TheChek News Flow

Meta, SoftBank and Nvidia Splurge on Holiday AI Deals

The dealmakers of the AI sector didn’t take a break this holiday season. Meta revealed this afternoon it had bought Manus, the Singapore-based developer of AI agents. The price wasn’t disclosed but seeing as Manus’ last fundraising valuation was $500 million, Meta might have spent $2 billion or so but not hugely more than that. (More on Manus here and here). SoftBank, meanwhile, today unveiled a $4 billion buyout of DigitalBridge, an investor in data centers, demonstrating that CEO Masayoshi Son’s corporate credit card remains active despite a spate of other purchases lately (including Ampere, ABB’s robotics group, OpenAI, etc). Does Son have a credit limit?

Then there’s Nvidia CEO Jensen Huang, who doesn’t need to borrow money to buy stuff, given that he has his own dollar printing plant set up in Nvidia’s basement (that’s a joke! We mean his AI chip business). Unlike SoftBank, Nvidia doesn’t want to shout about every deal it does. Take the $20 billion it has lavished on its far-behind rival, Groq, news of which broke on Christmas Eve. Judging by that timing and the parsimonious public statements from both companies, they were hoping this deal would disappear into the ether amid the holiday carousing. After all, Nvidia probably doesn’t want people asking too many questions about how the biggest-by-far seller of AI chips is allowed to get its hands on the technology and most of the staff of a rival. Ah well, reporters have little else to do but wonder about these things.

We should acknowledge at the outset that the details that have been reported so far are all “speculation,” as public relations folks like to say. (That means they’ve been leaked to reporters unofficially by someone or other). All Nvidia and Groq have said publicly is that Nvidia is taking a “nonexclusive license” to use Groq’s technology. In other words, others can negotiate with Groq to use that tech (if you feel like it, you can too!) Oh, and Groq acknowledged it was hiring Groq’s founder, president and other team members. Actually, Nvidia is hiring 90% of Groq’s staff, according to this Axios report over the weekend. Employees, like Groq’s investors, are getting paid out for their stock at that $20 billion valuation, Axios also said. Hmmm. If you’re a sentient being, you might think that amounts to an acquisition of most of Groq‘s assets. But why quibble over semantics! Notably, The Wall Street Journal reported today there’s likely to be a bidding war for Groq's remaining assets. (For more on why Nvidia would want Groq, and why Groq might want to sell, see here and here).

Lots of other companies have done faux-acquisitions in the AI sector like this, to be sure, including Google (with Character.ai), Amazon (Adept) and Microsoft (Inflection). As our story about the Groq deal last Wednesday also noted, Nvidia did another one of these with Enfabrica. By falling short of being an actual acquisition, these deals appear designed to avoid regulatory scrutiny. So far they’ve succeeded on that front, even if there’s been reports that regulators have taken a look at one or two. This Groq deal, though, is the biggest of all of these. If antitrust regulators in the Trump administration are doing any work at all, this should attract their attention.

Making Money With Intel

One place where both SoftBank and Nvidia have made a quick buck on investments is in shares of Intel. Both companies announced in the late summer they would invest in the ailing chipmaker, although Nvidia only completed its $5 billion investment on Friday. (It had to wait for regulatory approval for that one!) SoftBank completed its $2 billion investment in September.

Since the two deals were announced, Intel’s stock has risen more than 50%. SoftBank, for instance, bought 86.9 million shares for $23 apiece while Nvidia got 214.8 million shares for $23.28. Intel today closed at $36.68, which means SoftBank has already made $1.2 billion while Nvidia has made $2.9 billion. Not bad!

One point in Waymo’s favor is that it will likely draw revenue from sources beyond ride-hailing. Waymo co-CEO Tekedra Mawakana has said the company plans to expand beyond its robotaxi service to local delivery and long-haul trucking, and eventually licensing its autonomous-driving technology to automotive makers.

Its partnership with DoorDash to offer autonomous food delivery in Phoenix, announced in October, may be a preview of what comes next. At Uber, food delivery makes up one-third of revenue and is growing faster than ride-hailing. It also accounts for about 40% of Uber’s operating earnings, excluding stock compensation, depreciation and amortization.

