In partnership with
In Today’s Issue:
🫧 The AI bubble narrative begins to… pop! Tools like Claude Code deliver undeniable productivity gains
🧬 A leading immunologist predicts AI could help cure all known diseases
📉 xAI’s massive GPU cluster struggles with an 11% utilization rate
🎮 GameStop launches a massive $56 billion unsolicited takeover bid
✨ And more AI goodness…
Dear Readers,
Two Chinese AI startups are about to force Wall Street's hand, and the numbers are staggering. Morgan Stanley predicts that Zhipu AI and MiniMax joining the Hang Seng Tech Index on June 8 will funnel up to $1.75 billion in passive capital into Chinese AI model stocks, marking the moment these companies graduate from speculative bets to institutional-grade investments.
But the capital story is just one thread in today's issue: we're also unpacking why the AI bubble narrative is finally starting to crack as tools like Claude Code turn developer hype into real productivity gains, how an immunologist believes AI could help cure all diseases within a decade, and why xAI's massive GPU cluster is only hitting 11% utilization, exposing a brutal truth about the gap between owning chips and actually using them well. On the corporate side, GameStop just made an unsolicited $56 billion bid for eBay, a move that could either turbocharge or gut one of the most ambitious AI-powered marketplace transformations in e-commerce, and meanwhile, software engineering job postings have sunk to their lowest level since November 2023.
Whether you're tracking where global AI capital is flowing, how science is being reimagined, or what happens when meme stock energy meets real infrastructure, today's issue has something that will make you rethink what you thought you knew.
All the best,
Kim Isenberg
🫧 AI Bubble Doubts Start Shifting
The AI boom is beginning to look less like speculative excess and more like a business reality, as Claude Code and rival agentic tools rapidly turn developer enthusiasm into measurable productivity gains. Anthropic’s extraordinary revenue growth, alongside surging demand for data centers, chips, and cloud capacity, suggests that AI companies are no longer merely selling a futuristic promise, they are selling tools that businesses are already struggling to get enough of.
The unresolved question is whether this breakthrough in coding can spread across law, finance, consulting, marketing, and other knowledge-work fields. If AI agents can reliably read, think, write, and verify across those domains, the current build-out may prove justified; if not, the industry could still discover that today’s infrastructure race has simply inflated a larger and more painful bubble.
🧬 AI Reimagines Disease Cure Timeline
Immunologist Derya Unutmaz believes AI could help cure all diseases within a decade, arguing that models can now accelerate everything from hypothesis-building to drug discovery and experimental design. This is more than just tech optimism: examples like AI explaining cancer-treatment mechanisms in minutes suggest a real shift in how science may be done.
Still, the prediction depends on major leaps in digital twins, clinical-trial reform, regulation, and safety oversight. Unutmaz sees AI as a collaborator rather than a replacement for scientists, but the piece wisely leaves room for tension around misuse, institutional resistance, and whether medicine can move as fast as the technology itself.
🧠 GPU Glut Meets Reality Check
Even with one of the world’s largest disclosed AI chip clusters, xAI’s reported 11% Model Flops Utilization shows how difficult it is to turn massive GPU ownership into efficient AI training. The piece points to a broader industry problem: expensive Nvidia chips often sit underused because training is bursty, memory bottlenecks slow computation, and networking across thousands of GPUs remains fragile.
Meanwhile there is a wider AI landscape defined by constraint and competition: SenseTime is chasing a low-cost “DeepSeek moment” with an efficient multimodal model, Musk’s OpenAI trial is unfolding under judicial warnings about social media, and AI funding remains intense across defense, finance, healthcare, agents, and coding tools. The underlying theme is clear: AI’s frontier is no longer just about who has the most chips, but who can use them smartly, cheaply, and consistently.
Sundar Pichai argues that AI is moving from a tool that answers questions to an always-available assistant that helps people make decisions, manage daily tasks, connect with others, and navigate the world, but only if companies build it responsibly with stronger guardrails.
Chinese AI Stocks Get $1.75B Boost
The Takeaway
👉 Zhipu AI and MiniMax will join the Hang Seng Tech Index on June 8, mechanically forcing passive funds to buy in and driving up to $1.75 billion in new capital into Chinese AI model stocks.
