TheChek News Flow

How AI Dealmakers Are Sidestepping Regulators

Attention: AI startups. What’s the best way to sell your company without the prying eyes of antitrust regulators? Make yourself so small you won’t attract attention!

We’re serious, kind of. We’re hearing that a lesser-known loophole to the rules determining when companies have to notify antitrust regulators about an acquisition is based on the size of a company’s net assets or sales. To sum it up, if it’s below $23.9 million, you might be OK.

Specifically, I’m referring to the rules based on the Hart-Scott-Rodino Act, which governs antitrust reporting.

Right now, antitrust authorities don’t need to be notified if the value of an acquisition is below $119.5 million. If the deal’s value is higher—between $119.5 million and $478 million—the parties involved only have to notify authorities if one company (likely the acquirer) has at least $239 million in annual net sales or total assets and the other company (likely the target) has at least $23.9 million in annual net sales or total assets. (Very specific figures, we know!)

So to avoid notifying regulators ahead of time and putting their deal at risk, startups being acquired can use a number of tricks to get their total assets below that $23.9 million figure, a banker who’s been involved in several of these deals told me.

Some tactics this banker has seen include paying outstanding cloud computing bills that a startup might otherwise contest or paying the bankers early, before the deal closes, they said. Startups might also pay employees early or spend cash to buy cloud computing credits, they added.

In doing so, these startups can fly under the radar of antitrust regulators, and don’t have to go through the complicated licensing-and-hiring agreements that startups like Inflection, Adept and Character.AI have had to strike with Microsoft, Amazon and Google, respectively, to avoid antitrust scrutiny.

These strategies have been used to avoid the HSR thresholds in the past, but they’re becoming more popular now with AI companies looking to sell themselves, since the main acquirers that can afford high AI startup prices are big tech companies like Microsoft, Google and Amazon that regulators already have their eye on, the banker said.

There have been 54 AI M&A deals this year so far, worth a combined $2.8 billion, according to PitchBook. (This excludes Inflection-esque deals.) However, this year’s AI quasi-acquisitions likely didn’t have to worry about pre-notifying regulators, since they were technically licensing and hiring agreements rather than traditional mergers, a banker told me. (It seems like that hasn’t stopped the Federal Trade Commission from investigating some of these deals, though.)

Therefore, these tactics would have been more relevant for smaller deals this year, like the acquisitions of Rockset, Deci AI, OctoAI and Robust Intelligence, that could have fallen around these thresholds. (None of the companies had anything to add when we reached out to them about this.)

Don’t expect to see too many acquisitions for the next few weeks, though. Bankers, companies and VCs alike are looking to the results of next month’s presidential elections to see how the antitrust environment and M&A loopholes might change. But I’m sure there will be plenty more dealmaking to come in the months ahead. (Check out our generative AI takeover list for a sneak peek at who might be putting “for sale” signs up.)

OpenAI hires its first chief economist

OpenAI has hired its first chief economist: Aaron Chatterji, formerly the chief economist at the Commerce Department under President Joe Biden and a senior economist in President Barack Obama’s Council of Economic Advisers.

Chatterji, who’s also a professor of business and public policy at Duke, will study AI’s economic impacts at OpenAI, leading research into how AI might influence economic growth and job prospects.

Notably, Chatterji helped to coordinate the Biden administration’s implementation of the 2022 CHIPS Act, which authorized roughly $280 billion in funding for the development of computer chips in the U.S. His understanding of the project — and political connections — could help OpenAI as it undertakes chip design efforts of its own.

Chatterji wasn’t OpenAI’s only high-profile hire today. The company also appointed Scott Schools, former associate deputy attorney general and Uber’s compliance head, as its new chief compliance officer. He’ll work with OpenAI’s board and teams across the company on legal requirements and ethics, OpenAI said.

WhatsApp introduces contact storing directly within the app, teases usernames

WhatsApp on Tuesday announced a new feature that will allow users to save contacts within the app. That means that even if you lose your phone or link a new device to your primary number, you will see all your contacts stored within WhatsApp’s cloud storage.

Until now, WhatsApp relied on your phone’s contact book to sync contacts. If a new person messaged you on the app, and you wanted to save their details, the information would be stored locally on your device.

