In April 2025, Eric Glyman, co-founder and co-CEO of Ramp, received some bad news: A deal was falling apart.
Glyman had been hoping Ramp, a highly valued startup that sells corporate credit cards and expense management software, could strike a partnership with Airwallex—a fast-growing fintech founded in Melbourne, Australia—that would improve Ramp’s ability to handle international payments. But now Airwallex CEO Jack Zhang was pulling the plug on the proposed deal—partly because of pressure from investors who wanted his company to compete with Ramp, according to employees of both companies.
Glyman made an attempt to salvage the partnership, according to people familiar with discussions between the two companies. Zhang demurred. Soon after, Glyman warned Zhang that the relationship between their companies might not be so friendly in the future, according to people familiar with the situation.
In the wake of the canceled deal, some of Ramp’s high-profile, politically connected investors have launched a public and private campaign to urge the U.S. government to saddle Airwallex with a government investigation that probes any links the company may have with the Chinese government. Those investors have alleged that the Chinese government could access customer data passing through Airwallex subsidiaries in China.
The U.S. Treasury Department on Thursday proposed to expand existing outbound investment restrictions out of national security concerns, in a move set to further decouple venture capital ties between the U.S. and China.
The proposed regulation widens the scope to prohibit Americans from ...
DeepSeek has closed its second funding round, raising 106 billion yuan ($15.9 billion) in one of the biggest-ever fundraising deals in China’s AI sector, according to three people with knowledge of the matter.
One quotation has rattled around in my head for a while, which I first heard from early Anthropic investor Jaan Tallinn: “The future of the world will be decided by a group of people who fit in an elevator.” The enigmatic billionaire said on a podcast earlier this year that a friend had used the illustration to describe the concentration of power at the leading AI companies.
I’ve been reflecting on it as Anthropic prepares for its Wall Street debut because the company’s small group of powerful insiders, who are set to realize incredible wealth at unprecedented speed, face their biggest test in explaining themselves to the world outside their elevator. To raise the money Anthropic leaders need, and to keep raising it again and again, big fund managers need to fully buy in.
On their earnings calls earlier in the week, Meta Platforms and SpaceX delivered the same underlying message to investors: demand for AI is running ahead of the supply of AI compute. On Thursday, another important player in the AI ecosystem—Blackstone, the alternative asset management giant—repeated a version of that refrain.
On its own earnings call, the firm, a major investor at the center of the AI data center boom, faced pointed questions from analysts about rising financing costs and mounting investor concerns over the sustainability of the buildout. Its executives argued that the very factors threatening the AI industry’s expansion could actually work in Blackstone’s favor. CEO Steve Schwarzman noted that constraints on power availability and growing public opposition to new data centers are making it harder to bring new capacity online, potentially making existing facilities more valuable. Blackstone President Jon Gray also suggested the firm could seek higher returns on projects where it hasn’t signed leases yet in order to offset rising borrowing costs.
Blackstone’s third-quarter results offered some support for that argument: Its infrastructure investments were buoyed by what executives called “extraordinary” data center leasing momentum, significantly outperforming its broader private equity and real estate holdings. “Despite all the negative things you’re reading about AI today, when we look across our business, this continues to be the area where we’re driving the highest returns,” Gray said during the call.
But it’s hard to ignore the fact that the financing concerns and roadblocks threatening the broader AI infrastructure buildout are also hitting closer to home for Blackstone. Fears about rising financing costs have sent CoreWeave’s stock and bond prices plunging—the neocloud’s shares closed down another 10% Thursday after sliding last week—while Australian AI cloud provider Firmus abruptly pulled its initial public offering earlier this month. Blackstone has provided billions in loans to CoreWeave and has both invested in and lent money to Firmus.
Blackstone has also been pressing ahead with brand-new AI ventures. In May, the firm announced an initial $5 billion equity commitment to a joint venture with Google to create a new cloud company that will use Google’s TPU chips, with plans to bring 500 megawatts of capacity online by next year. The venture could need as much as $25 billion in total investment, including some borrowed money, and it has reportedly encountered delays at potential data center sites.
