TheChek News Flow

Hollywood studio wants Larry Ellison to offer an airtight personal guarantee, like the Tesla boss did with Twitter

Faced with a $108bn bid backed by Oracle chair Larry Ellison, the directors of Warner Bros Discovery seem to have found themselves thinking of Elon Musk. The Tesla boss signed a merger agreement for social network Twitter in 2022, and then tried to walk away from it. Ellison is not Musk, but he could do a little more to prove it.

WBD on Tuesday shared its reasons for declining a $30-per-share offer from Paramount Skydance, the media conglomerate controlled by Ellison’s family. WBD is instead sticking by its endorsement of a deal with Netflix, which offers $27.75 in the form of cash and Netflix stock, as well as equity in a spin-off cable networks business worth perhaps a few bucks.

Its reasoning rests partly on the structure of the funding. Of the $41bn equity contribution, only $12bn comes from Ellison, and the rest from various Middle Eastern wealth funds. Ellison has offered to “backstop” the whole equity portion, but WBD protests at the entity he is using, a “revocable trust”. It wants Ellison to offer an airtight personal guarantee, as Musk did, not one issued by an “estate-planning vehicle with no transparency”.

Paramount argues that the Ellison trust, which dates back to 1988 and holds his 40 per cent stake in Oracle, is simply the billionaire’s preferred way to handle his affairs. He has, after all, used it countless times over the past three decades to fund deals including his own investment in Musk’s Twitter bid, although that was much smaller than this current deal.

Still, while that bolsters the argument that Ellison won’t walk away from his promises, it does not guarantee that he cannot do so, to WBD’s satisfaction. In theory, he has the power to move assets in and out of the trust. WBD notes that, if a court cannot later force the Paramount buyout to close, the trust is only on the hook for monetary damages of less than $3bn.

Musk, of course, did try to walk away from Twitter, claiming the company had misrepresented its spam problems. But he ultimately closed the deal as it became clearer that he had little legal leg to stand on. WBD pointed out in Tuesday’s filing that Paramount’s legal adviser is Quinn Emanuel, the same firm Musk used to attempt his foiled escape from Twitter.

Faced with the option of $30 in cash, versus Netflix’s more complicated and seemingly less valuable alternative, WBD investors may prove to be more tolerant of the potential wrinkles in Paramount’s financing than the board. The comparison is further muddied by the other risks to closing, which include regulatory reviews in multiple jurisdictions.

Even so, it would be fairly easy for Ellison to increase his odds. As the owner of $206bn of Oracle shares, he can afford to tighten up the Paramount offer with a more direct backstop or even direct contribution, and raise the offer by a couple of dollars per share to make clear it’s financially superior too. Sure, he might not want to; in a more straightforward world he might not have to. But success in Hollywood is all about knowing how to engage an audience.

SpaceX, Anthropic and OpenAI might try to spread out selling by current shareholders to avoid a flood of stock hitting the market.

Wall Street is gearing up for three potentially record-setting initial public offerings over the next 18 months. The talk among bankers is not about what SpaceX, Anthropic and OpenAI will do on their big days but about what happens after that, when the companies’ existing shareholders try to unload their multibillion-dollar stakes.

Each of the three has raised some of the largest amounts ever for private companies. That creates the risk that when they go public, early investors will flood the market with so much stock that it will depress prices. That’s always a danger for companies going public, which is why early investors are typically prevented from selling for three to six months. But the sheer amount of stock held by early investors in SpaceX, Anthropic and OpenAI heightens the risk for the tech giants.

Investment bankers are starting to prepare for that reality, discussing proposed solutions to help win deals, even before some of the official IPO work has begun. At least two large banks largely ruled out a standard IPO lockup period of either 90 to 180 days and are discussing how to design a staggered lockup release for companies like OpenAI and Anthropic, bankers said.

Bankers have started floating that idea with existing investors in some of the companies, who would need to sign off on the plans, said an Anthropic investor.

“The bankers, given the size of the IPOs, have to rethink how traditional lockups are done,” said Jon Redmond, a former Morgan Stanley banker who is now portfolio manager at the hedge fund Discovery Capital.

Bankers and people close to the company expect SpaceX to lead the IPO parade, with a listing taking place next summer. Bankers vying to lead the IPO are scheduled to meet with the company this week, according to a person briefed on the plans. A SpaceX spokesperson didn’t return a request for comment.

The preparations for OpenAI and Anthropic are less advanced and their timing plans appear more fluid. But investors and advisers expect the companies to list either late next year or sometime the following year. An OpenAI spokesperson declined to comment. An Anthropic spokesperson pointed to comments made by chief communications officer Sasha de Marigny this month that the company was “keeping our options open” but had “no immediate plans to go public.” OpenAI declined to comment.

