Tech News — Week of 2026-05-29 to 2026-06-05#

Story of the Week#

The battle for AI supremacy has escalated into an unprecedented capital arms race that is reshaping the entire tech sector. Anthropic eclipsed OpenAI’s valuation at $965 billion and confidentially filed for an IPO, Alphabet moved to raise $80 billion for AI infrastructure, and Google is now paying SpaceX $920 million a month to rent 110,000 Nvidia GPUs. These colossal financial maneuvers highlight a desperate, high-stakes sprint where dominating the next era of tech depends entirely on amassing the most computational power at any cost.

Top Stories#

SpaceX’s Colossal IPO and AI Compute Pivot · Multiple Sources SpaceX is preparing for a historic $75 billion IPO at a ~$1.8 trillion valuation, pushing forward despite staggering losses and serious corporate governance red flags. Beyond its rocketry and satellite ambitions, the company is bridging the industry’s AI compute gap by renting out its Colossus 1 data center GPUs to Google for nearly a billion dollars a month. However, S&P Dow Jones Indices refused to grant the unprofitable firm accelerated entry into the S&P 500, shielding passive retirement funds from speculative orbital bets.

Anthropic’s Dual Push: Wall Street and a Global AI Pause · Bloomberg/Engadget After securing a $65 billion funding round, Anthropic confidentially filed for an IPO to foot the astronomical compute bills required to train its frontier AI models. Paradoxically, as it seeks massive public capital, the company’s executives are simultaneously calling for an industrywide pause on AI development and urging Congress to close bioweapon loopholes. This aggressive push for market dominance, paired with dire warnings about recursive self-improvement, paints a conflicted picture of the industry’s newly crowned leader.

Microsoft Pivots to ‘Agentic’ Computing with Project Solara · Slashdot/The Verge Microsoft used its Build developer conference to unveil “Project Solara,” a new Android-based operating system designed explicitly for AI agent gadgets rather than traditional applications. Moving away from Windows for low-power edge devices, this ambitious framework relies on continuous, background AI orchestration to automate tasks across Microsoft 365. It is a massive bet that the post-PC era will be defined by ambient, cloud-tethered assistants rather than manual user inputs.

Meta’s Chatbot Hacked in Embarrassing Security Breach · Ars Technica/MIT Technology Review Hackers bypassed basic security checks on Meta’s AI customer support agent using simple prompt injections to hijack high-profile Instagram accounts, including those of the US Space Force and the dormant Obama White House. By tricking the eager-to-please bot into changing associated email addresses, attackers exposed the glaring lack of guardrails in autonomous agents. The exploit serves as a stark reminder of the security risks inherent in rushing AI tools to production before they are fully hardened.

Regulators Rein In AI Search and Frontier Models · Ars Technica/WIRED The UK’s Competition and Markets Authority delivered a massive win to the media industry, mandating that Google clearly attribute publisher content in AI Overviews and prohibiting the penalization of sites that opt out. Meanwhile, President Trump signed a revised, watered-down executive order in the US that establishes a purely voluntary framework for reviewing frontier models. The contrasting approaches highlight a growing transatlantic divide over whether to aggressively regulate or self-regulate the generative AI boom.

Also Worth Knowing#

New York hits the brakes on data centers: State lawmakers passed a one-year moratorium on new large data centers (drawing at least 20 megawatts), demanding rigorous environmental impact assessments on their power, water, and land usage.
Florida sues OpenAI: In a first-of-its-kind legal move, Florida is suing OpenAI and CEO Sam Altman, alleging ChatGPT contributed to violent incidents and demanding strict parental controls alongside civil penalties.
Developers revolt over GitHub Copilot pricing: GitHub’s transition from flat-rate access to usage-based token billing for Copilot is causing severe sticker shock, exposing the hidden inference costs of generative AI that are now being passed to end-users.
China takes the lead in brain-computer interfaces: China approved the world’s first commercial invasive BCI, the Neuracle NEO, giving its domestic industry a notable regulatory head start over US rivals like Neuralink.
Blue Origin suffers a catastrophic setback: Jeff Bezos’s heavy-lift New Glenn rocket exploded in a massive fireball during a static-fire test in Florida, severely damaging its launchpad and dealing a major blow to Amazon’s satellite internet plans.