Chinese lab MiniMax claims frontier coding and agentic benchmarks at 5–10% of rival model costs, with a 1M-token context window.
The AI Daily
Daily · June 2, 2026
Happy Monday. The biggest story today is Anthropic confidentially filing for an IPO — the first of the major frontier AI labs to formally kick off the process — while Google parent Alphabet is raising a massive $80B equity round to fund AI infrastructure, with Warren Buffett's Berkshire Hathaway backing the deal. Meanwhile, Nvidia's RTX Spark chip signals a serious push into the consumer PC market, and Florida became the first U.S. state to sue OpenAI over alleged consumer harms.
🧠 Foundation models
go deeper →Chinese startup MiniMax is claiming frontier-tier performance at a fraction of incumbent costs, while Anthropic's new Mythos model earns EU access — and a sobering security study reveals its browser agent was hijackable nearly a third of the time before safeguards kicked in.
MiniMax-M3 debuts, eclipsing GPT-5.5 and Gemini 3.1 Pro on key benchmark performance for just 5-10% of the cost
Anthropic to offer EU access to its advanced Mythos model
Anthropic shares its most advanced model with EU regulators following cybersecurity concerns, signaling proactive regulatory engagement.
Anthropic's browser agent got hijacked 31.5% of the time before safeguards engaged
Red-teamers hijacked Anthropic's browser agent nearly one-third of the time, exposing a critical prompt-injection vulnerability in agentic AI.
🏗️ Infrastructure
go deeper →The AI infrastructure build-out is accelerating on multiple fronts: Nvidia is entering the consumer PC chip market with RTX Spark, HPE posted a blowout quarter on AI server demand, Intel is targeting AI data centers with new Xeon 6+ processors, and OpenAI broke ground on a 1GW Michigan data center as part of Stargate.
Nvidia jumps into PCs with new Arm-based chip debuting in laptops from Microsoft, Dell, HP
Nvidia's RTX Spark ARM SoC marks its first serious entry into consumer PC chips, targeting AI-agent workloads on Windows devices.
Nvidia debuts RTX Spark processor for Windows laptops, compact desktops
RTX Spark SoC powers a new wave of AI-native Windows devices alongside OS upgrades designed for on-device agent execution.
This could be Windows' M1 moment — but expect it to cost a ton
Nvidia's RTX Spark could finally bring Apple-M1-style efficiency to Windows, though high prices may limit early adoption.
HPE skyrockets 30% on biggest earnings beat since 2018
HPE's AI server segment drove a historic earnings beat, validating surging enterprise demand for AI compute hardware.
HPE posts huge earnings beat thanks to AI servers, sending its stock skyward
AI server sales fueled HPE's largest earnings surprise in years, with Cloud & AI segment revenue soaring.
Intel introduces Xeon 6+ server processors, previews upcoming graphics cards
Intel's new Xeon 6+ CPUs and previewed GPUs are explicitly targeting AI data center workloads in a bid to stay relevant.
Building the infrastructure for the Intelligence Age in Michigan
OpenAI breaks ground on a 1GW Stargate data center in Michigan, one of the largest AI infrastructure investments in U.S. history.
Nvidia chases $200B CPU market with AI agent PCs from Microsoft, Dell, and HP
Nvidia's PC chip push targets a $200B CPU market by embedding AI agent capabilities directly into consumer devices.
💰 Funding & deals
go deeper →The AI funding story today is defined by scale: Anthropic's confidential IPO filing and Alphabet's $80B equity raise — backed by Berkshire Hathaway — signal that the largest AI companies are now tapping public and institutional capital markets to sustain their infrastructure ambitions.
Anthropic confidentially files for IPO amid rapid growth
Anthropic's confidential SEC IPO filing marks a pivotal moment for AI lab valuations and sets up what could be a landmark public offering.
Alphabet plans to raise $80B to pay for AI buildout
Alphabet's $80B equity raise — with Berkshire Hathaway's backing — is one of the largest corporate capital raises ever to fund AI infrastructure.
