THE WEEKLY BRIEF

Anthropic’s Trillion-Dollar Week, Meta’s AI Breach, and the Ring That’s Too Small to See

Networkcraft Desk · The Weekly Brief · June 2, 2026

Five stories defined this week in tech. An AI lab crossed the trillion-dollar threshold and filed for an IPO. Meta’s own chatbot was used as a weapon against its users. A smart ring shrunk to jewellery proportions. The VC market officially split in two. And the biggest data breach report of the year confirmed what every professional suspected: the attackers are winning the speed race.

01 — Anthropic Files for IPO at $965B, Overtaking OpenAI

Anthropic closed a $65 billion Series H round at a $965 billion post-money valuation this week, vaulting past OpenAI ($852B) to become the most valuable private AI startup in the world. On June 1, the company confidentially filed for an IPO — joining OpenAI and SpaceX in what analysts are calling a once-in-a-generation public offering wave. Read Maya Chen’s full breakdown on Networkcraft.

Anthropic’s run-rate revenue crossed $47 billion in May, driven by Claude Code adoption — the agentic coding tool that lets developers ship software by describing what they want in plain English. Amazon committed $25 billion to Anthropic in April, in exchange for over $100 billion in AWS spend across the next decade.

02 — Meta’s AI Chatbot Used as a Weapon Against Its Own Users

Hackers exploited Meta’s AI-powered support chatbot to take over high-profile Instagram accounts — including the Obama White House archive and Sephora — simply by asking the bot to change the recovery email address. No password required. No human review. The AI approved the request, sent a verification code to the attacker’s email, and handed over the keys. Read Sara Voss’s full analysis on Networkcraft.

The breach, which Meta patched on May 29 and confirmed on June 1, exposes a structural flaw in AI-powered customer support: bots optimized for speed over security become attack surfaces. Meanwhile, the Verizon DBIR confirmed that software vulnerability exploitation now accounts for 31% of all breaches — overtaking credential theft for the first time.

03 — Oura Ring 5: The Smallest Smart Ring Ever Ships June 4

Oura launched the Ring 5 this week — 40% smaller than its predecessor at 6.09mm wide and just 2 grams. It adds GLP-1 metabolic insights, blood pressure trend monitoring, and an AI health escalation pipeline that can route users from symptom alerts to licensed physicians. Starting price: $399. Read James Okafor’s full review on Networkcraft.

The smart ring market shipped 4 million units in 2025, doubling for the second consecutive year. Oura has sold 5.5 million rings since founding and is targeting an IPO at an $11 billion valuation. Samsung has not announced a Galaxy Ring 2 for 2026, leaving Oura uncontested on the hardware front.

04 — The VC Market Splits: AI Takes 81% of Q1 Funding

PitchBook data confirmed this week that AI companies captured 81% of global venture funding in Q1 2026 — more than $240 billion. OpenAI’s $122 billion March round alone was 46% of all US deal value. Meanwhile, 220+ former unicorns — including Glossier, Savage X Fenty, and The Farmer’s Dog — are now classified as ‘fallen’ by PitchBook. Read Alex Rivera’s full analysis on Networkcraft.

Startups that last raised in 2021 are worth 68% less on average. The pattern is clear: AI-native companies are absorbing virtually all available venture capital, and the pre-ChatGPT cohort is being cut off from funding. The trillion-dollar IPO test — whether public markets can absorb Anthropic, OpenAI, and SpaceX in a single year — will define the next chapter.

05 — Verizon DBIR 2026: Software Exploits Overtake Credential Theft

Verizon’s 2026 Data Breach Investigations Report, released this week, contains a sobering finding: software vulnerability exploitation now accounts for 31% of all breaches — overtaking credential theft as the top attack vector for the first time. Ransomware represents 48% of all breaches. Only 26% of critical vulnerabilities were fully remediated in 2025.

IBM’s latest data puts the average US data breach cost at $10.22 million. Munich Re projects global cybercrime costs hitting $14 trillion by 2028 — exceeding the combined output of Germany, Japan, and India. The gap between vulnerability discovery and remediation is widening, and AI-powered attackers are exploiting it at machine speed.