Upgrade to ask Deep Research to…

How does Broadcom factor into the AI chip race with Nvidia, AMD and others?

View Response

Anthropic tentatively expects to list publicly toward the end of next year or in early 2027 and has been talking to banks informally about what an IPO would look like, an investor in the company said. OpenAI has been discussing whether to list in 2026 or 2027, an adviser said.

Anthropic and OpenAI have each raised tens of billions of dollars privately, far more than any other tech firm, while SpaceX has raised about $10 billion, according to PitchBook. OpenAI and Anthropic need to raise significantly more capital to fulfill expensive plans to build data centers, while SpaceX is trying to build larger rockets. Investment bankers thought each could raise between $25 billion and $50 billion in their IPOs, making them the largest tech offerings of all time.

Companies going public generally allow some large shareholders, including current and former executives, to sell a portion of their shares in the offering. But the vast majority of shares held by early investors can’t be sold for six months. Some companies in recent years, like Airbnb and Instacart, have made exceptions to those rules by allowing shareholders to sell earlier if the stock price hits certain milestones.

One model could be the 2012 IPO of Facebook, among the largest tech listings in history, which raised about $16 billion. The social network put in place a staggered lockup release to prevent shares from flooding the market. A more recent case could be Figma, which after its July listing got shareholders who own a majority of the company to agree to hold on to shares for more than a year.

With the coming crop of IPOs, investment bankers have internally discussed how to structure the lockup. One option is staggering the dates to prevent a wave of selling on one day. IPO bankers and lawyers said investors could be allowed to sell a portion of their holdings every 20 to 30 days. Releases can also be triggered when the stock hits a certain price, they said.

Another tricky issue with these giant offerings is how much to allocate to small investors, who have been unable to get a piece of these companies and are expected to gobble up the shares. IPO advisers don’t want to allocate too little and watch the price pop, only to tumble when new shares hit the market.

Then there’s the element of surprise. Redmond, the former banker, said banks should consider getting securities regulators’ blessing to allow companies to provide less disclosure about when lockups will end. “I think people are thinking about it,” he said. “I haven’t seen anything done in practice, but we haven’t seen IPOs like this in our lifetime.”

The banks that solve these problems could reap billions of dollars in fees. Investment banks typically take a single-digit percentage cut of IPO proceeds, which gets magnified in such large deals. “Fifty billion dollars times 3% is $1.5 billion in fees,” said one investment banker whose firm will try to be in on the deals.

Morgan Stanley and Goldman Sachs dominate the tech IPO business, with one or the other leading the vast majority of such listings. Morgan Stanley has been particularly close to SpaceX CEO Elon Musk lately. He hired Morgan Stanley for his bid to take Twitter private and for the merger between X and xAI. Musk also recently hired former Morgan Stanley banker Anthony Armstrong as chief financial officer of xAI, his AI firm. Goldman Sachs led Tesla’s IPO in 2010.

Goldman Sachs has recently been close to OpenAI, and the bank advised OpenAI on its restructuring from a nonprofit earlier this year. Sarah Friar, OpenAI’s CFO, is a former Goldman equity research analyst. Morgan Stanley, meanwhile, has courted OpenAI from its early days and signed on as one of its first clients, as well as collaborating with OpenAI to build products for Morgan Stanley’s research and wealth management team.

Spokespeople for Morgan Stanley and Goldman Sachs declined to comment.

Banks’ pitches will also likely include promises to lend money to the companies in the future and agreements to become larger corporate customers, former bankers said.

One former Morgan Stanley banker said the ultimate choice of a winner in the deals is unlikely to center on “who will tell the story best.” Instead, “it’s much more, ‘What have you done for us?’”

Gapping down
In reaction to earnings/guidance:

Other news:

Gapping up
In reaction to earnings/guidance:

Other news:

The Chinese government is betting that robots will drive economic growth. But the bots can’t really do much yet.

Robots made by Chinese start-ups have danced on television, staged boxing matches and run marathons. When one company debuted its most recent robot last month, people online in China thought it looked so much like a human that workers cut the robot’s leg open onstage to reveal its metal pistons.

Despite the public fascination, concerns are growing that China’s robotics industry is moving too fast. The robots can mimic human movement and even complete basic tasks. But they are not skilled enough to handle many tasks now done by people. And with so many companies rushing into the industry, Beijing is warning of a bubble.