Blackstone, for its part, says that broader anxiety about AI hasn’t undermined its ability to raise fresh cash for making its own investments. “We’re continuing to be able to raise capital. Obviously investors are asking the same question around AI, but what we continue to see is demand and usage of the AI going up faster than the compute,” Gray said.
While Blackstone says it wants to continue going big in AI, it can only finance a tiny portion of the industry’s broader infrastructure buildout itself. The trillions of dollars of financing needed for chips and data centers in the coming years requires raising money from a far broader pool of investors. Blackstone itself has said it’s not clear where all the money will come from.
Jas Khaira, the head of the team leading Blackstone’s push into AI investment, said late last month that the firm is still trying to determine which investors across the public debt and private credit markets will be able to provide the funding needed to sustain the buildout.
“We have not mapped out who’s going to buy all the debt. We have some hypotheses, but the honest answer is we don’t know where all of it will come from,” he said at The Information’s AI Agenda event in San Francisco.
Data Center Headwinds
Despite Blackstone’s overall positive note on AI demand, we got fresh evidence Thursday afternoon of the uncertainty in the IPO market for companies in the data center business. Phoebe Liu and Valida Pau had a scoop that SB Energy, a company controlled by SoftBank Group that is developing an OpenAI data center in Ohio, is discussing a plan to put off its IPO until December or earlier next year. Previously, SB Energy was aiming to go public in September. The best hope for an SB Energy IPO is that the supply of AI compute gets so tight that investors start warming back up to data centers.—Nick Wingfield
In Other News
Meta Platforms’ stock fell 7% as investor enthusiasm over its third-quarter earnings report on Wednesday gave way to concerns about the probability that it will need to issue new equity to pay for its AI spending.
Andreessen Horowitz renamed its flagship blockchain fund a16z Frontier, dropping “crypto” from its name (The Information’s earlier scoop on the change is here).
Google partnered with energy company Voltus to pay homes and businesses to use less electricity during periods of peak demand.
Today on The Information’s TITV
Check out today's episode of TITV, in which Akash Pasricha speaks with reporters Aaron Holmes and Leo Schwartz about the PHP security patch mayhem.
Recommended Newsletter
Start your day with Applied AI, the newsletter from The Information that uncovers how leading businesses are leveraging AI to automate tasks across the board. Subscribe now for free to get it delivered straight to your inbox twice a week.
The Information's WTF Summit returns to Napa Valley, bringing together the women shaping tech, media and finance. Confirmed speakers include Bari Weiss, Whitney Wolfe Herd (Bumble), Kim Posnett (Goldman Sachs) and Whitney Cummings. Expect candid conversations and a room built for meaningful exchange. Limited tickets remain—reserve your spot now.
Trust is quickly becoming the defining question of the AI era. Over dinner, Jessica Lessin, Founder & Editor-in-Chief of The Information, and Whitt Butler, EY Americas Vice Chair – Consulting, will talk candidly about how leaders can build that trust and keep it. Request an invite.
Join The Information and NYSE for a cocktail reception and conversation on what it really takes to go public, hosted by The Information’s Dakin Campbell. The panel will explore the trust, relationships and preparation companies need to build with advisers, investors and exchange executives, along with the many complexities founders and executives must navigate along the way. Sponsored by NYSE. Apply to attend.
OpenAI is aiming to close a $30 billion fundraising round that will value the AI company at up to $1.25 trillion, not including the new investment, before the end of next week, according to a person familiar with the deal.
New York-based investment firm Coatue Management and investors from the ...
Salesforce is leading the software industry’s attempt to collect fees from businesses that use Claude and other external AI tools to automate work in enterprise apps.
As soon as next month, Salesforce plans to require companies to register outside AI agents, a step toward charging them each time those agents update or analyze data stored in Salesforce. The fees let Salesforce earn money from customers who use such AI instead of paying for Salesforce’s built-in AI products.
