Quick answer
Jensen Huang resigned as Nvidia’s CEO, president and board member on 11 November 2026, and finance chief Colette Kress is now interim CEO. In a new video, Nate B Jones argues the real story is not the announcement but Nvidia’s next big deal: who gets scarce chips, who backs the financing and how Washington treats chip exports, now that the person who held all of that together has gone. His conclusion is cautiously hopeful for buyers. If rival chips get easier to use and Nvidia has to negotiate harder, more competition could slowly push down what we all pay for AI. Nothing changes overnight, but it is a good moment to avoid long, rigid AI contracts and keep your options open.
Key takeaways
- The facts are confirmed. Nvidia’s own announcement and SEC filing confirm the resignation, Kress as interim CEO and Harvey Jones as the first board chair.
- Nate’s focus is deal-making. Jensen was the link between chips, customers, financing and politics. Replacing that takes time.
- The CUDA moat is narrowing, not gone. Anthropic says AI helped bring part of a Claude model up on AMD chips in about three days, but that claim is narrow.
- Nvidia is deep in AI financing. It has guarantees capped at $105 billion on an Ohio data centre leased to OpenAI.
- For your bill, think gradual. More chip competition may mean cheaper AI over time, so keep contracts flexible rather than betting on a date.
🛠️ Practical guides
Ready to put this into practice? These step-by-step guides answer two common follow-up questions:
- Annual or Monthly AI Subscription: How to Choose When Prices May Fall — when an annual plan saves money and when staying monthly is the safer bet.
- Questions to Ask Your AI Tool Supplier Before You Renew — the questions that protect you on price, data and leaving, plus an email you can copy.
In this guide
- Watch: the source video
- What actually happened at Nvidia?
- Why does Nate think the next Nvidia deal matters most?
- Is the CUDA moat starting to shrink?
- What happens if AI companies split their chip orders?
- How is Nvidia tied up in AI financing?
- What about chip export rules?
- Will the Jensen Huang resignation make AI cheaper?
- What should you do about it now?
- Frequently asked questions
- The bottom line
- Related posts
Big company news often feels far away from a small business. A chip maker’s boardroom does not obviously affect the price of your ChatGPT, Claude or Gemini plan. But almost every AI tool you use runs on chips, and for the last few years most of those chips have come from one company.
That is why the Jensen Huang resignation is worth ten minutes of your time. In this article we sum up Nate B Jones’s argument, check his main facts against public sources, flag what is still uncertain and turn it into a few sensible steps. We covered the separate China chip-access row in our Nvidia China chips explainer, so we will only touch on it here.
Watch: the source video
What this video covers: Nate B Jones explains why Nvidia’s next deal matters more than the leadership announcement, what Anthropic’s Claude-on-AMD test says about the CUDA moat, how $105 billion of guarantees tie Nvidia to AI financing, where the AI Overwatch Act leaves chip exports and what it could all mean for the price of AI.
Video: Nate B Jones (@NateBJones), “Why Jensen Huang’s Resignation Could Change Your AI Bill”, published 13 November 2026. No paid sponsor is mentioned; Nate promotes his own prompt library and Substack newsletter. This summary, the fact-checks, the opinions and the practical advice are our own.
What actually happened at Nvidia?
Short answer: Jensen Huang, Nvidia’s co-founder and CEO for 33 years, has left the company completely. This part is not rumour.
- Nvidia’s SEC filing says the board accepted his resignation as president, CEO and director on 11 November 2026, effective immediately.
- The Nvidia newsroom announcement followed on 12 November.
- Colette Kress, CFO since 2013, is interim CEO and stays CFO while the board searches for a permanent leader, looking at internal and external candidates.
- Harvey C. Jones, on the board since 1993, becomes Nvidia’s first-ever chair.
- The board says it expects operations, customer commitments and the product roadmap to continue without interruption.
The resignation came days after press reports about Huang’s personal relationships, which led US senators to raise national security concerns. Nate does not dwell on that, and neither will we. His interest, and ours, is what happens next for people who buy AI.
⚠️ Warning: Expect a lot of hot takes this week, including names of possible successors. Until Nvidia names a permanent CEO, treat any shortlist you see as speculation.
Why does Nate think the next Nvidia deal matters most?
Nate’s core point: the board’s promise of “no interruption” is probably true for day-to-day work, but it misses the job Jensen actually did.
He asks you to imagine you are building an AI cloud and about to commit to years of Nvidia hardware. You need delivery dates, technical support and someone to call when things go wrong. Somebody has to decide which customer gets scarce chips, which new system gets priority and what terms make a giant deal possible.