👉 Chinese AI model pricing is rising fast, jumping from 5% to 17% of U.S. model costs in just one year, signaling that the "cheap Chinese AI" narrative is fading as demand scales and companies push toward profitability.
👉 Morgan Stanley forecasts each leading Chinese AI model company will cross $1 billion in annual revenue this year, with that figure more than doubling in 2027, suggesting a commercialization inflection point is arriving.
👉 Hong Kong is rapidly positioning itself as the global listing venue for pure-play AI companies, with tech representing 40% of IPO fundraising year-to-date, potentially drawing future AI listings away from U.S. exchanges.
Two Chinese AI startups are about to trigger a Wall Street money wave. Morgan Stanley predicts that the inclusion of Zhipu AI and MiniMax into the Hang Seng Tech Index on June 8 will drive between $1.25 billion and $1.75 billion in passive investment inflows. That's a massive vote of confidence for a sector that barely existed as a public market category six months ago.
Both companies went public in Hong Kong in January, making them the world's first two pure-play AI model companies to IPO, while U.S. rivals OpenAI and Anthropic are still preparing their own listings. Since then, their stocks have skyrocketed. Morgan Stanley nearly doubled its price target for Zhipu's parent Knowledge Atlas to 990 HKD from 560 HKD, and raised MiniMax to 1,100 HKD from 990 HKD. Zhipu is known for strong coding models, while MiniMax powers the popular Hailuo AI video generator and the Talkie companion app.
Here's what makes this fascinating: the cost of accessing Chinese AI models has surged to at least 17% of U.S. model pricing, up from just 5% a year ago. The era of dirt-cheap Chinese AI is ending, and these companies are building real revenue. Morgan Stanley expects each frontier Chinese AI model maker to hit at least $1 billion in revenue this year, more than doubling next year.
Why it matters: The index inclusion marks the moment Chinese AI model companies graduate from speculative bets to institutional-grade investments. With tech accounting for 40% of Hong Kong's IPO fundraising this year, this could fundamentally reshape where global capital flows in the AI race.
Sources:
🔗 https://www.cnbc.com/2026/05/03/chinese-stocks-are-about-to-get-a-big-ai-boost-morgan-stanley-predicts.html?taid=69f73061563ff900012b0284
Build AI Product Skills - Without Pressing Pause on Your Career
Build real AI product skills through a flexible, modular MBA designed for working professionals. Progress from AI fundamentals to applied product strategy, with mentor support and hands-on projects.
Software engineering jobs hit their lowest posting since november 2023.
GameStop Bids $56B for eBay
Ryan Cohen just dropped one of the most unexpected moves in e-commerce history. GameStop, the meme stock darling worth roughly $12 billion, has made an unsolicited $56 billion offer to acquire eBay, a company nearly four times its size. The bid offers $125 per share in a 50/50 cash-and-stock split, representing a 20% premium to eBay's Friday close. Cohen says he wants to turn eBay into a serious Amazon competitor.
EBay has been going all-in on artificial intelligence, with over 10 million sellers using its generative AI tools to create more than 300 million listings. The platform generates roughly 500,000 AI-assisted listings every single day. eBay even partnered with OpenAI on a program giving small businesses free access to ChatGPT Enterprise. A GameStop takeover could either turbocharge these AI investments, or stall them entirely if cost-cutting takes priority. GameStop has pledged to cut $2 billion in annual costs within 12 months of closing.
This deal would test whether AI-powered marketplace tools can survive a massive corporate restructuring. The outcome could set a precedent for how acquirers treat AI infrastructure as an asset, not just a cost line.
AI Agents Are Reading Your Docs. Are You Ready?
Last month, 48% of visitors to documentation sites across Mintlify were AI agents, not humans.
Claude Code, Cursor, and other coding agents are becoming the actual customers reading your docs. And they read everything.
This changes what good documentation means. Humans skim and forgive gaps. Agents methodically check every endpoint, read every guide, and compare you against alternatives with zero fatigue.
Your docs aren't just helping users anymore. They're your product's first interview with the machines deciding whether to recommend you.
That means: clear schema markup so agents can parse your content, real benchmarks instead of marketing fluff, open endpoints agents can actually test, and honest comparisons that emphasize strengths without hype.
Mintlify powers documentation for over 20,000 companies, reaching 100M+ people every year. We just raised a $45M Series B led by @a16z and @SalesforceVC to build the knowledge layer for the agent era.
