. L’opposition - qui assure avoir découvert le projet dans la presse - s’insurge. Et les questions fusent, sans réponse précise : «À quoi doivent s’attendre les automobilistes ? Quels sont les dispositifs de contrôle ? Comment sera budgété le projet ?». Aurélien Véron, élu Changer Paris de Paris Centre, s’en souvient bien : «Je leur demande s’il y a une enquête publique et une étude d’impact, obligatoires pour mettre en place ce projet, et je n’obtiens aucune réponse à mes questions». De «l’amateurisme complet», déplore-t-il, décrivant une «mesure radicale et complexe» dont les conséquences «n’ont pas été réfléchies».

Maire (LR) du 6e arrondissement de la capitale, Jean-Pierre Lecoq explique qu’on lui avait d’abord présenté ce projet comme une zone apaisée, dans laquelle la circulation pourrait être limitée à 30 km/h. «L’objectif était d’intégrer dans cette zone dite “tranquille” tout le quartier situé entre le boulevard Saint-Germain et les quais de Seine, mais il n’était pas du tout question d’y contrôler les entrées et les sorties», se remémore-t-il. L’élu déchante rapidement, dès le début des réunions officielles, au lendemain du Covid. «Le projet a commencé à se durcir, et au bout de quelques mois. On m’expliquait que pour que la ZTL fonctionne, il faudrait mettre le boulevard Saint-Germain à double sens et interdire la circulation sur les quais hauts de la rive gauche entre l’Assemblée nationale et l’Institut du monde arabe. Ce n’était pas du tout ce qu’on m’avait raconté [...] C’est là où j’ai compris que la ZTL s’accompagnait de la création d’une sorte de nouvel anneau périphérique à l’intérieur de Paris», tempête-t-il.

Pour le maire du 6e, en poste depuis 30 ans, ce projet a notamment une visée électorale claire : «éminemment politique», il est le fruit d’un accord entre l’équipe d’Anne Hidalgo et les Écologistes, qui s’allient au second tour des municipales pour peser contre la droite. Pour autant, un «grand flou» demeure quant aux moyens de contrôle envisagés. «Du déploiement de policiers municipaux à la création de vignettes résidentielles... Toutes les idées ont germé», ironise Jean-Pierre Lecoq. Le projet continue néanmoins d’avancer et se précise durant l’année 2021. Une consultation citoyenne est lancée, et l’inquiétude des commerçants - déjà durement touchés par les grèves dans les transports, les manifestations des gilets jaunes, celles contre la réforme des retraites et le Covid - croît. Au point de réclamer le report de la mise en place de la ZTL «à l’horizon 2024», notamment afin d’échanger davantage avec la Ville de Paris.

Mi-février 2022, Emmanuel Grégoire annonce que la ZTL sera reportée «à début 2024, avant les Jeux olympiques». L’exécutif souhaite peaufiner les détails, alors qu’Anne Hidalgo est prise par sa campagne présidentielle. L’Hôtel de ville se donne surtout «plus de temps pour discuter avec la préfecture de police», comme le souligne alors David Belliard. Car le préfet de police de l’époque, Didier Lallement, avec qui la maire entretient des relations conflictuelles, freine des quatre fers. Dans une lettre rédigée le 17 février 2022, ce dernier rappelle que le projet de la ZTL parisienne «suppose une décision conjointe de la maire et du préfet». La Ville ne peut plus faire cavalier seul, alors que l’institution s’insurge de ne pas avoir été «saisie d’un dossier précis comportant les impacts de cette mesure sur les flux de véhicules et les reports de trafic tenant compte de l’ensemble des aménagements réalisés ou projetés». Pire, le préfet émet de «fortes réserves» sur le projet «tel qu’envisagé» et explique que ses «caractéristiques actuelles» risquent même «de créer des difficultés pour la circulation des services de secours et de police (...) et pourrait avoir un impact négatif pour l’activité économique de la capitale».