For Waymo, licensing its software to automakers is likely further off. But those fees could eventually boost margins because the overhead cost to maintain and improve the software likely won’t be as onerous as the cost of putting more robotaxis on the road.

Fortunately for Waymo, if it proceeds with the fundraise, there’s some precedent for cash-burning ride-hailing startups eventually growing into their valuations.

In December 2014, then-private Uber raised money at a $40 billion valuation. That was double the valuation the startup secured in a previous round announced that June, and it was the highest for any venture-backed startup at the time—raising eyebrows among investors and analysts. Uber, then five years old, was generating only hundreds of millions of dollars in revenue at the time, though the specific amount wasn’t publicly known then.

Concerns about Uber’s valuation turned out to be overblown. The company generated about $1.5 billion in net revenue in 2015, roughly three times as much as in 2014. That suggests its backers ultimately invested at a multiple of about 27 times forward sales.

Although Waymo doesn’t need to pay human drivers to run its service, there are big questions lingering around other long-term costs. The cost of insurance for autonomous vehicles is one key area of uncertainty. Another concern is that the detection sensors and accompanying hardware installed on Waymo’s vehicles are expensive.

When it comes to those hardware costs, Waymo is at a disadvantage compared with Tesla. Tesla’s autonomous vehicles use less-expensive cameras to map roadways. For now, they are available to ride-hailing passengers in only a few cities, with a human driver behind the wheel for supervision. In their report this month, Morgan Stanley analysts estimate that Waymo’s cost per mile is around $1.43 currently, while Tesla’s is $0.81.

It remains to be seen whether that difference will translate to an enduring advantage for Tesla. The Morgan Stanley analysts say newer Waymo vehicles that the company plans to deploy in 2026 will likely cost less to operate, at about 99 cents to $1.08 per mile. That’s partly because Waymo is testing less-expensive base vehicles—cars before AV tech is incorporated—from Hyundai and Chinese firm Zeekr rather than the premium Jaguar I-Pace models currently used for most public Waymo rides. Insurance costs could also fall if robotaxi safety records continue to improve.

Costs aside, Waymo has a leg up over its robotaxi rivals when it comes to regulatory approval. Tesla, like Amazon’s robotaxi unit Zoox, is still awaiting clearance to charge riders for paid rides in cities where it operates. Longtime Tesla bull and venture capitalist Chamath Palihapitiya, for one, noted in a post on X that “it will be incumbent on Tesla to show their error tolerances are as good or better than Waymo.” He suggested Tesla would need that kind of performance to convince regulators to let it operate more broadly.

Even if Tesla or Zoox secures the right permits, however, Waymo has already built a solid lead by being first in several major cities. If it continues to focus aggressively on expansion and market share like Uber did in its early years of growth, its moat could prove even wider than that of Uber, whose market capitalization is over 20 times larger than its rival Lyft’s.

Uber was worth $82 billion when it went public in 2019, not far off from its highest private valuation. It’s now profitable, more diversified and worth about twice as much. Waymo’s investors seem poised to make a similar bet.

Investors including D1 Capital Partners, Fidelity and Ontario Teachers’ Pension Plan to purchase $1bn in equity

Octopus Energy has reached a deal to sell a minority stake in its Kraken Technologies unit, a transaction that would value the software business at $8.65bn and pave the way for its potential public listing.

The UK energy group is selling about $1bn of Kraken equity to a syndicate of investors including D1 Capital Partners, Fidelity and a unit of the Ontario Teachers’ Pension Plan, according to the company.

The vast majority of the $1bn proceeds from the investment in Kraken — about $850mn — will go to Octopus Energy, and the remaining $150mn to Kraken.

The deal underlines investor appetite for utility technology as they bet that power companies will need to upgrade their software systems to help customers move towards lower-carbon technologies such as solar panels and electric cars.

Greg Jackson, founder of Octopus Energy, said the investment would allow Kraken to “grow even faster”, adding that the software arm was “in a class of its own in terms of technology, capability and scale”.

Existing shareholders will retain a stake in Kraken, alongside Octopus Energy and the new investors.

Octopus Energy said the investment was a step towards Kraken’s formal independence and demerger from the group.

The move sets up an initial public offering of Kraken potentially within the next year or two and could enable an eventual flotation of Octopus Energy, the people added. Kraken could float in London or New York.