For decades, Nate says, Jensen had the authority and relationships to make those calls, and nobody questioned them. Nvidia now has to rebuild that coordination role while it searches for a permanent leader. And it has to do it at the moment customers have more credible alternatives than ever.
💡 Tip: Nate admits the obvious counterpoints: thousands of skilled people still work at Nvidia, and customers still badly need its chips. His argument is about negotiating power at the margin, not about Nvidia falling over.
Is the CUDA moat starting to shrink?
Short answer: a little, according to Nate, and AI coding tools are the reason.
CUDA (Compute Unified Device Architecture) is the software layer developers use to make Nvidia chips do AI work. Years of code have been written for it. Moving that code to another company’s chips has traditionally taken a lot of engineering time, so even good rival hardware lost sales because switching was slow and expensive. That switching cost is often called Nvidia’s “moat”.
Nate points to Anthropic’s IPO roadshow as a sign of change. Anthropic says its unreleased Mythos model got Claude running on AMD MI355X chips in about three days, with “basically no human intervention”, compared with months of work by dozens of engineers in the past.
🔎 Unconfirmed / rumour: This is Anthropic’s own claim, made while marketing its shares, and it has not been independently checked. The slide’s footnote narrows it to bringing up prompt processing for one Anthropic model at a useful level of performance. Prompt processing is the stage where a model reads your input, not the whole job of running it. Nate stresses this too: it is not proof that all of Claude could move to new chips over a weekend.
Nate also notes that Anthropic has invested in several chip platforms for years, so it was unusually well prepared. His fair summary: the CUDA moat is still there and will be next year, but a small new risk is emerging as AI takes on more of the tedious low-level engineering that made switching so costly.
What happens if AI companies split their chip orders?
Short answer: Nvidia may keep its customers but earn a smaller share of each deal.
Nate’s example is simple. If you are building a private cloud, AMD does not have to beat Nvidia at everything. If some of your work can run economically on AMD, you can split your next order between suppliers. Nvidia might still win most of it, but it may have to negotiate harder to keep you.
| What changes | Who feels it first | What it could mean for you |
|---|---|---|
| Buyers can credibly split chip orders | AI labs and cloud providers | More pressure on chip prices over time |
| Software is easier to move between chips | Companies running their own models | Less dependence on one supplier’s roadmap |
| Nvidia negotiates harder to keep business | Large buyers signing multi-year deals | Savings that may later reach API and plan prices |
| Demand for AI keeps growing | Everyone | Prices may not fall fast, even with more competition |
From a customer’s point of view, Nate says, this is good news: more hardware choice can mean more computing power at lower cost. The catch is time. Chip deals are signed for years, so any saving reaches your monthly bill slowly.
Paying for several AI tools? This is a good week to check you are not locked in for longer than you need. Our guide on annual versus monthly AI subscriptions walks you through the decision in ten minutes.
How is Nvidia tied up in AI financing?
Short answer: very deeply, and Nate thinks this is the part the “no interruption” message underplays.
Building a data centre needs chips, but it also needs money. Lenders want confidence that tenants will pay rent, and tenants want confidence the hardware will arrive and work. Under Jensen, Nate says, Nvidia became part of all of those promises at once.
- The verified example: Nvidia’s August 2026 filing describes guarantees on SB Energy’s PORTS campus in Pike County, Ohio, with an OpenAI affiliate as tenant. Nvidia’s obligation is capped at $105 billion and only kicks in if OpenAI defaults on or cannot pay its leases, starting once the buildings are ready for service, expected from 2028.
- Nate’s wider claim: he says Nvidia’s quarterly reporting describes earlier deals aimed at raising more than half a trillion dollars of outside capital over time. We have not checked that figure line by line, so treat it as his reading of the filings.
Nate’s worry is not that these deals collapse. It is that uncertainty makes everything slightly harder: a lender asks for better terms, a client delays signing, a project takes a little longer to build. Across many projects, that could slow how fast new AI capacity arrives.
What about chip export rules?
Short answer: they are likely to get stricter, and Nvidia has lost its most powerful voice in that debate.
Jensen argued that selling to China keeps Chinese developers on American technology, which keeps American influence. Lawmakers who want tighter limits see the same dependency as a security risk. Nate describes the debate as heated and getting hotter.
The bill he names is the bipartisan AI Overwatch Act, introduced by Senators Jim Banks and Elizabeth Warren, which would give Congress a stronger role over advanced chip exports. In a 10 November statement, the two senators cited the concerns around Nvidia as a reason to pass it through the defence spending bill. Whether it becomes law is still open.