Quelques mois plus tard, en novembre, Anne Hidalgo reporte encore une fois la ZTL, cette fois-ci «après les Jeux olympiques». Elle met de l’eau dans son vin, et semble moins catégorique sur le périmètre du projet, qui reste «à définir». Le dialogue avec le nouveau préfet de police Laurent Nuñez, arrivé en juillet, semble bien plus fluide qu’avec son prédécesseur, mais les discussions pour parvenir à un accord n’en sont pas moins intenses. Car la question du périmètre pose toujours problème : la maire tient à garder la rive gauche dans son projet dont les aménagements selon elle «n’empêcheront pas» Laurent Nuñez «de continuer à mener sa mission de protection des Parisiens». Le préfet, quant à lui, s’inquiète de la bonne circulation des véhicules d’urgence et défend l’attractivité économique de la zone. Interrogée récemment à ce sujet, la préfecture de police de Paris se targue d’avoir obtenu gain de cause sur ces deux points : la rive gauche est retirée du projet et les modalités de la zone sont redéfinies avec les acteurs économiques. «Après plus de trois années d’échanges entre les services de la Ville et de la préfecture de police, le périmètre retenu est celui que nous avons proposé, c’est-à-dire une ZTL qui couvrira Paris Centre, hors quais hauts rive droite et hors îles», souligne l’institution. Avant d’ajouter que ces «nouveaux échanges avec la Ville de Paris ont permis d’aboutir à la garantie, notamment, que les autocars de tourisme ne feraient l’objet d’aucune restriction spécifique».

Un «long travail de négociation», reconnaît Emmanuel Grégoire, désormais député à l’Assemblée nationale. Et si l’élu socialiste a quitté la mairie, il défend toujours ardemment le projet. «Il ne faut pas tomber dans le piège de penser que la ZTL serait attentatoire à la circulation», plaide celui qui qualifie cette zone de «levier de transformation des habitudes», alors que la capitale est depuis trop longtemps «utilisée comme lieu de transit». Entre 350.000 à 550.000 véhicules y passeraient chaque jour, dont près de la moitié sans s’y arrêter, selon les chiffres de la Ville. Dans l’entourage de la maire, le lancement de cette ZTL est défini comme une réussite, à l’image de la fermeture des quais bas et de la rue de Rivoli à la circulation de transit. «Tout le monde disait “ça va être le chaos” mais ceux qui en bénéficient, ce sont surtout les piétons (...) On ne reviendra pas en arrière», assure-t-on.

Pour autant, le plus important reste à faire : définir les modalités de contrôle de la zone. Si la solution d’un «QR Code» semble être plébiscitée, il faudra encore définir précisément les justificatifs à présenter aux policiers municipaux. En attendant, les discussions se poursuivent à ce sujet et la Ville promet qu’aucune amende ne sera infligée pendant six mois, le temps de «faire un peu de pédagogie». «La ZTL, c’est de la communication, ce qui compte, c’est ce qu’il y a dedans», peste Aurélien Véron, persuadé que ce système «incite les automobilistes à tricher», «comme dans le Marais où tous les sens de circulation ont été repensés». Résultat selon lui : «il y a ceux qui connaissent et qui prennent les rues en sens interdit» et «ceux qui ont peur et qui contournent». Face à ce risque, les élus de la droite parisienne réclament un bilan chiffré dans les prochains mois, afin notamment de constater les conséquences de la ZTL sur les commerces. Le bras de fer est loin d’être terminé.

The Republican president-elect is inaugurating a new era in US politics and perhaps for the world as a whole

Cybersecurity continues to command a lot of attention from enterprises looking for better protection from malicious hackers, and VCs want in on the action. In the latest example, TechCrunch has learned and confirmed that Upwind — a specialist in assessing and securing cloud infrastructure — is closing in on a $100 million round at a valuation of between $850-$900 million, post-money.

The round includes a mix of new and existing investors such as Craft Ventures, Greylock, CyberStarts, Leaders Fund, Omri Casspi’s Sheva Fund, and basketball star Steph Curry’s investment fund Penny Jar. The round is in its final stages of getting closed — this could come in a matter of days — and could include further investors.

