TOP TECH NEWS — FRIDAY, JUNE 5, 2026

The Week's Biggest Tech Stories: SpaceX's $75B IPO, Anthropic's Security Revolution & Apple's Next Chapter

From record-breaking IPOs to AI-powered cybersecurity, the last 48 hours in tech have been extraordinary.

🤖 Anthropic Raises $65B and Launches Project Glasswing — AI Now Hunts Zero-Day Vulnerabilities

BREAKING AI Artificial intelligence technology concept

In what analysts are calling a watershed moment for both AI investment and cybersecurity, Anthropic announced this week that it has closed a $65 billion funding round at a staggering $965 billion valuation — placing it within striking distance of joining the trillion-dollar club. The round, co-arranged by Apollo and Blackstone, underscores how the appetite for frontier AI companies remains insatiable even as the broader market faces headwinds.

Simultaneously, Anthropic unveiled Project Glasswing, a sweeping AI-powered cybersecurity initiative developed in partnership with AWS, Apple, Google, Microsoft, NVIDIA, and more than 40 additional organizations. The centrepiece is the company's Claude Mythos Preview model, which has already autonomously discovered thousands of high-severity software vulnerabilities — including decades-old zero-day flaws buried in every major operating system and browser. Anthropic is committing $100 million in usage credits and $4 million in open-source security donations to support the effort.

The implications are profound. For years, security researchers have relied on manual code audits and limited automated scanning tools. Project Glasswing represents a qualitative leap: an AI that can reason across millions of lines of legacy code, identifying attack surfaces that human teams would require months to uncover. Critics have noted the dual-use nature of such capabilities, but Anthropic has responded by making all discovered vulnerabilities immediately available to affected vendors through coordinated disclosure.

$65B

Funding Raised

$965B

Valuation

40+

Partner Organizations

$100M

Security Credits

"Claude Mythos Preview has already identified more critical vulnerabilities in its first operational weeks than most security teams discover in a year." — Anthropic, Project Glasswing launch statement, June 2026

🚀 SpaceX Files Record $75B IPO to Fund Space-Based AI Data Centers — Nasdaq Debut Set for June 12

BREAKING SPACE / AI Rocket launching into the sky

SpaceX has filed for what would be the largest technology IPO in history, targeting a $75 billion raise on the Nasdaq under the ticker symbol "SPCX," with trading expected to begin on June 12. The offering dwarfs previous tech IPOs and reflects both the stratospheric valuation of Elon Musk's space company — now valued in excess of $350 billion — and the convergence of aerospace and artificial intelligence infrastructure.

The IPO prospectus reveals the company's ambitious plan to deploy a constellation of AI data centers in low Earth orbit, leveraging the Starlink satellite network as a distributed computing backbone. The pitch to investors is compelling: by locating compute infrastructure in space, SpaceX argues it can sidestep terrestrial constraints on land, power, and cooling — three of the most pressing bottlenecks facing hyperscale data centre operators on the ground. The company has already struck preliminary agreements with several Fortune 500 firms interested in orbital compute access.

The announcement has sent ripples through both the aerospace and semiconductor sectors. Alphabet responded almost immediately with an $80 billion stock issuance of its own, widely interpreted as a defensive capital-raising move ahead of what many analysts expect to be an intensifying competition for next-generation AI infrastructure. Microsoft, meanwhile, disclosed internal plans for on-device AI agents this week, signalling a parallel push toward the edge that could complement SpaceX's orbital strategy.

$75B

IPO Target

SPCX

Nasdaq Ticker

Jun 12

Trading Begins

🍎 Apple Week: iPhone Ultra Breakthrough, Nvidia Chips for Siri & Vision's Scaled-Back Future

APPLE

Apple has had a remarkably eventful few days, with a trio of significant reports reshaping expectations ahead of WWDC 2026. The most striking concerns the iPhone Ultra: according to reliable supply chain sources, the forthcoming flagship will incorporate a display technology that has been more than 15 years in development — a micro-LED advancement that promises dramatically improved brightness, contrast, and power efficiency compared to the OLED panels used in current Pro models. The iPhone Ultra is expected to debut later this year as Apple's premium tier above the standard Pro lineup.

Separately, a detailed report from 9to5Mac revealed that Apple has struck a deal to use Nvidia chips to power its new Gemini-based Siri infrastructure. The move is notable on two fronts: it marks Apple's first significant reliance on Nvidia hardware for cloud AI services, and it confirms that the forthcoming AI-enhanced Siri will run on Google's Gemini models — a partnership that deepens the already billion-dollar relationship between the two companies. The chips will power the server-side inference that Siri offloads for complex requests.

Finally, Apple hardware chief John Ternus has reportedly scaled back the Vision product roadmap. According to internal sources, the reduced scope reflects both challenging production economics and tepid consumer adoption of the first-generation Vision Pro. Fewer new Vision SKUs are now expected over the next two years, with Apple redirecting engineering resources toward its core iPhone and Mac lines — which continue to perform strongly, with MacBook Neo doubling production after unexpectedly robust sales.

"iPhone Ultra might have tech innovation over 15 years in the making." — 9to5Mac, June 2, 2026

💻 GitHub Copilot Abandons Flat-Rate Pricing — Developers Now Pay Per Token

SOFTWARE AI TOOLS Code on a computer screen

Effective June 1, GitHub Copilot has transitioned from its familiar flat monthly subscription model to token-based billing — a move that affects millions of developers worldwide and signals a broader industry shift in how AI coding tools are monetised. Under the new structure, users are billed based on the actual volume of AI inference they consume, rather than paying a fixed rate regardless of usage.

The change has been met with mixed reactions in the developer community. Casual or occasional Copilot users may find themselves paying less than before, while power users — those relying heavily on code generation, extensive refactoring suggestions, and multi-file context analysis — could see bills increase substantially. GitHub has provided usage dashboards and budget caps to help teams manage costs, but the shift places the burden of monitoring on the user in a way the previous model did not.

The pricing pivot mirrors moves already made by several competitors, and it aligns with the economics of running large language models at scale: inference costs are genuinely variable, and providers argue that consumption-based billing is fairer and more sustainable long-term. For enterprise engineering teams evaluating AI assistant ROI, the new model offers clearer cost attribution per developer or project — a metric that budget owners have long requested. Expect other AI development tools to follow suit over the coming months.

Search This Blog

GADGETGLOW BYTES