The new feature, rolling out soon, lets you store contacts on any device, including WhatsApp for the web and Windows. You can choose to sync contacts saved within WhatsApp to your phone, as well.

Last year, the company began allowing users to log into two WhatsApp accounts on one device. Saving contacts linked to one WhatsApp account is handy if you want to keep personal and business contacts separate. Plus, if you share the device with others, you can maintain your own contact list linked to your number.

WhatsApp said it developed a new encrypted storage system called Identity Proof Linked Storage (IPLS) to enable secure contact saving. When users save a contact, the system generates an encryption key on the device. The retrieval is based on the client authenticating its primary device identity.

Additionally, the app is also partnering with Cloudflare to sign any changes to the cryptographic properties of a user’s directory. According to Meta, this step ensures that someone has not edited contacts saved in WhatsApp.

The company said that the tech behind the contact storing feature would make it possible to save contacts by usernames.

“Usernames on WhatsApp will add an extra degree of privacy so that you don’t need to share your phone number when messaging someone,” the company noted in a blog post.

The feature has been in development for some time. In May 2023, a report by WABetaInfo highlighted that WhatsApp has been working on introducing usernames to the app. WhatsApp rivals, Signal and Telegram, both offer a way for users to share their username with someone without sharing their contact details.

Expulsion of chief marketing officer is latest blow to German carmaker in world’s largest auto market

A senior Volkswagen executive has been deported from China, in another setback for the German group fighting for survival in the world’s biggest car market.

VW’s chief marketing officer and head of product strategy for China, Jochen Sengpiehl, is back in Germany after being detained in China for about 10 days, according to two people with knowledge of the details.

He allegedly tested positive for drugs following a holiday abroad, they said. VW is reviewing the matter internally, one of them said.

Germany’s Federal Foreign Office on Tuesday confirmed that its embassy in Beijing had “provided consular assistance and was in regular contact with the concerned individual and his family as well as with the local authorities”.

The incident is another blow to VW in its most profitable market. Europe’s largest automaker has in recent years been losing grip of its decades-old market dominance in China and has announced investments of more than €5bn there to catch up to Chinese competitors, such as BYD.

VW, which is trying to push through historic job cuts at home, is also locked in sensitive negotiations with its Chinese joint venture partner SAIC over the future of a controversial plant in Xinjiang, the western province where Beijing has been accused of human rights abuses.

The VW executive, who spent just over two years in China, had tested positive for cannabis and cocaine — both illegal in the country — following his return from a holiday in Thailand about two weeks ago, one of the people added.

German newspaper Bild first reported the deportation. It was not yet clear why the executive was tested. Thailand decriminalised cannabis use in 2022, while Germany legalised it earlier this year. Cocaine remains illegal in both countries.

VW said it could not comment due to “contractual and data protection confidentiality obligations”.

Sengpiehl was emailed for comment but had not responded by time of publication. According to his LinkedIn profile, he had been in his China role since 2022. Prior to that, he was VW’s global marketing officer for some five years, following earlier roles at Hyundai, WPP, Daimler, BBDO and Nissan.

Following a series of raids on western consultancies and the introduction of tougher rules on espionage and state secrets, foreign investors and business people have become increasingly uneasy over their personal safety in China.

China has a criminal conviction rate of around 99 per cent and the country’s courts have been required to hand down tough sentences for drug convictions in recent years.

In a pitch to investors last spring, Anthropic said it intended to build AI to power virtual assistants that could perform research, answer emails, and handle other back-office jobs on their own. The company referred to this as a “next-gen algorithm for AI self-teaching” — one it believed that could, if all goes according to plan, automate large portions of the economy someday.

It took a while, but that AI is starting to arrive.

Anthropic on Tuesday released an upgraded version of its Claude 3.5 Sonnet model that can understand and interact with any desktop app. Via a new “Computer Use” API, now in open beta, the model can imitate keystrokes, button clicks, and mouse gestures, essentially emulating a person sitting at a PC.