The round, a Series B, comes on the heels of the company picking up “dozens” of Fortune 500 companies and growing its staff to about 160 people, a source said.

It’s a significant leap for Upwind, which had previously raised just over $77 million, including a $50 million round in September 2023. Upwind’s last valuation, from that most recent round, was $300 million. It will be using some of the funding for R&D, and some for hiring, with plans to add around 100 people across Israel, San Francisco… and Iceland.

Upwind was founded by Amiram Shachar, who sold his previous company, the cloud spend management startup Spot.io, to NetApp for $450 million. It is part of the guard of cybersecurity startups founded in Israel by teams that cut their teeth originally working in areas like military intelligence.

In its case, it is also one of the many companies in the space focusing on vulnerabilities in the cloud through a platform approach. Specifically, Upwind aims to tackle the blizzard of alerts that are typically raised by threat detection tools. It claims to reduce these alerts by 90% to focus security operations teams more closely on understanding and responding faster to actual threats.

The company’s tech covers cloud services (covering areas like vulnerability management and identity security), workloads (which includes container security and detection and response) and apps (including areas like API vulnerability management). To some extent, these are all interrelated, one reason a platform approach makes sense.

New code introduced in the latest version of Apple’s mobile operating system could make it more difficult for both police and thieves to unlock iPhones.

404 Media reported Thursday that law enforcement officials were warning each other that phones being stored for forensic examination seemed to be rebooting themselves — something that security experts confirmed in a follow-up story. After the reboot, it’s harder for those phones to be unlocked by password-cracking tools.

Apple did not immediately respond to TechCrunch’s request for comment.

This reboot appears to take place after iPhones with iOS 18.1 have not been unlocked for a set period of time. According to Chris Wade, founder of mobile analysis company Corellium, iPhones seem to reboot after their fourth day of in a locked state.

Matthew Green, a cryptographer and assistant professor at Johns Hopkins University, described this as “a huge improvement in terms of security” that “probably doesn’t inconvenience anyone” — though the police might disagree.

With the U.S. election result and the German government’s collapse, the lagging European economy is at a crossroads

Wall Street’s verdict is clear: A second Trump presidency is likely to deliver a blow to an export-dependent European Union that is struggling with sclerotic economic growth and ever-multiplying political crises. Whether it will finally spark some change is the question for patient investors.

Since Wednesday, the day after the election, the S&P 500 has gained 3.7%. Meanwhile, the Euro Stoxx 50 and the FTSE 100 are down. Among those to shed the most market value have been clean-energy firms such as Vestas, carmakers such as BMW, consumer-goods companies such as Nestlé and Unilever and sellers of pharmaceuticals such as Roche. They all sell a lot to the U.S.

The U.S. is the top goods-export market for the European Union, and for Germany, with pharmaceuticals, machinery and vehicles topping the export list.

During the campaign, President-elect Donald Trump floated a 60% tariff on Chinese imports and a 10%-to-20% levy across the board. The think tank German Economic Institute estimates that such a measure could make the German economy between 1.2% and 1.4% smaller than it would have been by 2028.

The core of the European Union’s export machine has been plunged into difficulties because of the end of cheap Russian energy, delays in joining the electric-vehicle revolution and an overreliance on selling to China.

Volkswagen last week announced the closing of at least three plants in Germany. According to FactSet, American customers make up 18% of its sales, about the same as the German market.

“I want German car companies to become American car companies,” Trump said last month while holding a rally in Savannah, Ga. “If you don’t make your product here, then you will have to pay a tariff, a very substantial tariff,” he added.

On Wednesday, Oliver Zipse, chairman of German carmaker BMW, underscored that the company has a plant in Greer, S.C.

“The most demanded vehicles in the United States, we produce there,” he told analysts Wednesday in a conference call. “So there is some natural cover against possible tariffs.”

