The next big thing for Nvidia will be CPUs for AI agents, $200 billion worth, CEO Jensen Huang predicts.

Politico: Sources: the Pentagon is launching a task force to study how to safely deploy leading AI tools with hacking capabilities across Cyber Command and NSA missions  —  These people, like others in this report, were granted anonymity because they were not authorized to speak publicly about the sensitive effort.

Anthropic has told its investors that it will more than double revenue to around $10.9 billion in its second quarter.

Build a live Greeks dashboard and ask Claude what to do before you log off.

Grok Imagine Video 1.5 is an exciting video model release from xAI. Here's everything you can do with it.

Elon Musk faced backlash after his NSFW AI, Grok, generated sexualized AI images of women, including minors. SpaceX listed those possible generation of such images as a risk factor in its S-1. Bloomberg/Getty Images SpaceX warned investors about the risk that Grok could generate "potential nonconsensual or exploitative imagery." The company's S-1 paperwork cited investigations into Grok "content representing children in sexualized contexts." Elon Musk previously pushed back on claims that Grok created sexualized AI images of minors after public backlash. Elon Musk bet big on "spicy" AI. Now, it's officially a risk factor. SpaceX acquired xAI three months before it filed its pre-IPO S-1 paperwork. That brought in-house xAI's social media platform, consumer chatbot — and not-safe-for-work Grok AI features. In the Wednesday filing, SpaceX said that such features could pose "heightened risks" and "reputational harm" because the Grok NSFW modes are "more irreverent and harsher than our standard offerings." The filing warned about the possible "generation of potentially explicit content," as well as "potential nonconsensual or exploitative imagery" and content that infringes on intellectual property. Such content could also be viewed as "harmful, harassing, abusive, or discriminatory," the S-1 read. It's standard practice for companies to flag potential risks to its business in its pre-IPO paperwork, including any ongoing investigations or lawsuits that could have a material impact. In January, xAI faced backlash after Grok created non-consensual sexualized AI images of women, including minors. The images sparked public condemnations from government officials, policy changes, and lawsuits. SpaceX referenced these lawsuits in the S-1 filing. The company is subject to "investigations and inquiries" in the US concerning "allegations that our AI products were used to create nonconsensual explicit images or content representing children in sexualized contexts, and similar matters," it read. "The Company and certain subsidiaries have been named as defendants in multiple lawsuits arising from Grok's image-generation and editing features," the S-1 reads. SpaceX said it intends to "defend itself vigorously in these actions." Elon Musk has previously pushed back on claims that Grok generated sexualized AI images of minors, stating in January that he was "not aware of any naked underage images generated by Grok." The company made changes to its Grok AI following the backlash, limiting image generation to paying customers. The company also warned that it may be subject to "additional litigation in the future" surrounding such topics. The company also framed Grok as "among the fastest-advancing frontier models relative to peers, including OpenAI, Anthropic, and Google." The SpaceX filing also referenced the Irish Data Protection Commission's inquiry into the AI company. "This inquiry involves the processing of personal data of European Union data subjects, including children, using generative AI functionality associated with the Grok model within the X platform," the filing read. "Ani", the chatbot from XAI's Grok app that is capable of NSFW interactions. illustration by Cheng Xin/Getty Images XAI also produces a variety of AI companions, like its AI anime character, Ani. In the filing, SpaceX warned about an inquiry from the Federal Trade Commission about how AI companies "have evaluated the safety of their chatbots when acting as companions to children and teens." While SpaceX had doubled down on Grok's "irreverent" AI offerings, its competitors have shied away. OpenAI CEO Sam Altman announced a push into "erotica for adults" in October; the plans are on hold. Read the original article on Business Insider

Tim De Chant / TechCrunch: SpaceX S-1: xAI plans to buy another $2.8B worth of turbines for its data centers, including a $2B deal for mobile gas turbines, the type it's being sued over  —  Elon Musk's xAI has gotten itself in hot water over its use of polluting generators at its data center near Memphis, Tennessee.

The investment comes as Elon Musk’s AI unit faces complaints about the carbon-emitting units and looks to become a big player in cloud computing.

Rebecca Bellan / TechCrunch: SpaceX S-1: xAI had a $6.4B operating loss on $3.2B in revenue in 2025; Grok and X had 550M MAUs combined as of March 2026, and 117M used Grok's AI features  —  Elon Musk's xAI lost $6.4 billion from operations on just $3.2 billion in revenue in 2025, according to SpaceX's IPO filings.

SpaceX has finally made the contents of its IPO filing public, weeks ahead of what is expected to be the largest IPO ever and one that will make Musk the CEO, CTO, and chairman of the board.

"We believe we have identified the largest TAM in human history."