“We trained Claude to see what’s happening on a screen and then use the software tools available to carry out tasks,” Anthropic wrote in a blog post shared with TechCrunch. “When a developer tasks Claude with using a piece of computer software and gives it the necessary access, Claude looks at screenshots of what’s visible to the user, then counts how many pixels vertically or horizontally it needs to move a cursor in order to click in the correct place.”

Developers can try out Computer Use via Anthropic’s API, Amazon Bedrock, and Google Cloud’s Vertex AI platform. The new 3.5 Sonnet without Computer Use is rolling out to Claude apps, and brings various performance improvements over the outgoing 3.5 Sonnet model.

Automating apps

A tool that can automate tasks on a PC is hardly a novel idea. Countless companies offer such tools, from decades-old RPA vendors to newer upstarts like Relay, Induced AI, and Automat.

In the race to develop so-called “AI agents,” the field has only become more crowded. AI agents remains an ill-defined term, but it generally refers to AI that can automate software.

Some analysts say AI agents could provide companies with an easier path to monetizing the billions of dollars that they’re pouring into AI. Companies seem to agree: According to a recent Capgemini survey, 10% of organizations already use AI agents and 82% will integrate them within the next three years.

Salesforce made splashy announcements about its AI agent tech this summer, while Microsoft touted new tools for building AI agents yesterday. OpenAI, which is plotting its own brand of AI agents, sees the tech as a step toward super-intelligent AI.

Anthropic calls its take on the AI agent concept an “action-execution layer” that lets the new 3.5 Sonnet perform desktop-level commands. Thanks to its ability to browse the web (not a first for AI models, but a first for Anthropic), 3.5 Sonnet can use any website and any application.

“Humans remain in control by providing specific prompts that direct Claude’s actions, like ‘use data from my computer and online to fill out this form’,” an Anthropic spokesperson told TechCrunch. “People enable access and limit access as needed. Claude breaks down the user’s prompts into computer commands (e.g. moving the cursor, clicking, typing) to accomplish that specific task.”

Software development platform Replit has used an early version of the new 3.5 Sonnet model to create an “autonomous verifier” that can evaluate apps while they’re being built. Canva, meanwhile, says that it’s exploring ways in which the new model might be able to support the designing and editing process.

But how is this any different than the other AI agents out there? It’s a reasonable question. Consumer gadget startup Rabbit is building a web agent that can do things like buying movie tickets online; Adept, which was recently acqui-hired by Amazon, trains models to browse websites and navigate software; and Twin Labs is using off-the-shelf models, including OpenAI’s GPT-4o, to automate desktop processes.

Anthropic claims the new 3.5 Sonnet is simply a stronger, more robust model that can do better on coding tasks than even OpenAI’s flagship o1, per the SWE-bench Verified benchmark. Despite not being explicitly trained to do so, the upgraded 3.5 Sonnet self-corrects and retries tasks when it encounters obstacles, and can work toward objectives that require dozens or hundreds of steps.

But don’t fire your secretary just yet.

In an evaluation designed to test an AI agent’s ability to help with airline booking tasks, like modifying a flight reservation, the new 3.5 Sonnet managed to complete less than half of the tasks successfully. In a separate test involving tasks like initiating a return, 3.5 Sonnet failed roughly a third of the time.

Anthropic admits the upgraded 3.5 Sonnet struggles with basic actions like scrolling and zooming, and that it can miss “short-lived” actions and notifications because of the way it takes screenshots and pieces them together.

“Claude’s Computer Use remains slow and often error-prone,” Anthropic writes in its post. “We encourage developers to begin exploration with low-risk tasks.”

Risky business

But is the new 3.5 Sonnet capable enough to be dangerous? Possibly.

A recent study found that models without the ability to use desktop apps, like OpenAI’s GPT-4o, were willing to engage in harmful “multi-step agent behavior,” such as ordering a fake passport from someone on the dark web, when “attacked” using jailbreaking techniques. Jailbreaks led to high rates of success in performing harmful tasks even for models protected by filters and safeguards, according to the researchers.

One can imagine how a model with desktop access could wreak more havoc — say, by exploiting app vulnerabilities to compromise personal info (or storing chats in plaintext). Aside from the software levers at its disposal, the model’s online and app connections could open avenues for malicious jailbreakers.