Volkswagen and Mercedes-Benz have factories in Chattanooga, Tenn., and Vance, Ala., respectively. Manufacturers Airbus, Siemens and BASF also service the U.S. market from within, as do Nestlé and Unilever.

Much depends on details. In early 2021, Airbus’s assembly line in Mobile, Ala., was forced to pay tariffs for its shipments of fuselage, wing and tail components from France and Germany, as part of a World Trade Organization dispute. An agreement was quickly reached to suspend them.

Regardless, building up capacity to service all types of American-based demand would be hard. The Mobile plant makes A220 and A320 jets, but A330 and A350 wide-bodies are assembled in France. Volkswagen uses Chattanooga for the Atlas SUV, the Passat sedan and the electric ID.4, but the bestselling Tiguan and Jetta are built in Mexico. Roughly a quarter of U.S. imported cars originate there, and Trump has suggested that a 200% tariff could be slapped on them.

And when it comes to high-performance models, most EU firms still make them domestically and ship them over. Exports to the U.S. amounted to about 800,000 cars in 2023.

To be sure, EU leaders have struck a conciliatory tone with Trump this week, suggesting that a more amicable endgame such as the 2018 trade deal between the U.S., Canada and Mexico is possible.

Another risk is that China would send even more cheap goods to Europe if the U.S. ratchets up its trade war with Beijing. Yes, recent experience shows that China often just reroutes exports through third countries—and, as of recently, faces higher tariffs for electric vehicles in the EU anyway—but even small shifts could have big effects.

For a decade and a half, the 27-nation bloc has limped along, fostering just enough political change to avoid a painful breakup during the debt crisis of the 2010s and the 2020 pandemic, but never enough to truly invigorate its economy. Attempts by France’s Emmanuel Macron and Germany’s Olaf Scholz to change course have ended in paralysis. Scholz’s three-party government collapsed this week, after years that saw the coalition’s pro-austerity member blocking efforts to spur domestic industry with public spending.

Yet the first Trump presidency did galvanize some early support for a cohesive industrial strategy in Europe. The long-term bull case for European equities is that Trump 2.0 will be a catalyst for further transformation. European Central Bank President Mario Draghi published a report in September urging less red tape, state aid to key sectors and, where appropriate, harsher tariffs, all of which has buy-in from officials in Brussels.

On a small scale, the impulse toward a European industrial policy is already playing out. European defense contractors such as BAE Systems, Rheinmetall and Thales have seen their shares jump on the expectation that less American military involvement in Europe will force governments there to rely on their own capabilities. By 2030, the EU wants members to direct 50% or more of their procurement budgets toward European contractors.

Elsewhere, substituting foreign markets for domestic consumers will prove much harder, though providing advantages to buyers of electric vehicles has proved extremely effective in Norway. They now outnumber cars that run on gasoline.

Caught between the U.S. and China, Europe’s economic strategy is soon to face its biggest challenge since the eurozone crisis. Investors are right to be wary.

AI groups rush to redesign model testing and create new benchmarks

Rapidly advancing technology is surpassing current methods of evaluating and comparing large language models

Tech groups are rushing to redesign how they test and evaluate their artificial intelligence models, as the fast advancing technology surpasses current benchmarks.

OpenAI, Microsoft, Meta and Anthropic have all recently announced plans to build AI agents that can execute tasks for humans autonomously on their behalf. To do this effectively, the systems must be able to perform increasingly complex actions, using reasoning and planning.

Companies conduct “evaluations” of AI models by teams of staff and outside researchers. These are standardised tests, known as benchmarks, that assess models’ abilities and the performance of different groups’ systems or older versions.

However, recent advances in AI technology have meant many of the newest models have been able to get close to or above 90 per cent accuracy on existing tests, highlighting the need for new benchmarks.

“The pace of the industry is extremely fast. We are now starting to saturate our ability to measure some of these systems [and as an industry] it is becoming more and more difficult to evaluate [them],” said Ahmad Al-Dahle, generative AI lead at Meta.