The Salesforce tollgate has implications for the rest of the industry, which is coming to terms with the idea that so-called AI agents may be able to use software firms’ apps more efficiently than humans do, meaning their seat-based or per-user pricing model might not have staying power. The question is whether the new AI tollgates—which Atlassian, Workday, ServiceNow and Datadog are also erecting to various extents—and other pricing shifts will more than make up for an eventual decline in per-seat subscription revenues.
The vast majority of workers haven’t yet outsourced such work to AI, of course, and companies like Salesforce are trying to get more revenue from them as well. They charge fees based on how much they use the apps’ built-in AI tools and are also experimenting with outcome-based pricing, meaning customers only pay when such tools actually complete a task successfully or help the customers make more money.
For now, Salesforce and some other software firms developing tollgates for customers’ AI agents haven’t started charging the tolls. They’re letting customers experiment for free with the technology a little while longer. (The customers still have to pay providers such as Anthropic to use the AI agents in the first place.)
Starting around November, according to the company’s website, Salesforce will require customers to register any externally-made AI agents if they plan to use them to take actions on their behalf in Salesforce apps. Each action an agent takes in Salesforce counts as a billable “interaction,” according to the company. The move has lifted investor confidence in Salesforce recently.
Salesforce describes the registration requirement as a security measure because the company gives each agent its own identity and narrower data-access permissions, compared to humans. It isn’t clear how Salesforce will clamp down on customers using their own AI agents, such as OpenClaw-powered ones that run on a person’s computer and can access any app that runs on it. (Salesforce did not respond to requests for comment.)
Salesforce hasn’t announced prices tied to such agent usage, but customers could eventually end up paying a lot more money than they do now, given how many of them are already using their own AI on top of Salesforce apps and bypassing Salesforce’s own built-in AI tools, according to software consultants and chief information officers.
Nick Hamm, a software consultant, said that most of the 150 enterprises he works with are currently using AI from Anthropic and OpenAI to automate work inside Salesforce apps, for free. But they’re reconsidering that usage in anticipation of the upcoming fees. If the fees are too high, the customers may move their business data outside of Salesforce’s apps so they can use AI to analyze the data without paying an extra cost, he said.
New customers will have a three month grace period after the new pricing is announced to register their AI agents, and the pricing won’t affect existing customers until their contracts are renewed.
In addition to the billable “interactions,” customers have the option to buy unlimited agent usage by purchasing a “headless license,” a Salesforce account executive toldHamm in a LinkedIn message viewed by The Information.
Some customers such as Lumen, a fiber-optic network provider to enterprises and data centers, say they are waiting to test Salesforce’s new connectorsuntil pricing is disclosed.
So far, Lumen has leaned on Salesforce’s built-in AI. A longtime customer of Salesforce’s customer relationship management software, Lumen made a “major” investment in using Salesforce’s AI sales agents in the last 12 months, according to CIO Jim Fowler. The sales agents help with tasks such as collecting customer information and generating sales quotes for the customers to buy fiber connections between data centers, saving Lumen’s human salespeople considerable time, he said.
So far, Fowler said, Salesforce has been “super flexible” about letting Lumen use existing subscription fees to also cover its use of built-in AI features such as the sales agents.
Early customer reactions to the pricing announcement will likely determine how flexible Salesforce is with customers that start renewing their contracts after November. If its past AI pricing efforts are any guide, Salesforce could easily switch things up depending on the feedback, or any blowback, it receives.
The Information's WTF Summit returns to Napa Valley, bringing together the women shaping tech, media and finance. Confirmed speakers include Bari Weiss, Whitney Wolfe Herd (Bumble), Kim Posnett (Goldman Sachs) and Whitney Cummings. Expect candid conversations and a room built for meaningful exchange. Limited tickets remain—reserve your spot now.
Trust is quickly becoming the defining question of the AI era. Over dinner, Jessica Lessin, Founder & Editor-in-Chief of The Information, and Whitt Butler, EY Americas Vice Chair – Consulting, will talk candidly about how leaders can build that trust and keep it. Request an invite.