Nate thinks this matters more for Nvidia’s customers outside the US than those inside it. If rules tighten, cloud providers may face harder choices about where to put hardware and who they can sell to, even when chips and money are available. For UK businesses, the practical effect is indirect: it could shape where providers build and which services they offer in which regions. Our China chips article goes deeper on access.
Will the Jensen Huang resignation make AI cheaper?
Short answer: possibly, over time, but nobody can promise when or by how much.
Nate’s chain of reasoning goes like this. If Nvidia’s next leader cannot match Jensen’s deal-making, rivals win more business, hardware becomes more varied, margins shrink and running AI models (inference) gets cheaper for everyone. He is open that he does not know the answer, and that he expects Nvidia to remain a great company.
Here is how we would read it for a small business:
- Chips are only one cost. The price of your AI plan also reflects energy, staff, research and each company’s own pricing choices.
- Savings travel slowly. Multi-year hardware deals mean any fall in chip prices reaches API rates and monthly plans gradually.
- Competition already helps. Prices for the same level of AI ability have generally been falling as labs compete, so this is one more push in that direction, not a new trend.
- Uncertainty cuts both ways. Slower data-centre financing could tighten capacity in the short term, even if competition helps in the long term.
Nate points to Nvidia’s earnings call on 17 November as the next thing to watch. The numbers will cover a quarter that ended before the resignation, so he is more interested in three questions: who will be the permanent CEO, whether chip export limits become a bigger issue, and how Nvidia’s role in funding AI infrastructure changes after Jensen.
What should you do about it now?
Short answer: nothing dramatic. Just make sure you can benefit if prices fall, and are not hurt if a supplier changes.
- Do not panic-buy or panic-cancel. Your AI tools will work tomorrow exactly as they did yesterday.
- Check your renewal dates. Before you commit to a year, weigh it up using our guide to choosing an annual or monthly AI subscription.
- Ask your suppliers the right questions. Find out how they handle price changes, which models they use and how you leave. Our list of questions to ask your AI tool supplier before you renew includes an email you can copy.
- Keep a plan B. Make sure your prompts and workflows are not tied to one tool. Our AI backup plan guide shows how.
- Measure cost per finished job. When cheaper models appear, you will only notice the saving if you track it. See how to compare AI models on token cost.
💡 Tip: Put a 15-minute “AI costs” review in your calendar once a quarter. Note what you pay, what you use and whether a cheaper option now does the job. If prices fall, that is the moment you will catch it.
Frequently asked questions
Did Jensen Huang really resign from Nvidia?
Yes. Nvidia’s SEC filing says its board accepted his resignation as president, CEO and director on 11 November 2026, and the company announced it on 12 November.
Who is running Nvidia now?
Colette Kress, Nvidia’s CFO since 2013, is interim CEO while the board searches for a permanent replacement. Harvey C. Jones is the company’s first-ever board chair.
Will the Jensen Huang resignation make AI cheaper?
Not immediately. Nate B Jones argues that weaker deal-making power and stronger rivals could increase competition and lower AI costs over time, but that is a forecast, not a certainty.
What is the CUDA moat?
CUDA is Nvidia’s software for programming its chips. Because so much AI code is written for it, switching to another chip maker has been slow and costly, which protects Nvidia’s position.
Did Claude really move to AMD chips in three days?
That is Anthropic’s claim from its IPO roadshow. Its own footnote limits it to prompt processing for one model on AMD MI355X chips, not a full move of Claude.
Should I change my AI tools because of this?
No. Keep using what works, but avoid locking into long contracts without a review, and keep your workflows portable. Our guides on annual versus monthly plans and questions to ask suppliers can help.
What is the AI Overwatch Act?
A bipartisan US bill from Senators Jim Banks and Elizabeth Warren that would give Congress more say over exports of advanced AI chips. It had not become law at the time of writing.
The bottom line
The Jensen Huang resignation is confirmed, but most of what follows is still uncertain. Nate B Jones makes a reasonable case that Nvidia’s biggest challenge is replacing a deal-maker, and that more competition could slowly make AI cheaper. For a small business, the sensible response is calm and practical: keep using your tools, avoid locking yourself in, and review your AI costs regularly so you notice when better prices arrive.
Your next step: list every AI tool you pay for and its renewal date. Then use our guide to annual or monthly AI subscriptions and our supplier questions before your next renewal.
Related posts
- Nvidia China Chips Standoff: What It Means for Your Access to AI
- How to Build an AI Backup Plan in Case Your Main Tool Gets Pricier or Slower
- AI Data Centre Backlash: Is the AI Boom Over? What It Means for Your Business