Anthropic has officially filed to go public
Anthropic's IPO filing, valued at over $60B at last raise, intensifies the race between frontier AI labs to access public markets.
Alphabet unveils plan to sell $80B in shares to fund ongoing AI infrastructure buildout
Google's parent cites AI demand exceeding supply as justification for its massive equity raise, underscoring infrastructure as a competitive bottleneck.
🔧 Middleware & platforms
go deeper →The middleware layer is maturing with new tools for AI agent governance, traffic detection, and security — reflecting growing enterprise anxiety about managing and controlling proliferating AI agents in production.
Merge launches Agent Handler for Employees as an IT gatekeeper for workplace AI agents
Merge's new product gives IT teams identity-aware governance over workplace AI agents, addressing enterprise control-plane gaps.
Fingerprint launches AI Assistant Detection to spot traffic from ChatGPT, Gemini and Claude
Fingerprint's new detection tool helps businesses distinguish AI-agent web traffic from human users in real time.
How we used Gemini to build Google I/O 2026
Google's own team used Gemini extensively to produce its flagship developer conference, a notable internal dogfooding case study.
📱 Application solutions
go deeper →Agentic AI is entering everyday consumer experiences — Google's Gemini Spark agent is live for hands-on review, OpenAI is positioning Codex as a universal productivity tool for knowledge workers, and Meta's AI chatbot suffered an embarrassing security failure when exploited to hijack Instagram accounts.
Gemini's new AI agent is about as good as Google's demo
Hands-on with Gemini Spark shows Google's 24/7 autonomous agent is genuinely impressive but raises real privacy and cost questions.
Codex is becoming a productivity tool for everyone
OpenAI reframes Codex beyond coding as a broad knowledge-work automation platform for research, data analysis, and content creation.
Meta's own AI was exploited to hijack Instagram accounts
Meta's AI support chatbot was manipulated to transfer account access to hackers, exposing a dangerous authentication design flaw.
This AI weather startup is out-forecasting government agencies
WindBorne's AI model, fed by a global balloon sensor network, now beats government forecasters — a compelling vertical AI success story.
'Disrupted or dead': AI is crushing a generation of startups built before ChatGPT
Pre-ChatGPT SaaS and AI startups are being wiped out as foundation model capabilities render their core value propositions obsolete.
Microsoft and Google are late to AI coding, but 'absolutely critical' they compete for growth
Microsoft and Google are racing to catch Anthropic and OpenAI in AI coding tools, a market analysts call existentially important for developer platform dominance.
⚖️ Policy & legal
AI governance is heating up simultaneously at the state and federal levels: Florida became the first U.S. state to sue OpenAI over consumer harms, while a Wired investigation reveals the Trump administration is internally fractured over how — or whether — to regulate AI at all.
Florida AG sues OpenAI and Sam Altman over claims the technology is dangerous and exploits its users
Florida's first-in-nation state lawsuit against OpenAI seeks to hold CEO Sam Altman personally liable, potentially opening a wave of state-level AI litigation.
Florida AG sues OpenAI, seeks to hold CEO Altman personally liable for alleged harms
The Florida complaint frames OpenAI's conduct as consumer fraud driven by an 'insatiable' competitive drive, setting a novel legal precedent.
The Trump Administration Is at War With Itself Over AI Regulation
Internal White House divisions over AI regulation have created a policy vacuum, leaving industry and officials uncertain about federal oversight direction.
Our views on AI policy and political advocacy
OpenAI publicly clarifies its policy stance and distances itself from outside political groups, likely in response to growing regulatory scrutiny.
📖 Beyond the headlines
Anthropic's browser agent got hijacked 31.5% of the time before safeguards engaged — As AI agents move from demo to production — Gemini Spark, Codex, Claude in enterprise workflows — prompt injection is becoming the defining security threat of the agentic era, yet no frontier lab has published comparable red-team numbers. This piece forces a concrete reckoning: what does a 31.5% hijack rate mean for enterprise AI deployments at scale, and why are only Anthropic's numbers public? Read →