Join The Information and NYSE for a cocktail reception and conversation on what it really takes to go public, hosted by The Information’s Dakin Campbell. The panel will explore the trust, relationships and preparation companies need to build with advisers, investors and exchange executives, along with the many complexities founders and executives must navigate along the way. Sponsored by NYSE. Apply to attend.
SB Energy, the SoftBank Group–controlled developer of a massive planned Ohio data center project for OpenAI, is considering going public in December or possibly waiting until next year, according to people familiar with the discussions.
That is a later timeline than previously expected. The data center developer was aiming to go public as soon as September, The Information reported in August. But since then, the public market has soured on initial public offerings from data center and AI cloud companies. Australian cloud firm Firmus yanked its IPO plans earlier this month, and others are considering delaying their offerings.
Enterprise software firm Datadog recently offered to buy cloud cybersecurity startup Upwind Security for nearly $4 billion, the latest sign of white-hot interest in such firms, according to someone with direct knowledge of the talks.
Upwind, which hasn’t accepted the offer, is finalizing a $300 million funding round at a roughly $3.8 billion valuation, meaning the offer doesn’t carry much of a premium over the private valuation, said two people with knowledge of the situation. The startup sells software to keep tabs on applications, including so-called AI agents, that run in the cloud and to detect and automatically respond to potential evidence of hacks or breaches.
Tractian, a 7-year-old physical AI startup based in Atlanta, Georgia, has raised $130 million at a valuation of $1.4 billion in a Series D funding round led by Sapphire Ventures. Tractian provides sensors to assess the condition of machines, predicting how long it will take for them to fail and ...
Tractian, a 7-year-old physical AI startup based in Atlanta, Georgia, has raised $130 million at a valuation of $1.4 billion in a Series D funding round led by Sapphire Ventures. Tractian provides sensors to assess the condition of machines, predicting how long it will take for them to fail and ...
Google said it is partnering with energy company Voltus to pay homes and businesses to curb electricity use during peak hours, in the latest effort to mitigate the data center boom’s impact on power demand.
The companies on Thursday announced what they call a 1-gigawatt “virtual power plant,” ...
OpenAI’s systems had multiple vulnerabilities that would allow AI agents to circumvent monitoring and take arbitrary unauthorized actions, evaluators from AI safety nonprofits METR and Redwood Research found during a three-week pilot with OpenAI from late July to early August. The nonprofits ...
While much of the AI industry has fixated on new personal AI assistants from Meta and OpenAI, video AI is enjoying a quiet renaissance. Hollywood mogul Jeffrey Katzenberg is teaming up with the former head of OpenAI’s Sora app to launch a new AI film company. Startup Higgsfieldquadrupled its valuation to $5.4 billion in August. And Kling AI, the video unit of Chinese social media giant Kuaishou Technology, raised nearly $3 billion at a $15 billion valuation over the summer.
The latest video AI company to seize this moment is Utopai Studios, a 4-year-old film startup based in Mountain View, Calif. Utopai closed a funding round for $60 million that valued it at $1 billion and has already received term sheets for a subsequent round that would value the company at $2 billion to $3 billion, according to someone close to the company.
The enthusiasm for video AI startups follows improvements in the quality and length of the clips that video AI models can generate. In September, Higgsfield announced that it had crossed $1 billion in annualized revenue, showing that the technology has serious commercial potential.
Israeli security firm Lava said it found 32 Nvidia Blackwell Ultra B300 chips through monitoring systems linked to the network of China Telecom, a state-owned telecommunications company. The B300 is one of Nvidia’s most powerful data-center chips and is barred from being sold to China by ...
Andreessen Horowitz, which made one of the boldest bets on blockchain investments in the venture world, is dropping the term crypto from the name of its fund focused on the category.
A16z crypto, which has raised nearly $10 billion across five funds, has informed some of its limited partners that it plans to change its name to a16z frontier and broaden its investment strategy beyond crypto, according to a person briefed on the matter. One of the firm’s general partners, Ali Yahya, also informed some investors that he is exiting the firm to work on his own startup, the person said. Yahya joined a16z in 2017 as the first full-time member of its crypto team.