Over 150 manufacturers are vying for a piece of the market, the Chinese government said last month, warning that the industry was at risk for a crowd of “highly repetitive products.”

“China has an attack-first approach when it comes to the adoption of new technology,” said Lian Jye Su, a chief analyst at Omdia, a tech research firm. “But this generally leads to a large number of vendors fighting for small chunks of market.”

As it did with electric vehicles, China has gained an early global lead in making robots. China is using more robots in factories than the rest of the world combined, moving farther ahead of Japan, the United States, South Korea and Germany. Robots have transformed Chinese factory lines, doing things like welding car parts and lifting boxes onto conveyor belts.

It’s not unusual to run into a robot in Beijing. Robotic machines deliver room service in hotels and buff the floors in airports. Four-legged robots help deliver packages on university campuses. Robots cooked and served food in canteens during the 2022 Winter Olympics.

But China is also working on the next frontier of robotics: robots that not only look but think and act like people. Public and private investors spent over $5 billion this year on start-ups making humanoid robots — the same amount spent in the last five years combined.

Chinese robot makers have significant advantages. They are able to draw on the world’s strongest manufacturing sector and the backing of multiple levels of government. They are getting better at making parts like the motors and specialized screws in robot joints.

What Chinese robot start-ups have not been able to do is make humanoid robots that could transform the economy.

Experts say the humanoid robots that have been released so far struggle with unpredictable situations. They can be programmed to follow patterns, but they have a hard time reacting to events as they happen.

Chinese companies are realizing that making robots is not enough, said P.K. Tseng, a research manager at TrendForce, a market research firm in Taipei, Taiwan. “Without use cases, even if they can ship the products, they don’t know where to sell them,” he said.

Company founders and investors believe that artificial intelligence will be the answer and that humanoid robots could be how A.I. becomes a physical force in the world.

In Silicon Valley, tech executives often talk about achieving what they call artificial general intelligence. There is no settled definition, but for many it is the idea that A.I. could match the powers of the human mind.

In China, robotics companies claim they will make A.G.I. a reality.

“For people in China, A.G.I. should be something that benefits people in their everyday life,” said Sunny Cheung, a fellow at the Jamestown Foundation, which studies Chinese government influence. “Robotics is a testament of applied A.I. in real life.”

But there is a big gap between this vision and the current abilities of robots. Many Chinese robotics start-ups are working on software they hope will transform robot behavior the way large language models have transformed A.I.

One way that robots can learn to act more like people is by repetitively doing basic tasks. For example, a limited number of robots made by UBTech Robotics, which is based in Shenzhen like dozens of other start-ups, have been lifting boxes over and over again at electric vehicle factories.

Another way to train robots is by simulation, in which they watch a lot of videos of the thing they will do. Many of China’s leading robotics start-ups use software and chips made by the Silicon Valley company Nvidia to run their robots’ simulation training, Mr. Cheung said.

While no one is certain how useful humanoid robots will turn out to be, China has already put two million manufacturing robots to use. Factories in China installed nearly 300,000 new robots last year, while American factories installed 34,000.

Chinese factories have also gotten better at making robots, a major advantage over foreign firms that struggle to manufacture them in large numbers.

The start-up Unitree Robotics has announced plans to do an initial public offering, which could provide the capital it needs to help it become China’s leading humanoid robotics maker. Its latest basic humanoid robots are priced at about $6,000 in China, a fraction of the price of robots made by Boston Dynamics, long the leading American player in the industry. Boston Dynamics was acquired by the South Korean giant Hyundai Motor in 2020.

Major A.I. research labs, universities and start-ups in the United States have bought Unitree robots in recent months to test the robots’ abilities and interactions with their software.

Chinese robot makers can offer lower prices in part because they are getting a lot of funding from municipal governments and state-backed hedge funds. The Beijing government has started a $14 billion fund to invest in A.I. and robotics. Shanghai set up an embodied A.I. fund with an initial investment of about $77 million.

In Hangzhou, a tech hot spot, Unitree and a rival, Deep Robotics, are part of a group of A.I. and robotics start-ups that the Chinese media has crowned the “six dragons.” The A.I. start-up DeepSeek is another.

This month, Deep Robotics said it had raised $70 million in its latest funding round.

The humanoid robot maker Robotera said in November that it had raised more than $140 million from investors, including the venture capital arm of Geely, an electric carmaker, and the Beijing city government’s dedicated investment funds for robotics and artificial intelligence.

In late November, the Chinese central government set up a committee to establish industry standards. Members included founders, university research labs, municipal government hedge funds and Chinese state officials working on cryptography.