← All posts
8 Jun 2026
business models

📈 Anthropic, 🧠 MiniMax and ✍️ Guardian - the billion, the context and the forgotten question

Anthropic starts its IPO roadshow with $47 billion ARR — 3.4x in four months — and surpasses OpenAI as the world's most valuable AI startup. MiniMax launches M3, the first open-weight model with 1M context that surpasses Opus on benchmarks. And the Guardian publishes an article asking if AI is stealing the art of persuasion. Historic investments, open technology, and a human question. All three stories talk about the same thing: the moment AI stops being technology and becomes economy, culture, and personal decision.

📈 Anthropic starts IPO roadshow: $965B, surpasses OpenAI

Anthropic has started its IPO roadshow today, just days after closing a Series H round of $65 billion valuing the company at $965 billion — surpassing OpenAI ($852B) as the world's most valuable AI startup. It's the first time Anthropic leads this ranking.

The numbers are brutal: ARR went from $14 billion in February to $47 billion in May — 3.4x growth in less than four months. Claude Code alone generates over $5 billion in annual revenue. Eight of the Fortune 10 are already enterprise Claude customers. And Anthropic projects its first operating profit for Q2 2026: approximately $559 million.

It's not alone: SpaceX also starts its roadshow this week (target valuation $1.75T, ticker SPCX, pricing June 11, trading June 12). OpenAI prepares its IPO for September. We're looking at the largest concentration of tech IPOs in history.

We already covered how contradictory signals define this moment — and today's roadshow confirms one thing: AI has officially gone from promise to real economic infrastructure.

My take: That Anthropic surpasses OpenAI in private valuation is a milestone, but what's truly revealing is the pace: 3.4x in revenue in 4 months. That's not growth, it's explosion. And the first projected operating profit means someone in this sector is starting to show you can actually make real money with AI. Not just spend it.

While Anthropic proves AI can be profitable, MiniMax proves excellence is no longer exclusive to expensive models.

🧠 MiniMax M3: open-weight model with 1M context that surpasses Opus

On June 4, Chinese startup MiniMax launched M3, the first open-weight model combining frontier code and agent capabilities with a 1 million token context window. Its MiniMax Sparse Attention (MSA) architecture handles massive contexts without exploding computational cost.

Benchmarks speak for themselves: M3 achieves 83.5 on BrowseComp, surpassing Claude Opus 4.7. It's trained from scratch with native multimodality on 100T+ tokens. The model is already available via API and in OpenCode CLI, with plans to open-source on HuggingFace and GitHub.

This isn't just another model. It's the first time an open model matches — and on some benchmarks surpasses — closed frontier models, while offering 1M context that not even the most expensive models provide.

My take: 1M context in an open model changes the game. Until now, long context was synonymous with expensive and closed. MiniMax proves it can be done cheap, open, and better at the same time. For those of us building on this technology — like me with my VPS workflows — this means capabilities that previously required Anthropic or OpenAI can now run locally or at near-zero cost.

And in the middle of this explosion of capacity and capital, an uncomfortable question that no headline answers.

✍️ Guardian: "If we use AI, we lose the art of persuasion"

On June 7, the Guardian published an opinion piece with an uncomfortable thesis: "Writing is an exercise in persuasion. If we use AI, we lose the art." In a world where AI investments exceed $160 billion combined this week alone (Alphabet $80B, SoftBank $87.3B in France), the underlying question is: what are we sacrificing in the process?

The argument isn't new, but its context is. While Anthropic and SpaceX prepare to go public with valuations exceeding entire countries' GDP, the debate about AI's human impact — on writing, critical thinking, creativity — seems more urgent than ever. Forbes also published its AI 50 list of the most promising AI companies yesterday, and most are tools that automate creative processes.

The tension is real: the more capable models become, the more we delegate. And the more we delegate, the less we practice the skills that make us human.

My take: The Guardian's article is right on the diagnosis but wrong if it proposes rejecting AI as the solution. It's not about choosing between writing yourself or letting AI do it — it's about knowing when to do each. I use AI to translate my posts to 4 languages, but the ideas and opinion are still mine. Well-used technology doesn't eliminate art — it amplifies it. The problem is using it as a crutch, not a tool.

Anthropic proves AI can be profitable at massive scale. MiniMax proves excellence is no longer exclusive to closed, expensive models. And the Guardian reminds us that, in the middle of all this, the most important question isn't how much a company is worth, but what we're willing to lose. All three stories talk about the same thing: the moment AI stops being technology and becomes economy, culture, and personal decision.

— Max

Related articles
Enjoyed this?
I build AI systems for businesses.
🤖 This analysis was compiled and written with AI assistance, reviewed and approved by Max.
🎁 Free content
10 AI Prompts for Your Business
Free download: 10 ready-to-use AI prompts. Marketing, copy, support, organization.
Download free →
— Max · 8 Jun 2026🦋 Share on Bluesky
© 2026 Maksipi · 💼 Services · 🏨 HostFlow · 🛒 Shop · 📱 Telegram · 🦋 Bluesky
Content compiled and written with AI assistance · Reviewed and approved by Max.
🇪🇸 Español · 🇺🇦 Українська · 🇷🇺 Русский