Anthropic doesn’t deny that there’s risk in releasing the new 3.5 Sonnet. But the company argues that the benefits of observing how the model is used in the wild ultimately outweigh this risk.

“We think it’s far better to give access to computers to today’s more limited, relatively safer models,” the company wrote. “This means we can begin to observe and learn from any potential issues that arise at this lower level, building up computer use and safety mitigations gradually and simultaneously.”

Anthropic also says it has taken steps to deter misuse, like not training the new 3.5 Sonnet on users’ screenshots and prompts, and preventing the model from accessing the web during training. The company says it developed classifiers to “nudge” 3.5 Sonnet away from actions perceived as high-risk, such as posting on social media, creating accounts, and interacting with government websites.

As the U.S. general election nears, Anthropic says it is focused on mitigating election-related abuse of its models. The U.S. AI Safety Institute and U.K. Safety Institute, two separate but allied government agencies dedicated to evaluating AI model risk, tested the new 3.5 Sonnet prior to its deployment.

Anthropic told TechCrunch it has the ability to restrict access to additional websites and features “if necessary,” to protect against spam, fraud, and misinformation, for example. As a safety precaution, the company retains any screenshots captured by Computer Use for at least 30 days — a retention period that might alarm some devs.

We asked Anthropic under which circumstances, if any, it would hand over screenshots to a third party (e.g. law enforcement) if asked. A spokesperson said that the company would “comply with requests for data in response to valid legal process.”

“There are no foolproof methods, and we will continuously evaluate and iterate on our safety measures to balance Claude’s capabilities with responsible use,” Anthropic said. “Those using the computer-use version of Claude should take the relevant precautions to minimize these kinds of risks, including isolating Claude from particularly sensitive data on their computer.”

Hopefully, that’ll be enough to prevent the worst from occurring.

A cheaper model

Today’s headliner might’ve been the upgraded 3.5 Sonnet model, but Anthropic also said an updated version of Haiku, the cheapest, most efficient model in its Claude series, is on the way.

Claude 3.5 Haiku, due in the coming weeks, will match the performance of Claude 3 Opus, once Anthropic’s state-of-the-art model, on certain benchmarks at the same cost and “approximate speed” of Claude 3 Haiku.

“With low latency, improved instruction following, and more accurate tool use, Claude 3.5 Haiku is well suited for user-facing products, specialized sub-agent tasks, and generating personalized experiences from huge volumes of data–like purchase history, pricing, or inventory data,” Anthropic wrote in a blog post.

3.5 Haiku will initially be available as a text-only model and later as part of a multimodal package that can analyze both text and images.

So once 3.5 Haiku is available, will there be much reason to use 3 Opus? What about 3.5 Opus, 3 Opus’ successor, which Anthropic teased back in June?

“All of the models in the Claude 3 model family have their individual uses for customers,” the Anthropic spokesperson said. “Claude 3.5 Opus is on our roadmap and we’ll be sure to share more as soon as we can.”

Private equity management fees fall to lowest level since records began

Fund managers are fighting to secure investors’ money in a tough environment

Management fees on private equity buyout funds have fallen to their lowest levels since records began in 2005 as fund managers fight to attract investors in a tough fundraising environment.

According to industry specialist Preqin, the average management fee for buyout funds that closed this year or were still raising money in June was 1.74 per cent of investors’ committed capital. The previous low was 1.85 per cent in 2023.

In the past two years, private equity firms have struggled to sell out of their investments. The usual exit routes of stock market IPOs and industry dealmaking have been limited by higher interest rates, disagreements about valuations and general economic uncertainty.

Firms have returned less money to their investors as a result, in turn leaving those investors with less cash to reinvest in new buyout funds.

“Because of that pressure on fundraising, that’s why [buyout managers] are going to make concessions on fees and terms,” said Greg Durst, a senior managing director at the Institutional Limited Partners Association, which represents the industry’s investors.

“They’re being very slow and judicious about how they’re going to be making new commitments.”

He added that over the past couple of years, the cash returned by private equity managers to their investors, known as limited partners, has been “way, way off what LPs had grown accustomed to planning around” and that had been “a challenge”.