Meta Platforms’ decision to revamp its stock structure was shaped by factors including expectations that the company could have to raise hundreds of billions of dollars in coming years and concerns that a proposed wealth tax in its home state of California could prompt Chief Executive Mark ...
Microsoft said it has observed cyberattacks by a Chinese state-aligned actor known as Storm-0940 that used a Chinese AI agent called Artex to automate the attacks targeting hundreds of organizations globally across a wide range of industries.
The attacks took place as early as September. In ...
Tesla reported flat revenue while surprising analysts by reporting free cash flow for its third quarter.
On Wednesday, the company said its revenue for the period was $28 billion, unchanged from a year earlier. Instead of burning $4.92 billion in cash for the quarter, as analysts had forecast, ...
At first, the X headlines this morning about the latest AI-powered cybersecurity incident seemed dire. But by the time I had a cup of coffee and began parsing all the cyber jargon, I was reassured that an apparent solution was on its way to plug a major new security hole that could make websites vulnerable to cyberintrusions. Since August, a coalition led by OpenAI has been working to address critical vulnerabilities in PHP, an open-source programming language that undergirds the web. But there was bad news, too: This type of long-simmering response to the latest AI-discovered vulnerabilities may not be viable for much longer.
For months, OpenAI has been harping on what it calls the “defender’s window”—the critical period of time when cybersecurity experts, armed with advanced AI models, can identify and patch vulnerabilities before attackers exploit them first. The issue is that the gap is rapidly narrowing between cybercapable models from leading U.S. frontier labs and Chinese open-source models, which lack the restrictions U.S. labs impose on their models’ capabilities. Anthropic said as much in a blog post from late September warning that Chinese lab Z.ai’s GLM-5.3 model only lags the U.S. frontier models by about four months.
The good news: The Trump administration has a shiny new organization created for moments just like these. It’s called the Super Intelligence Force, a name that may elicit eyerolls from some of our more jaded readers. Still, I have heard over and over that all the AI industry wants is coordination and clarity from the White House. That, presumably, is what the SI Force will offer, led by heavy hitters including National Intelligence Director Jay Clayton and senior Defense Department official Emil Michael.
An OpenAI spokesperson told me today that they had briefed several agencies, including the Cybersecurity and Infrastructure Security Agency, about the PHP vulnerability, as well as briefing Clayton and Michael late last week, representing the first “major engagement” between OpenAI and the SI Force. The spokesperson added that the SI Force provided a central point to coordinate their communication across the broader government. According to a White House official, the SI Force’s work involved helping to put pressure on the “relevant parties” to deploy a vulnerability patch.
While the SI Force’s coordination function may not seem terribly ambitious, at least it’s a start. It’s worth remembering that the White House essentially had to reorient its entire AI policy in April with the announcement of Mythos. The arrival of Chinese cybercapable open models could represent a Mythos 2.0 moment. I spoke with one former White House official who told me solutions from even a few months ago, like the July Gold Eagle clearinghouse that allows organizations and government agencies to share and fix software vulnerabilities, might already be outdated. It’s up to the SI Force to create a path forward. The next major internet bug likely won’t have the luxury of a two-month-long defender’s window.
Meta’s Performance
On Sunday, in a preview of this week’s Meta Platforms earnings report, your regular Briefing columnist, Martin Peers, declared: “If Meta can stay on investors’ good sides, an equity offering seems in the cards.” Martin nailed it on both accounts. On Wednesday afternoon, Meta CEO Mark Zuckerberg delivered a bunch of (mostly) good news that had investors in the company purring like kittens. Meta also announced the creation of a new category of common stock that will allow it to raise capital to fund its hefty AI investments without sacrificing Zuckerberg’s control.
The highlights: For the third quarter, its revenue grew 26% to $64.6 billion, slightly above what the company had forecast. It’s trimming the losses from its Reality Labs unit. While it showed negative free cash flow of $2.12 billion rather than the positive free cash flow of $10.63 billion in the year-earlier period, that was due to its AI investments, and the number wasn’t as bad as some analysts had expected. The company’s anticipated capital expenditures of $135 billion to $145 billion for the full year, while eye-popping, were still within the range it had previously forecast. That was reassuring for investors who’ve gotten spooked in the past by sudden leaps in tech company spending on AI.
