SpaceX has landed another massive compute deal, this time with Google. According to a regulatory filing published Friday, Google will pay SpaceX $920 million per month from October 2026 through June 2029 for access to around 110,000 Nvidia GPUs and supporting hardware. The total value of the agreement over its full term runs into the tens of billions.

The deal follows a similar arrangement SpaceX announced two weeks ago with Anthropic, which agreed to pay $1.25 billion per month through 2029 for the full output of the Colossus 1 data center in Memphis. Google's deal appears to cover roughly half that amount of compute, and SpaceX has not confirmed which specific data center Google will be drawing from. Elon Musk has previously indicated Colossus 2 would be reserved for xAI's own use.

Google's reason for the deal is notable. A company spokesperson described it as a short-term bridge arrangement to handle demand for Gemini Enterprise that came in higher than the company expected following its I/O launches in May. That is a meaningful admission from a company that is already considered one of the world's largest single owners of AI computing infrastructure, having committed more than $180 billion in capital expenditures this year alone.

Both sides have built in flexibility. Google and SpaceX can each terminate the agreement with 90 days notice after December 31, 2026, and Google's access to the data center will ramp up gradually through September at a reduced fee.

Meta has launched a new AI creator assistant on Facebook, and it is built specifically to replace the time creators currently spend digging through analytics dashboards trying to figure out what is working. Instead of checking multiple charts and reports, creators can now just ask the assistant directly and get a plain-language answer based on their own account data.

The assistant is conversational, so creators can ask things like when to post for maximum reach, what topics their audience is responding to, or how their follower base has shifted over the past few months. Follow-up questions work naturally, making it feel more like a conversation than a search. Content ideas are also part of the mix, with the assistant able to suggest trending audio, cultural moments worth tapping into, and formats that are performing well for similar creators.

Most creators already turn to external tools like ChatGPT to help with brainstorming and strategy. By building this directly into Facebook, Meta keeps that activity inside its own platform and makes itself harder to leave. Alongside the assistant, Meta also announced it is expanding its AI translation feature for Reels to five new languages including Arabic, French, and Vietnamese. The feature, which preserves a creator's original tone and optionally syncs translations to lip movements, is now being watched by over 500 million Facebook users weekly.

The creator assistant is rolling out now in the US, Canada, and India, with more countries to follow.

Apple's Worldwide Developers Conference kicks off Monday June 8, and the expectations this year are higher than they have been in a while. Simply out, Siri is getting a proper AI overhaul, and almost every Apple app is getting touched in the process.

The headline announcement will almost certainly be a completely rebuilt Siri. According to multiple reports and leaks, the new Siri will be powered by Google Gemini under the hood, making it capable of handling multi-step tasks, understanding context across apps, and holding more natural back-and-forth conversations. A standalone Siri app is also expected to launch, designed to go head to head with ChatGPT and Claude as a full AI chatbot experience. Leaked renders suggest conversations can be set to auto-delete after 30 days, a year, or kept indefinitely, a nod to privacy concerns that Apple has long positioned itself around.

Beyond Siri, Apple is expected to introduce an AI agent integration with the App Store, letting users delegate real-world tasks like booking reservations, managing smart home devices, and handling everyday admin through third-party agents. The Camera app is getting a dedicated Visual Intelligence mode sitting alongside Photo, Video, and Portrait, powered by Google Image Search for object recognition. Photos is getting smarter editing powered by Apple Intelligence, with natural language commands for things like removing objects or adjusting scenes.

The Image Playground app is also in line for an upgrade, with higher quality generation, better character consistency, and a simplified interface. The Wallet app is expected to gain a bill-splitting feature that reads receipts and sends payment requests automatically, plus a tool for creating digital passes from physical items like concert tickets or gym cards.

WWDC kicks off Monday at 10am PT and will be streamed live across Apple's website, the Apple Developer app, and YouTube.

Lovable, the vibe coding startup that crossed $400 million in annual revenue earlier this year with just 146 employees, has signed a new multi-year agreement with Google Cloud that will expand its footprint on the platform by five times. The dollar figure was not disclosed, but the scale of the expansion signals that Lovable's growth is showing no signs of slowing down.

As part of the deal, Lovable will gain expanded access to both Anthropic's Claude models and Google's own Gemini models, giving the platform more AI horsepower to handle the volume of projects its 8 million users are building. The Claude access piece is particularly interesting given Google's $40 billion investment in Anthropic, a portion of which was tied to Anthropic hitting certain performance targets. More Claude usage flowing through Lovable directly helps Anthropic hit those targets, creating a neat loop that benefits all three parties.

Beyond raw compute, the deal also plugs Lovable into Google's enterprise ecosystem in a meaningful way. Lovable's AI agent will be listed in the Gemini Enterprise Agent Gallery, Google Cloud's marketplace for enterprise AI tools, which simplifies procurement and billing for large companies wanting to use Lovable at scale. The platform will also integrate with Wiz, the cloud security company Google acquired for $32 billion earlier this year, allowing security issues in AI-generated code to be identified and fixed in real time.

For a startup that went from near zero to hundreds of millions in revenue in under two years, landing a deal of this size with Google is less a milestone and more a signal of where Lovable is heading next.

Video Rebirth bridges imagination and production. We empower creators with fluid generation — delivering visuals that are coherent, controllable, and instantly commercial-ready.