In addition to the difficult fundraising environment, another factor affecting fees is the size of the fund manager.

According to Preqin, as fund sizes have increased over the past 20 years some of the bigger managers, who receive a larger volume of fees, have chosen to cut their rates. Some smaller firms have cut their fees to try to compete.

“Many investors have concentrated on relationships with the largest fund managers,” said one London lawyer who advises mid-market private capital funds. “This means that smaller managers at the lower end of the market are having to work harder.”

A lot of the larger firms manage funds across multiple strategies, such as private credit and buyout, and “will offer a fee break across all of them”, said Durst. “If you’re in one, you’re in for a 2 per cent management fee. If you do three, you’re in for 1.75 [per cent].”

Bigger limited partners also have more leeway to negotiate on fees, according to one London family office manager.

The family office manager, who allocates relatively small amounts of client money to “some of the biggest and most prestigious [private equity firms] on the planet”, said they were not paying less.

“I suspect it’s only one category [of investor] paying lower fees and that it’s the big boys who write $25mn to $100mn cheques.”

Despite the falls in management fees, buyout fund performance fees — or the share of profits that fund managers get to keep on their successful investments, also known as carried interest — have barely changed, Preqin found.

Over the past 20 years, that figure has hovered around an average of 19.5 per cent of fund profits, after a minimum return for limited partners is met.

The Preqin data also shows that there has been no notable downward pressure in management fees for private debt funds.

Private debt management and performance fees can be lower than for other private asset classes depending on the risk and potential returns, the data provider said.

There has been an upsurge in investor interest in private debt investment.

The Preqin data refers to the management fees for the period in which fund managers are actively investing the capital, usually the first 3 to 5 years of a fund.

Like-for-like third quarter revenue at world’s biggest beauty company misses expectations

L’Oréal reported disappointing sales in the third quarter as demand for skincare and make-up waned in the French beauty group’s crucial Chinese market.

Sales at the group, whose brands include Garnier, La Roche-Posay and Kiehl’s, grew 3.4 per cent on a like-for-like basis in the three months to September 30, reaching €10.28bn, significantly below analyst estimates of a 6 per cent rise.

The third-quarter figures marked a further slowdown from the second quarter, in which sales grew 5.3 per cent on a like-for-like basis, slowing from 9.4 per cent growth in the first quarter.

Like-for-like sales in north Asia, which mainly constitutes China, plunged 6.5 per cent — well below the consensus estimate of a 2.9 per cent rise.

The French beauty giant’s earnings are seen as a bellwether for the health of the global beauty industry, which has fallen victim to weak demand, particularly in China, where consumer confidence has cratered alongside the country’s housing slump. 

“The situation in the Chinese ecosystem has become even more challenging,” said chief executive Nicolas Hieronimus. “But we believe in the future of this market and hope that the governmental stimulus will help improve consumer confidence.”

Sales in China have also been hit by a government crackdown on daigou, shoppers who buy cosmetics in lower-tax areas in order to sell them for a profit in mainland China.

Last week luxury behemoth LVMH reported weaker than expected sales on the back of the Chinese slump, prompting shares in the group and its peers Cartier, Hermès and Kering to tumble.

The L’Oréal growth shortfall was also driven by a heavy sales miss in its higher-end dermatological beauty division, which includes brands such as SkinCeuticals and CeraVe. Sales grew 0.8 per cent in the third quarter, compared to an expected 10.8 per cent rise.

The world’s biggest beauty company by sales, L’Oréal has for the past three years enjoyed healthy profits as a result of the “lipstick effect”, in which consumers opt to buy small-ticket luxury items even as they cut back on everyday goods as the cost of living has soared since the Covid-19 pandemic.

However, the gloss has started to come off the high-margin “prestige” beauty category, which has been outpacing sales growth of mass-market beauty, in a sign that even higher-income consumer spending is faltering. 

“The last time L’Oréal reported quarterly organic sales growth lower than this was Q3 2020 in the darkest days of Covid,” wrote RBC Capital analyst James Edwardes Jones, although he added the group was still outperforming the rest of the beauty market.