Perhaps equally comforting were Zuckerberg’s reassurances that investors will see a payoff from some of its most expensive AI initiatives. He described Muse, Meta’s new agent app, as a bona fide hit for the company, one that has gained 10 million users in the U.S. and Canada in its first six weeks on the market—a faster pace of growth than for its Threads app. While Zuckerberg said the product will remain free for most people, he described a number of ways Meta plans to eventually get paid for it, including commerce transaction fees paid by businesses and money power users will pay for access to advanced models and extra tokens. Meta shares rose 2% in after-hours trading.
Meta also warned of some potentially expensive new legal costs on the horizon, including from lawsuits related to child safety legal cases. During the third quarter, Meta said, its operating income declined 37% from a year earlier to just under $13 billion, due in large part to a nearly $10 billion charge related to legal proceedings. In August, the company settled a child online safety case with 52 attorneys general for $18.59 billion. And on Wednesday, a New Mexico judge ordered Meta to pay $25 billion in a case related to the Cambridge Analytica scandal.
Meta did not say when it plans to raise capital through a new equity offering, or how much it might raise. Presumably, that will happen after Jan. 19, the effective date for issuance of the new nonvoting Class C common stock the company announced today. The company’s existing shareholders will receive a dividend of one Class C share for each share of Class A or B common stock they hold. And crucially, when the company raises money in the future with its Class C stock, that won’t dilute Zuckerberg’s voting power. “The creation of the Class C common stock will continue the founder-led model that has driven the company’s success since its inception,” Meta said.—Nick Wingfield
Tongue-Tied at Tesla
On Tesla’s earnings call, Elon Musk had much less to say on the topic of a potential merger of the EV maker with his other company, SpaceX, than he did Tuesday afternoon on SpaceX’s earnings call. He told investors he had said everything he wanted to on the topic for now. Tesla’s automotive revenues fell 3% to $20.56 billion while its total revenues remained flat at $28 billion compared to a year earlier. Tesla shares fell 3% in after-hours trading.—Nick Wingfield
In Other News
Electronic Arts plans to shed around 400 jobs, or about 3% of its staff, and shut down its BioWare studio, Bloomberg reported.
Microsoft asked shareholders to approve a plan to lower CEO Satya Nadella’s pay by 6% due to falling revenue in its Xbox unit. The company also nominated Stripe co-founder John Collison to its board.
Google will stop requiring local government officials to sign nondisclosure agreements when it plans to build data centers in their areas.
Today on The Information’s TITV
Check out today’s episode of TITV in which Akash Pasricha speaks with analyst Tim Farrar about SpaceX’s quarterly results.
Recommended Newsletter
Start your day with Applied AI, the newsletter from The Information that uncovers how leading businesses are leveraging AI to automate tasks across the board. Subscribe now for free to get it delivered straight to your inbox twice a week.
The Information's WTF Summit returns to Napa Valley, bringing together the women shaping tech, media and finance. Confirmed speakers include Bari Weiss, Whitney Wolfe Herd (Bumble), Kim Posnett (Goldman Sachs) and Whitney Cummings. Expect candid conversations and a room built for meaningful exchange. Limited tickets remain—reserve your spot now.
Trust is quickly becoming the defining question of the AI era. Over dinner, Jessica Lessin, Founder & Editor-in-Chief of The Information, and Whitt Butler, EY Americas Vice Chair – Consulting, will talk candidly about how leaders can build that trust and keep it. Request an invite.
Join The Information and NYSE for a cocktail reception and conversation on what it really takes to go public, hosted by The Information’s Dakin Campbell. The panel will explore the trust, relationships and preparation companies need to build with advisers, investors and exchange executives, along with the many complexities founders and executives must navigate along the way. Sponsored by NYSE. Apply to attend.
Login