Octopus Capital, one of Octopus Energy’s largest investors, and other shareholders would provide $320mn to it for “innovation and growth”, Octopus Energy also announced on Monday.

It confirmed this year that it was one of three UK retail energy companies that had not yet met regulator Ofgem’s capital adequacy targets, intended to ensure utility companies are financially robust.

Octopus Energy announced plans in September to spin out Kraken as a standalone company, a move that executives said would allow the business to speed up its global growth and pursue its ambition of serving 1bn people within the next decade. After the split, Octopus Energy will retain a 13.7 per cent stake in Kraken.

The transaction marks the first standalone investment round into Kraken since the software business was developed inside Octopus Energy.

Amir Orad, chief executive of Kraken, said becoming an independent company would give the business “the focus and freedom to scale as a neutral, global operating system for utilities”.

Kraken’s software was developed in-house by Octopus Energy and is now licensed to rival utilities, turning it into one of the group’s most valuable assets.

The platform is licensed to companies including EDF and Eon. It is used by energy and water groups to manage everything from customer accounts to batteries, renewable generation and electric vehicle charging.

Kraken is contracted to serve more than 70mn customer accounts worldwide, according to the company.

Analysts at UBS said recently that Kraken was on track to exceed its target of serving 100mn customer accounts by 2027, though they cautioned that the software’s competitive advantage could narrow over time.

Octopus Energy has steadily taken steps to distance Kraken from its retail energy business to avoid perceptions of conflicts of interest. Former BT Group chief executive Gavin Patterson was appointed chair of Kraken in 2023, and the software arm has built its own management team and governance structure.

Founded in 2015, Octopus Energy overtook British Gas this year to become the UK’s largest household energy supplier. Its rapid growth and technology-led model have attracted investors from around the world, including the Canada Pension Plan Investment Board and Generation Investment Management, which is chaired by former US vice-president Al Gore.

Trump threatens strikes against Iran if it restarts nuclear programme

US president warns Hamas it has a ‘very short period of time to disarm’ after talks with Israeli leader Benjamin Netanyahu

Donald Trump has threatened to launch fresh strikes against Iran if it is found to be rebuilding its nuclear programme, as he held talks with Israel’s Prime Minister Benjamin Netanyahu.

Trump also said Hamas has a “very short period of time to disarm” in Gaza, where a US-brokered ceasefire has largely held since October, or the Palestinian militants would have “hell to pay”.

The warnings come as the US president’s efforts to broker peace across the Middle East have faltered. The implementation of the US peace plan in Gaza has stalled and tension with Iran appears to be rising.

Trump claimed that Iran’s nuclear enrichment programme was “obliterated” after the US briefly joined Israel’s 12-day war with the Islamic republic in June, before declaring an end to the conflict.

“Now I hear that Iran is trying to build up again and, if they are, we have to knock them down . . . We’ll knock the hell out of them,” Trump told reporters on Monday at his Mar-a-Lago resort in Florida.

Standing beside Netanyahu, Trump said: “Iran may be behaving badly. It hasn’t been confirmed. But if it’s confirmed . . . the consequences will be very powerful.”

The shift in Trump’s tone marks a victory for Netanyahu, whose government has argued that Israel’s regional adversaries, including Hamas and Iran, are uninterested in the peace proposals that Trump is pushing.

Israel alleges that Iran is rapidly rebuilding its ballistic missile arsenal. Netanyahu has also suggested that the Trump administration’s push to implement the next phase of Trump’s Gaza peace plan is unrealistic because Hamas is unwilling to disarm.

Trump’s peace plan has faltered over issues such as Hamas’s disarmament, the deployment of an international security force and Gaza’s future governance.

The fragile ceasefire between Israel and Hamas in Gaza, which ended two years of heavy Israeli bombardment in the Palestinian territory, has been tested by near daily eruptions of fighting and Israeli strikes.

Palestinian health officials say more than 400 Palestinians have been killed since the October ceasefire.

US officials have insisted that “phase 2” of Trump’s 20-point peace plan will begin in January.

That would include the unveiling of the international “board of peace” and an executive committee to oversee the strip, a Palestinian technocratic committee on the ground to handle daily governance, as well as the deployment of a multinational peacekeeping force.