In North America, like-for-like sales rose 5.2 per cent, better than an anticipated 3.7 per cent. European sales were slightly below expectations, growing 5.6 per cent.

At the end of June, Hieronimus tempered expectations for growth in the global beauty market, saying he expected 4.5 to 5 per cent compared with the 5 per cent previously forecast.

Shares in the beauty group, which reported after the close of the French market, have fallen about 17 per cent so far this year. Bernstein’s Callum Elliott said: “For such a well-loved name, the concurrent slowdown across the main growth drivers of the business risks meaningfully intensifying fears around the long-term sustainable growth rate, and we expect the stock to react accordingly.”

L’ex-PDG et principal actionnaire du distributeur demande la désignation d’un juge d’instruction pour enquêter sur ceux qu’il estime responsables de sa chute.

Jean-Charles Naouri a perdu le contrôle de Casino, mais pas l’envie de se battre contre ceux qu’il estime responsables de sa chute, à commencer par Carson Block. Avec sa société Muddy Waters, ce financier américain avait spéculé, à partir de décembre 2015, sur l’effondrement du cours de Bourse du distributeur stéphanois. Et, à l’été 2018, il avait lancé une nouvelle offensive de déstabilisation.

PDG et principal actionnaire du groupe (Casino, Monoprix, Franprix, Vival, Cdiscount…) pendant plus de trois décennies, l’ex-directeur de cabinet de Pierre Bérégovoy au ministère des Finances, qui fut associé gérant à la banque Rothschild avant de se lancer dans les affaires à la fin des années 1980, a dû lâcher les rênes au printemps dernier. Le groupe Casino, ou plutôt ce qu’il en reste après la vente contrainte de ses filiales en Amérique du Sud et de la quasi-totalité de ses hypers et supermarchés français, est passé fin mars 2024 sous le contrôle d’un consortium dirigé par le milliardaire tchèque Daniel Kretinsky.

Quelques mois à peine après sa mise à la retraite forcée, l’ex-PDG, qui a longtemps contrôlé plus de 51 % de capital de Casino via une cascade de holdings, est reparti à l’attaque sur le front judiciaire. Le 9 juillet dernier, Jean-Charles Naouri, désormais simple actionnaire ultraminoritaire de Casino, a déposé plainte avec constitution de partie civile auprès du doyen des juges d’instruction du tribunal judiciaire de Paris, Serge Tournaire, contre Carson Block, Muddy Waters « et tous autres ».

Spéculations sur la chute du groupe

La liste des méfaits présumés est impressionnante : « Manipulation de cours, diffusion d’informations fausses et trompeuses, délit d’initié, corruption active, association de malfaiteurs et recel de ces infractions », indique la plainte de 51 pages, que Le Figaro a pu consulter. Les avocats de Jean-Charles Naouri s’attendent à ce que Serge Tournaire désigne prochainement un juge d’instruction.

« La prophétie de Carson Block, telle que formulée depuis 2015, et les agissements entrepris dans ce cas de façon continue depuis lors auront conduit au dépeçage annoncé du groupe Casino, par une entente préalablement établie à cet effet, engendrant des conséquences économiques et sociales considérables », résume la plainte, avant de marteler le fond de la pensée de Jean-Charles Naouri : « Contrairement aux critiques portées par Carson Block, cette issue n’a aucunement procédé de prétendues carences du groupe. Malgré une dette en baisse (passant de 12 milliards en 2014 à 6 milliards en 2024), les frais financiers explosaient, passant de 100 à 500 millions par an, ce qui ne s’expliquait que par les attaques. »

En clair, ce n’est pas l’immense dette de Casino qui a provoqué sa chute, mais l’utilisation qu’en ont faite les spéculateurs : en pointant du doigt un risque de faillite, et en spéculant sur une chute du cours, Muddy Waters a entraîné une dégradation de la note de la dette par les agences, et donc une flambée de son coût, qui a fini par avoir raison du groupe…