To deal with this issue, several tech groups including Meta, OpenAI and Microsoft have created their own internal benchmarks and tests for intelligence. But this has raised concerns within the industry over the ability to compare the technology in the absence of public tests.

“Many of these benchmarks let us know how far away we are from automation of tasks and jobs. Without them being made public, it is hard for businesses and wider society to tell,” said Dan Hendrycks, executive director of the Center for AI Safety and an adviser to Elon Musk’s xAI.

Current public benchmarks — Hellaswag and MMLU — use multiple-choice questions to assess common sense and knowledge across various topics. However, researchers argue this method is now becoming redundant and models need more complex problems.

“We are getting to the era where a lot of the human-written tests are no longer sufficient as a good barometer for how capable the models are,” said Mark Chen, SVP of research at OpenAI. “That creates a new challenge for us as a research world.”

One public benchmark, SWE-bench Verified, was updated in August to better evaluate autonomous systems based on feedback from companies, including OpenAI.

It uses real-world software problems sourced from the developer platform GitHub and involves supplying the AI agent with a code repository and an engineering issue, asking them to fix it. The tasks require reasoning to complete.

On this measure OpenAI’s latest model, GPT-4o preview, solves 41.4 per cent of issues, while Anthropic’s Claude 3.5 Sonnet gets 49 per cent.

“It is a lot more challenging [with agentic systems] because you need to connect those systems to lots of extra tools,” said Jared Kaplan, chief science officer at Anthropic.

“You have to basically create a whole sandbox environment for them to play in. It is not as simple as just providing a prompt, seeing what the completion is and then evaluating that,” he added.

Another important factor when conducting more advanced tests is to make sure the benchmark questions are kept out of the public domain, in order to ensure the models do not effectively “cheat” by generating the answers from training data rather than solving the problem.

The ability to reason and plan is critical to unlocking the potential of AI agents that can conduct tasks over multiple steps and applications, and correct themselves.

“We are discovering new ways of measuring these systems and of course one of those is reasoning, which is an important frontier,” said Ece Kamar, VP and lab director of AI Frontiers at Microsoft research.

As a result, Microsoft is working on its own internal benchmark, incorporating problems that have not previously appeared in training to assess whether its AI models can reason as a human would.

Some, including researchers from Apple, have questioned whether current large language models are “reasoning” or purely “pattern matching” the closest similar data seen in their training.

“In the narrower domains [that] enterprises care about, they do reason,” said Ruchir Puri, chief scientist at IBM Research. “[The debate is around] this broader concept of reasoning at a human level, that would almost put it in the context of artificial general intelligence. Do they really reason, or are they parroting?”

OpenAI measures reasoning primarily through evaluations covering maths, STEM subjects and coding tasks.

“Reasoning is a very grand term. Everyone defines it differently and has their own interpretation . . . this boundary is very fuzzy [and] we try not to get too bogged down with that distinction itself, but look at whether it is driving utility, performance or capabilities,” said OpenAI’s Chen.

The need for new benchmarks has also led to efforts by external organisations.

In September, the start-up Scale AI and Hendrycks announced a project called “Humanity’s Last Exam”, which crowdsourced complex questions from experts across different disciplines that required abstract reasoning to complete.

Another example is FrontierMath, a novel benchmark released this week, created by expert mathematicians. Based on this test, the most advanced models can complete less than 2 per cent of questions.

However, without explicit agreement on measuring such capabilities, experts warn that it can be difficult for companies to assess their competitors or for businesses and consumers to understand the market.

“There is no clear way to say ‘this model is definitively better than this model’ [because] when a measure becomes a target, it ceases to be a good measure” and models are trained to pass the set benchmarks, said Meta’s Al-Dahle.

“It is something that, as a whole industry, we are working our way through.”

Billionaire activist investor’s holding company grasps for revival as it cuts dividend and weighs asset sale

Carl Icahn’s efforts to stabilise his corporate empire rest in part on a small Midwestern refinery and a scrapyard on the outskirts of Nashville, Tennessee.