Israeli troops — which at present still hold about half of Gaza — would gradually withdraw as a multinational force enters the strip to monitor Hamas’s disarmament and the transition to postwar reconstruction.

Israeli officials worried ahead of the meeting that Trump might try to push Netanyahu to agree to further troop withdrawals.

But as the two leaders emerged from their meeting, Trump declined to talk about an Israeli withdrawal, and said Hamas was failing to comply with the deal.

“I’m not concerned about anything that Israel is doing,” Trump said when asked about the pace of implementation. “They’ve lived up to the plan.”

Israeli officials have dismissed the likelihood that an international or Palestinian security force would undertake Hamas’s disarmament.

Trump said Monday that other countries “will come in and do it”, insisting that dozens of governments had backed his 20-point peace plan. But no country has committed to sending troops into Gaza.

Trump said he and the Israeli leader spoke at length about the occupied West Bank, where the US has opposed Israel’s rapid expansion of Jewish settlements.

“I wouldn’t say we agree on the West Bank 100% but we will come to a conclusion on the West Bank,” Trump said, adding that Netanyahu “will do the right thing”.

Trump also urged Netanyahu to “get along” with Syria and its new President Ahmed al-Sharaa, who he said “is working very hard to do a good job”.

Israel has continued to occupy additional swaths of south-western Syria, despite US pressure to withdraw, since the overthrow of Syrian dictator Bashar al-Assad late last year, as it seeks to bolster its security.

Chinese toymaker plans store on New York’s Fifth Avenue as it faces pressure to continue momentum of its popular dolls

Pop Mart is set to significantly expand its global presence next year as the Chinese toymaker faces pressure to continue the momentum of its popular Labubu dolls and ensure it is not a one-hit wonder.

The Beijing-based company is planning to open a flagship store on Fifth Avenue in New York, according to a person familiar with its plans, as well as a separate location in Times Square. They are part of dozens of expected openings in the US, where it already has close to 60 stores.

This month it opened a 600-square-metre store in Sydney, its largest so far in the southern hemisphere, and expects to increase its footprint in Europe and Japan, the person added.

Pop Mart’s revenues and valuation have soared this year on the back of its furry elf character Labubu. Celebrities including Dua Lipa, Rihanna and Lisa of K-pop group Blackpink have been spotted carrying the palm-sized plush toys, and scuffles have broken out in stores over limited editions.

Their popularity has taken Pop Mart’s market capitalisation to $35bn, double that of US groups Hasbro and Mattel combined.

But the company is now under pressure to maintain its momentum, as recent declines in the toys’ resale prices raise the prospect of “peak Labubu”. Its Hong Kong-listed shares have fallen about 40 per cent from their all-time high in August, and analysts have pointed to the need for new products or characters.

“Labubu doesn’t actually have that much of a story behind it, it’s just a strange-looking doll,” said Chris Pereira, founder of consultancy iMpact. “I don’t think it’s what they were expecting, for it to go so crazily viral.”

He added: “The question is ‘what next’? The direction I expect them to be looking at is cross-branding . . . say with Disney or with some other movie chain.”

Pop Mart sells many of its products in “blind boxes” that typically retail for about $10 in China. Customers do not know which specific character in a series they have bought until after unboxing them, driving repeat purchases and a vibrant secondary market, particularly among collectors of full sets.

Investors, analysts and toy enthusiasts have closely watched the resale value of Labubu dolls for signs of whether their popularity is ebbing. Pop Mart has increased its production of plush dolls, including Labubu, 10-fold this year, the company has said, and produces about 30mn plush dolls a month.

The company is one of the few Chinese consumer brands to gain global recognition. Facing intense competition and economic pressures at home, companies such as Luckin Coffee, a lower-priced competitor to Starbucks, milk tea chain Chagee and electric vehicle maker BYD have looked abroad for growth.

“A lot of them, when they reach out to me, they’ll say the [Chinese] economy’s not great [nor are] the prospects for growth in mainland China,” said Pereira, who works with about 400 Chinese companies, including Pop Mart, on foreign expansion.

“Because of the competitive nature of the Chinese economy, going overseas really is easier for these companies. They’re more efficient, more scrappy than their western counterparts.”