« Avant l’attaque, il était reconnu que les actifs de Casino étaient les meilleurs du marché français, ajoute la plainte. Les prix pratiqués étaient au niveau de ceux affichés par l’enseigne Leclerc. Aucune critique sur le management n’était formulée et Jean-Charles Naouri était élu à plusieurs reprises le meilleur dirigeant de la distribution en Europe. » Avec autant d’atouts en main, comment le groupe Casino a-t-il donc bien pu être acculé, quelques années plus tard, à une quasi-faillite ? « D’évidence, la spirale infernale initiée en 2015 conduisait à pratiquer des hausses de prix et engendrait une désaffection des clients, malgré une dette en forte baisse, expliquent les avocats du PDG de l’époque. Cette mécanique, marquée par des attaques successives en meute, évidemment coordonnées, a conduit à la mort du groupe au terme de huit années, avec un impact social considérable. »

En juillet 2018, Jean-Charles Naouri avait déposé plainte auprès du procureur de la République, déjà pour manipulation de cours, diffusion d’informations fausses, délit d’initié et association de malfaiteurs. « Depuis lors, aucune avancée judiciaire n’était constatée en dépit d’une enquête préliminaire de six années, toujours en cours », relèvent ses avocats. C’est ce qui a incité l’ex-PDG de Casino à former une nouvelle plainte, avec constitution de partie civile cette fois…

En fait, une autre raison a poussé le dirigeant déchu à cette nouvelle offensive : il est lui-même sous la menace d’un procès pour corruption et manipulation de cours pour s’être attaché, entre septembre 2018 et mars 2019, les services de Nicolas Miguet. Cet homme d’affaires, à la tête d’organes d’informations boursières, aurait fait à ses abonnés des recommandations d’achat du titre du distributeur stéphanois.

« Casino a été contraint de solliciter l’aide de différents professionnels dans le cadre de la mise en œuvre d’une stratégie globale de protection », expliquent les avocats de Jean-Charles Naouri pour tenter de convaincre le Parquet national financier (PNF) de joindre les deux affaires. Las. Les avocats de Jean-Charles Naouri s’attendent à ce que le PNF, qui avait été saisi en 2020 par le gendarme de la Bourse, renvoie l’affaire Miguet devant le tribunal judiciaire.

L’ex-PDG convaincu qu’un complot a été ourdi par des rivaux

Si Jean-Charles Naouri repart à l’offensive avec une nouvelle plainte, c’est aussi qu’il est intimement convaincu que c’est un complot ourdi par des rivaux qui a conduit à la chute de Casino. Selon ses avocats, c’est l’ancien associé brésilien de Casino, Abilio Diniz, qui serait à l’origine de cette « opération de déstabilisation dénommée Napoléon » visant « à racheter la société Casino à vil prix et à conduire ainsi à son dépeçage ». Abilio Diniz, fils du fondateur du distributeur brésilien Grupo Pao de Açucar (GPA), et Jean-Charles Naouri avaient été au centre d’un violent conflit entre 2011 et 2013.

Pour empêcher la prise de contrôle de GPA par Casino, prévue par des accords signés entre les deux groupes en 2005, Abilio Diniz s’était allié à Carrefour. Face à la combativité de Jean-Charles Naouri, le Brésilien n’a jamais pu arriver à ses fins. Il avait fallu des tractations menées par William Ury, un ancien négociateur entre le gouvernement colombien et les Farc, pour que les deux hommes soldent leurs litiges, en septembre 2013.

« En février 2015, William Ury présentait Abilio Diniz à Carson Block, assure la plainte. Des informations confidentielles relatives à Casino étaient manifestement échangées entre messieurs William Ury, Carson Block et Abilio Diniz, l’année même au cours de laquelle ce dernier devenait grand actionnaire de Carrefour. » Dans sa plainte, Jean-Charles Naouri incrimine également les ex-patrons d’une banque anglaise et d’un cabinet d’analyse financière, qui ont écrit des notes défavorables à Casino, ainsi que le patron d’un distributeur et un banquier d’affaires. Persuadé de leur collusion avec Carson Block, il préconise même au futur juge d’instruction une liste de questions à poser à chacun d’entre eux. Abilio Diniz, lui, est décédé le 18 février dernier, quelques semaines seulement avant que Jean-Charles Naouri ne perdre définitivement le contrôle de Casino.