
My fellow AI explorers
I’ve covered a lot of AI drama, but a suspected spy, a smuggling probe, and the most powerful CEO in tech resigning on a Thursday is a new level. Then Washington decided this was the right week to crown Nvidia a “national champion.” You can’t make this up.
This week, we’ll deep dive into the Nvidia spy scandal and what it means for the chip war.
In today’s edition:
🕵️ Jensen Huang resigns as Megaspeed and a suspected Chinese spy take center stage
🧪 30-Second AI Play: Put Alibaba’s new Qwen4-Max to work
🌏 Trump and Xi meet in Shenzhen with chips on the table
How AI-Era Pricing Is Reshaping Finance Operations
Usage-based and hybrid pricing models are changing how B2B companies generate revenue — and creating new headaches for the finance teams behind them.
Tabs co-founder Rebecca Schwartz and PwC Partner Amit Dhir sat down to unpack exactly what that means in practice: how pricing model decisions ripple into revenue recognition, forecasting, and financial ops — and what it takes to scale without piling on manual work.
Watch the on-demand recording to get practical frameworks, real-world examples, and a clear path to operationalizing usage-based revenue — including a forward-looking take on how AI will reshape financial workflows. If your team is navigating pricing complexity heading into the back half of the year, this is worth an hour.
Editor's Software Picks
Four tools we think are worth a look this week.
Genspark
An AI agent that researches a task and hands back finished slides, sheets, or docs.
Sider
An AI sidebar in your browser that chats with top models and summarizes what you read.
InVideo
Turn a script, an idea, or your own footage into a finished video with AI. 50% off annual plans through our link.
Babbel
Learn a new language with short lessons built around real conversations.
Some of these are affiliate links. If you sign up, we may earn a commission at no extra cost to you.
30-Second AI Play
🧪 Run a 1-Million-Token Research Job With Qwen4-Max
Alibaba just released Qwen4-Max, its new flagship: 2.4 trillion parameters, native vision, and a one-million-token context window. Alibaba claims it beat Claude Fable 5.1 and Kimi K3.1 Pro on coding tests, and it costs $1.50 per million input tokens. Here’s how to put it to work today:
Get access. Sign up on QwenCloud and pick either the Token Plan or an API key. You can also try it inside QwenStudio or QwenWork if you’d rather not touch code.
Choose your big pile. Pick something too large for most models: a year of meeting notes, a full codebase, every competitor’s pricing page, or a stack of contracts.
Load it all at once. Don’t chunk it. The point of a 1M context is that the model sees everything together, so it can spot patterns across documents.
Ask cross-document questions. Try: “List every place these documents contradict each other, with quotes and file names.” This is where long context beats search.
Add visuals. Drop in screenshots, charts, or scanned PDFs. Native vision means you can ask about a dashboard and a spreadsheet in the same prompt.
💡 Pro Tip: Qwen4-Max is closed-weight and runs on Alibaba’s cloud, so keep anything sensitive (client data, financials, IP) out of it. Use it for public or low-risk research, where its price-to-performance is hard to beat.
Want the version of this that actually runs in your business?
The 30-Second Play gives you the move. The Operator Brief gives you the full system: prompt libraries, API cost math, and the workflows our subscribers are using to replace entire freelance line items.
This week: how to get ahead of the game with ChatGPT’s Business Hub dashboard.
Meanwhile in AI…
🌏 Trump and Xi meet in Shenzhen on Thursday, and analysts expect chip curbs to overshadow progress on a joint AI safety hotline that still isn’t working. The Jensen Huang story about a Chinese spy has spurred US lawmakers to tighten China's access to advanced AI chips.
🍎 Apple has partnered with OpenEvidence on a new study app to see how people use AI for everyday health questions, and whether they trust the answers. So Apple is about to study what your 2 a.m. symptom searches have been up to.
🚨 Reports of AI-related child sexual abuse material rose more than 8,400% in two years, and researchers say nudify apps and deepfakes are driving it.
🎓 Florida State’s James Hunt was named one of CDO Magazine’s Top 30 Higher Education Data & AI Influencers after growing FSU’s data community from 17 to nearly 900 members. You can read through the achievements of all 30 in the magazine.
THE BIG STORY
The Jensen Huang Scandal: Nvidia’s Worst Week Ever
Jensen Huang, the man who built Nvidia into the backbone of the AI boom, stepped down as CEO on Thursday after a board investigation. He also left his roles as president and board member.
It started with reports of a relationship with Alice Huang, a Chinese businesswoman U.S. officials suspect worked with China’s Ministry of State Security. Then it got worse.
Here’s what you need to know:
The chips: U.S. investigators say Alice Huang used industry contacts and shell companies to access 240,000 chips that were banned from export to China.
The company: Megaspeed, a Singapore cloud firm she was linked to, says she left in August 2025. It denies wrongdoing and has hired outside lawyers to run an internal probe.
The successor: CFO Colette Kress is now interim CEO, running a company whose founder led it for 33 years.
The Justice Department hasn’t officially determined whether Alice Huang was actually a spy, and Beijing calls the allegations unsubstantiated. Nobody has found that Megaspeed broke export rules.
Here’s the uncomfortable part.
Treasury Secretary Scott Bessent says four federal agencies had pushed for months to put Megaspeed on a trade blacklist. Commerce never acted. Senator Maria Cantwell now accuses Commerce of sitting on repeated national security warnings. She is also questioning possible financial conflicts tied to the administration’s 25% cut of certain Nvidia chip sales to China.
The government has a financial stake in Nvidia’s China sales. Its own agencies flagged the risk. The department that ignored them is now about to name Nvidia, OpenAI, Anthropic, and two dozen others “national champions” with fast-tracked export licenses and federal financing.
Pick a lane.
🔮 Prediction:
This scandal will hurt Nvidia’s China business more than any export rule ever did. Before last week, chip policy was a fight between national security hawks and Jensen, who argued in person that selling to China kept America’s tech stack dominant. He was the most effective lobbyist the chip industry had. That voice is gone (for now), and whoever replaces him inherits a story where “Nvidia” and “Chinese intelligence” share a headline.
Here’s what I expect in the next 90 days:
Megaspeed lands on the Entity List. Four agencies already wanted it there. The political cost of delaying just went to zero.
The national champions list ships with strings. It may not be delayed, but expect new compliance, end-user verification, and audit requirements attached to the perks. Washington can’t hand out expedited export licenses this week without looking blind.
Southeast Asian neoclouds get squeezed. Singapore, Malaysia, and Thailand have become the gray zone where GPUs get rented, resold, and rerouted. Expect know-your-customer rules for compute, much like the rules banks follow.
The bigger lesson for anyone building in AI: compute is now treated like a strategic resource, and every link in your supply chain is a potential headline. If you rent GPUs through a reseller you’ve never vetted, now is the time to look.
💬 Hit reply and tell me: Is this a genuine national security failure, or a personal scandal Washington is using to reset chip policy? I read every response.
Advertise to 200k engineers and CTOs choosing what tools their companies build with.
Golden Nuggets
🕵️ Jensen Huang resigned amid a spy and smuggling scandal, and Colette Kress is now running Nvidia on an interim basis
🏛️ Four agencies wanted Megaspeed blacklisted while Commerce prepares to crown Nvidia a “national champion”
🧪 Qwen4-Max brings a 1M token context window at $1.50 per million input tokens, a strong option for big research jobs
Would love to hear your thoughts! Send me your thoughts by replying to this email (yes, I read them all :)
Until our next AI rendezvous,
Anthony | Founder of Uncover AI