The billionaire activist investor slashed the dividend of his publicly listed holding company in half on Friday, sending its stock tumbling and adding to a litany of financial troubles for the 88-year-old financier.

Icahn Enterprises, his listed conglomerate, has seen its shares plunge this year due to losses in its investment portfolio and operational problems across its businesses spanning car repair dealerships, sausage packaging operations and oil refineries, which have eaten into its cash flows.

It has been an ignominious stretch for the legendary corporate raider, who made his name in the 1980s with huge stakes in oil majors, airlines and consumer products companies.

Executives said on Friday they had fired senior managers of its Pep Boys car repair unit, after it ran out of tyres last quarter — due to a cost-cutting effort to buy cheap tyres from a new vendor. The operational crisis forced Icahn Enterprises to write down the value of Pep Boys by $193mn.

Icahn’s representatives did not respond to interview requests on Friday.

Last month, its largest investment, a controlling interest in refinery CVR Energy, was pummelled after the group eliminated its dividend to conserve cash amid power outages and heavy spending needs. That cut off a vital source of funds for Icahn Enterprises.

On Friday, Icahn Enterprises followed suit, halving its quarterly payout from $1 per share to 50 cents, in a move designed to save cash and double down on its refining operations.

But it scared shareholders, many of whom were attracted to the conglomerate for its high yields. Icahn Enterprises’ stock fell 6 per cent, putting year-to-date losses at about 38 per cent.

In addition to cash flow shortfalls that drove the dividend cut, Icahn Enterprises has seen the value of its assets quickly erode. Its net asset value per share has declined 35 per cent this year.

Those losses are putting financial strain on Icahn, who borrowed billions in personal loans against his shares in Icahn Enterprises, using the stock as collateral for as much as $5bn from a group of banks in recent years.

Icahn’s heavy borrowing came during a decade-long stretch of losses in his investment portfolio due to large bets he held against booming financial markets. The feared activist has pledged substantial assets as collateral against the loans and now must repay the debts in coming years after some of their values have fallen.

Now Icahn is betting that he can pull off a revival by selling a metals scrapyard he acquired more than a decade ago.

In 2007, Icahn acquired a scrap metals company called PSC Metals with a large junkyard in Nashville. When Icahn sold the company in 2021, he held on to PSC’s 45-acre scrapyard in the city’s industrial “East Bank” neighbourhood, betting he could eventually flip the property when the area was gentrified.

On Friday, executives at Icahn Enterprises told shareholders they had put the property up for sale and hoped to make a 10-fold gain on its $25mn carrying value.

The National Football League’s Tennessee Titans are building a stadium adjacent to the scrapyard and the area is part of a giant urban renewal plan in one of America’s fastest growing cities. It has turned the industrial wasteland into a coveted property, according to local real estate executives.

Icahn also told shareholders on Friday he decided to cut his conglomerate’s dividend because he wants to buy additional shares in CVR Energy. He has launched a tender offer to buy $263mn in CVR stock, boosting Icahn Enterprises’ ownership from 66 per cent to 81 per cent.

Icahn Enterprises characterised the manoeuvre as opportunistic after an about 20 per cent tumble in CVR’s shares last month, giving it greater ownership in a refining operation that has paid out over $3bn in dividends to the conglomerate since 2012.

“[We] think the decisions we made today will actually help cash flow in the future,” said Andrew Teno, chief executive, referring to the tender offer.

The trade could benefit from a new administration in the US. CVR carries $374mn of liabilities due to a federal mandate to mix biofuels into petrol supplies or buy credits to make up the shortfall.

But the refinery has sued the federal government over the mandate and recently blasted the Biden administration's “egregious conduct” in keeping refineries that had won some legal challenges to the payments hanging. “[The] financial impact of their actions threaten the very existence of small refineries like ours,” said CVR chief executive David Lamp.

But the new administration could offer the relief CVR is seeking, removing a big financial burden.

“Rarely have I seen a stock market with such extreme valuations — with some companies trading at unjustifiable premiums and others being massively undervalued,” Icahn said in a prepared statement on Friday.