AI News: June 9, 2026

1. Moonshot AI Seeks $30 Billion Valuation in Third Round in Six Months

Moonshot AI. The Beijing-based developer of the Kimi chatbot is in early talks to raise more than $1 billion, and as much as $2 billion, at a $30 billion valuation, according to a Bloomberg report. The target would mark a roughly seven-fold jump from the just-over-$4 billion valuation it held in December and arrives even as Moonshot is closing a separate Meituan-led round that valued it at $20 billion. The company says its annual recurring revenue topped $200 million in April, and it is unwinding its offshore structure ahead of a possible Hong Kong IPO, underscoring how aggressively capital is flowing to the leading Chinese labs challenging OpenAI and Anthropic. Source

2. PhysicsX Raises $300 Million at a $2.4 Billion Valuation for Physics AI

PhysicsX. The London-based engineering startup closed a $300 million Series C led by Singapore’s Temasek, more than doubling its valuation to $2.4 billion less than a year after a Series B priced it just under $1 billion. New backers M&G Investments and Intrepid Growth Partners joined, while existing investors including NVIDIA, Applied Materials, Atomico, General Catalyst, and Siemens increased their stakes in the oversubscribed round. PhysicsX builds AI models that replace conventional physics simulations, which can take hours or days, with results delivered in seconds, and says it has doubled recognized revenue and headcount over the past year. Source

3. Google and Nvidia Explore Intel as a TSMC Backup for AI Chips

Intel. Google and NVIDIA are evaluating Intel’s foundry as an alternative to TSMC for manufacturing AI chips as TSMC’s capacity strains under AI demand. Google has reportedly ordered more than three million of its TPUs for 2028, while NVIDIA is testing Intel’s manufacturing for its Feynman GPU architecture without yet committing to an order, and SK Hynix is assessing compatibility with Intel’s packaging technology. The interest, which sent Intel stock up more than 10 percent, follows TSMC’s own warning that global chip supply will not meet AI-driven demand for years and signals a potential turnaround for Intel’s struggling foundry business. Source

4. Mercor CEO Accuses Sequoia of Inflating Startup Valuations Through Dual Pricing

Mercor. Brendan Foody, co-founder of the $10 billion AI talent platform Mercor, publicly accused Sequoia of using a “dual-pricing” tactic that inflates headline valuations, claiming he had seen roughly half a dozen rounds structured in two tranches. He pointed to Serval, announced at a $1 billion valuation days after a Series A extension valued it under $400 million, and Aaru, where lead investor Redpoint valued the company at $450 million against a $1 billion headline. Sequoia partner Shaun Maguire pushed back, framing the structure as a market response to competition for hot AI companies, in a dispute that highlights how opaque valuation mechanics are distorting signals in the AI funding market. Source

5. Instagram AI Support Chatbot Breach May Have Exposed Over 20,000 Accounts

Meta. Meta disclosed that a flaw in its AI-driven “High Touch Support” account recovery tool may have allowed attackers to hijack up to 20,225 Instagram accounts over roughly seven weeks. The system sent password reset links to any email address without verifying ownership, and exploitation that began around April 17 went undetected until May 31, potentially exposing contact details, birth dates, posts, direct messages, and profile information. Meta has disabled the chatbot, removed the faulty code, forced affected users into password resets, and says it will fix email verification and audit similar systems across its platforms before reactivating the tool. Source

6. Sam Altman’s Eye-Scanning Venture Cuts Staff as It Struggles for Revenue

Tools for Humanity. Tools for Humanity, the company behind Sam Altman’s World identity-verification project and its Worldcoin cryptocurrency, is conducting layoffs as it struggles to generate revenue, according to a report. The cuts land the same day OpenAI filed confidentially for an IPO and follow regulatory setbacks for the iris-scanning effort, including a ban in Kenya and an $830,000 fine in South Korea over alleged privacy violations. The venture, previously backed at a $2.5 billion valuation, has yet to find a sustainable business around its orb-based human-verification hardware. Source

7. KPMG Survey Finds Most Companies Lack Visibility Into AI Spending

KPMG. A KPMG survey found that only 26 percent of companies have complete transparency into their AI spending, while 50 percent have only partial oversight and 22 percent have no visibility until the bills arrive. The shift to token-based billing is creating budgeting shocks, with one analyst warning that “a lot of CFOs are going to see their Anthropic bill and freak out this quarter” and KPMG documenting clients whose token usage rose sixfold and exhausted annual budgets within months. Analysts compare the dynamic to the pandemic-era cloud boom, suggesting rapid AI spending could be followed by sharp cutbacks unless organizations build tracking infrastructure. Source

8. Microsoft Tightens Cloud Rules for Conflict Zones After Azure Military Probe

Microsoft. Microsoft moved to tighten its rules governing cloud and AI use in conflict zones following an external investigation into the Israeli military’s use of Azure. The change reflects mounting scrutiny of how hyperscalers’ infrastructure is deployed in active conflicts and the governance obligations that attach to providing compute and AI services to military customers. The episode adds to a broader debate over the accountability of cloud providers whose platforms underpin sensitive government and defense workloads. Source