NurPhoto/Getty Images We’ve heard a lot about the artificial intelligence (AI) boom and how enormous amounts of money are being poured into companies building ever more powerful technologies. That boom is now taking a new turn as major AI players edge closer to becoming publicly-traded companies. According to reports, OpenAI is preparing to file confidentially for a public listing that could value the ChatGPT maker at hundreds of billions of dollars. Rivals including Anthropic (Claude) and Elon Musk’s SpaceX – which just absorbed xAI (Grok) – are also moving toward the stock market. What many people may not realise, however, is that, through retirement funds, pensions and other managed investments, they could end up owning shares in these giants – whether they choose to or not. And while people might have moral concerns with the AI companies they’re tied to, the greater issue at play is about money, risk and who ends up holding it. Where AI’s billions go Building a cutting-edge AI system requires vast numbers of specialised computer chips, running nonstop in data centres that consume enough electricity to power a small city. OpenAI plans to spend around US$50 billion on computing power in 2026 alone. In 2017, that same company spent roughly US$30 million, a 1,600-fold increase in less than a decade. OpenAI is targeting roughly US$600 billion in compute spending – or that in areas such as processing power, data storage and cloud infrastructure – through to 2030. It’s not just OpenAI. The big technology companies are collectively expected to invest around US$650 billion in AI infrastructure in 2026. That’s roughly one-third of Australia’s annual GDP – or two-and-a-half times New Zealand’s – being committed to one technology bet in a single year. All these expenses must be covered before the companies earn consistent profits. This is why they keep raising money – and why the question of where that money will eventually come from is enormously important. In 2025, total investment in AI companies reached US$217 billion. Then, in just the first three months of 2026, private AI companies raised a further US$226 billion, surpassing the entire 2025 total in a single quarter. Much of this was concentrated in three transactions: US$122 billion for OpenAI, US$30 billion for Anthropic and US$7.5 billion for xAI. Together, these three deals alone accounted for 71% of all AI funding that quarter. Mega-rounds above US$100 million now make up 94% of all AI investment by value. The funding has mostly come from large institutions: venture capital firms, sovereign wealth funds, and technology giants that can afford to take the risk. The gains and the losses stay within a small, specialist group. When the AI sector goes public Once a company is publicly listed, anyone can buy its shares. More importantly, large index-tracking funds – widely used in Australia’s Super system and New Zealand’s KiwiSaver funds – automatically gain exposure to companies once they become large enough to enter the indices those funds follow. In other words, they don’t get to decide whether it looks like a good investment; the index simply decides for them. That matters for ethical investors, too. The thought of AI raises concerns about privacy, labour, misinformation and security. But unlike tobacco or gambling, it may prove difficult to exclude because it is being increasingly woven into the world’s largest listed companies. There is, therefore, a case for asking whether these companies should trigger an opt-out mechanism for fund managers and regulators before the listings arrive. Neither OpenAI, Anthropic, nor xAI has formally announced a stock market listing, and timelines remain uncertain. When that shift comes, the risk also shifts. The investors who funded the early stages of this race knew what they were getting into. The people who will end up holding the shares through their pension funds or index trackers may not. Index providers are rewriting the rules Here is where the picture becomes more complex. Major index providers are changing their rules so newly listed mega-cap AI companies can enter key benchmarks much faster. Nasdaq has already adopted a fast-track rule that allows a newly listed mega-cap company to join the Nasdaq-100 after just 15 trading days. S&P Dow Jones Indices is consulting on similar changes that would reduce the waiting period and waive profitability requirements for mega-caps. These changes are reshaping the index system to funnel passive money into AI giants almost as soon as they list – and before most investors have had time to decide whether they belong in their portfolios at all. So, what can ordinary investors do? As they likely won’t be making the call themselves on whether to invest in an AI stock market float, they can put questions to the fund managers doing so on their behalf. Those might be questions about whether the company is becoming more efficient, what their customer retention looks like, or how their leadership holds up under pressure. OpenAI’s 2023 board crisis showed how unusual governance structures can create sudden instability. There is no doubt the AI revolution is real and is changing economies in the same way it is changing our everyday lives. But whether the AI boom will create lasting value for ordinary investors – or mainly provide an exit for early-stage insiders – is a question fund managers and regulators cannot afford to leave unanswered. Before the listings arrive, they need to decide: should ordinary investors be automatically swept into the AI gamble, or should they have a choice? Sara Ali does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

SpaceX filed its S-1 on Wednesday, offering a glimpse into its finances. Bloomberg/Getty Images SpaceX filed its S-1 on Wednesday, a required step ahead of an IPO. The document offers a glimpse into the company's financials. Here are the biggest takeaways from the filing. Elon Musk's SpaceX officially kicked off its march to Wall Street on Wednesday, filing the S-1 paperwork required for its hotly anticipated IPO. Although the S-1 didn't reveal key figures investors are waiting on — such as SpaceX's estimated stock value and anticipated public share price — it did offer the first close-up look at the financial details and strategies behind Musk's rocket company and AI business, xAI. Here are the biggest takeaways from the filing. There's an incentive to build a 'human colony' Making life "multiplanetary" is at the forefront of SpaceX's mission, and Musk has a massive financial incentive to make it so: 1 billion shares of SpaceX. The filing shows that, among market-cap milestones, SpaceX will have to establish a "permanent human colony on Mars with at least one million inhabitants" for Musk to access the package. Musk will also still have to be employed by SpaceX at the time of this milestone achievement, and the shares would be split into 15 tranches. Still, it's a highly unusual compensation trigger for an otherwise not normal IPO. Musk's pay package could be worth up to $737B Musk could get a hefty payday if he builds interplanetary cities and orbital data centers. Based on the implied share count in the filing, Musk's two performance awards could be worth roughly $737 billion if fully vested. That includes about $583 billion from the one-billion share award that vests if SpaceX reaches a $7.5 trillion market cap and establishes a permanent Mars colony. It also includes about $154 billion from a separate award of about 302 million shares tied to orbital data centers that deliver 100 terawatts of compute a year, along with a SpaceX market-cap milestone of about $6.6 trillion. Anthropic is paying a lot for SpaceX's compute Anthropic is paying SpaceX $1.25 billion a month through May 2029 for access to the rocket company's compute capacity, the filing showed. That's about $15 billion in revenue coming from Anthropic. The compute refers to SpaceX's Colossus and Colossus II, massive supercomputers and data centers designed for training AI. The filing said the agreement for cloud and compute services begins in May 2026. Musk has massive voting power Voting rights in stocks mean each share owned gives the shareholder one vote on eligible matters. The S-1 revealed that Musk, who serves as the company's chief executive, Chief Technical Officer, and Chairman of the board, holds 85.1% of the combined voting power on company matters, based on his 5,569,053,075 shares. A screenshot of the SpaceX S-1 filing shows how much voting power Elon Musk has by virtue of the number of shares he holds. Cadie Thompson Grok's CapEx was higher than Starship's "For the three months ended March 31, 2026, capital expenditures for our Space segment was $1,052 million, for our Connectivity segment was $1,332 million and for our AI segment was $7,723 million," the S-1 reads. That means spending on xAI's projects, like Grok, was higher than SpaceX's rocket division's spending, which includes its Super Heavy and Starship rockets. The numbers were starker in 2025, with the Space segment spending $3,832 million and the AI segment spending $12,727 million, according to the filing. Read the original article on Business Insider

The video clipping startup raised a $7 million seed round led by Slow Ventures.

<blockquote cite="https://www.sec.gov/Archives/edgar/data/1181412/000162828026036936/spaceexplorationtechnologi.htm"><p>We have the ability to use compute resources to support our proprietary AI applications (such as Grok 5, which is currently being trained at COLOSSUS II), while also providing access to select compute capacity to third-party customers. For example, in May 2026, we entered into <strong>Cloud Services Agreements with Anthropic PBC</strong> (“Anthropic”), an AI research and development public benefit corporation, with respect to access to <strong>compute capacity across COLOSSUS and COLOSSUS II</strong>. Pursuant to these agreements, the customer <strong>has agreed to pay us $1.25 billion per month</strong> through May 2029, with capacity ramping in May and June 2026 at a reduced fee. The agreements may be terminated by either party upon 90 days’ notice.</p></blockquote> <p class="cite">&mdash; <a href="https://www.sec.gov/Archives/edgar/data/1181412/000162828026036936/spaceexplorationtechnologi.htm">SpaceX S-1</a>, highlights mine</p> <p>Tags: <a href="https://simonwillison.net/tags/anthropic">anthropic</a>, <a href="https://simonwillison.net/tags/grok">grok</a>, <a href="https://simonwillison.net/tags/generative-ai">generative-ai</a>, <a href="https://simonwillison.net/tags/ai">ai</a>, <a href="https://simonwillison.net/tags/llms">llms</a></p>

SpaceX's IPO filing reveals xAI lost $6.4 billion in 2025 while planning a massive Grok expansion — offering the first public look at Elon Musk's AI financials and more details about his ambitions.

Nvidia announced another record revenue figure after market close on Wednesday, but forecasted that revenue growth would slow in the following quarter.

Discussion | Link

Officials denied refusing entry, but dodged questions on why Americans didn't return.

Elon Musk's xAI said it will buy $2.8 billion worth of natural gas turbines over the next three years, according to SpaceX's IPO filing.