Pop Mart said revenues in the Americas increased by 1,270 per cent year on year in the third quarter against overall revenue growth of 250 per cent.

The company did not disclose revenue by value for the third quarter but said its half-year revenues were just under Rmb14bn ($2bn), three times higher than the same period a year earlier.

“The pace of store openings has lagged behind the surge in consumer demand, particularly in the US,” said Lina Yan, a consumer analyst at HSBC, who noted there were more than 100 Pop Mart stores in Beijing and Shanghai each.

“Were it to mirror the strategy deployed in Beijing and Shanghai, it could open as many as 200 more stores in the country,” she said.

The deal for more than $2 billion is one of the first in which a major U.S. tech company has bought a startup with Chinese roots

Meta Platforms has agreed to acquire AI startup Manus, a Singapore-based company with Chinese founders that conducts deep research and performs other tasks for paying users.

Meta META is closing the deal at more than $2 billion, according to people familiar with the acquisition. Manus was seeking a fresh round of fundraising with a valuation of $2 billion when Meta approached the startup, some of the people said.

Manus’s co-founder and chief executive, Xiao Hong, who often goes by the nickname “Red,” will report to Javier Olivan, chief operating officer of Meta, some of the people said.

The terms of the deal weren’t disclosed. The acquisition is among the highest-profile examples of a major U.S. tech company buying an AI product developed in Asia’s AI and startup ecosystem.

Manus gained a wide following after previewing an AI agent in March that was capable of producing detailed research reports and building custom websites, using AI models developed by companies such as Anthropic and China’s Alibaba. That demo followed the release of DeepSeek, a made-in-China AI model that rocked Silicon Valley due to its advanced capabilities, coupled with claims by its developer that it was developed with far less computing power than American rivals.

The deal is a move in a new direction for Meta, which is investing aggressively in AI to compete with Google, Microsoft and OpenAI. The deal would help the social-media giant cement its position in the product segment of AI agents, an increasingly intense battlefield of AI companies that make tools to conduct complex tasks with minimal human input. Microsoft has operated a popular AI assistant Copilot.

Meta says it plans to continue to operate and sell Manus’s service and integrate it into its suite of social media products. Meta has previously touted so-called “open source” models that are largely free to access, modify or distribute.

“We plan to scale this service to many more businesses,” Meta said in its announcement about the deal.

Meta’s existing AI offerings are widely available for free in services including Instagram and WhatsApp, and the company has also fully incorporated AI into its advertising in ways that have fattened its bottom line, according to analysts.

After Meta faced unexpected challenges earlier this year while preparing to roll out a new model, Chief Executive Mark Zuckerberg went on a recruiting blitz to build an AI dream team, offering top executives and researchers multimillion-dollar paydays.

The company acquired a 49% stake in startup Scale AI that valued Scale at $29 billion, and Scale founder Alexandr Wang joined the social media giant as its chief AI officer.

Manus has garnered millions of users since its spring launch, including some who pay subscriptions to use its models for analysis, coding and other tasks.

In April, Manus raised $75 million in a fundraising round led by venture firm Benchmark. As part of the deal, Benchmark’s general partner Chetan Puttagunta joined the company’s board. After the investment, Manus’s valuation was boosted to $500 million, people familiar with the matter said. It also counts HSG, ZhenFund and Tencent as major investors.

In December, the startup announced it had crossed $100 million in annual recurring revenue, eight months after it launched. Around the same time, Meta started negotiating with the company for the acquisition, some of the people said.

“Joining Meta allows us to build on a stronger, more sustainable foundation without changing how Manus works or how decisions are made,” Xiao said in an announcement Monday. Xiao is one of two Chinese co-founders of the company.

The parent company behind Manus, Butterfly Effect, was founded in 2022 and had offices in Beijing and Wuhan. In October 2024, the company started developing Manus. Although most of its researchers and engineers were based in China, Manus was launched outside the country as it used many American AI models that aren’t available in China.

After securing investment from Benchmark, the company officially moved its headquarters to Singapore. U.S. lawmakers have criticized Benchmark for backing an AI company with ties to China. Manus has also shelved its plan to develop a version for the Chinese market, people familiar with the matter said.

Manus has around 100 employees, mainly in Singapore.

Other news: