Michael Burry doubles down on NVIDIA short after Wall Street stunt with 'Shades of ENRON'

quote:

---

Almost all of the convictions were overturned

---

Arthur Anderson got off the hook.. that is it

The CEO got one charge dropped but was still convicted on securities fraud, conspiracy, and insider tradin. Served 14 years in prison

I guess Kenneth Lay the founder got off because he died of a heart attack, I am sue you are counting that.

The CFO also served 6 years in prison..

That being said....

The circular financing structure and the fact that 70% of all AI spend is dependent on OpenAI and Anthropic is a massive red flag. WE all know OpenAI is a financial disaster... If even one of these gets shaken it will be catastrophic to the valuations of most of the related companies.

Technology is here to stay but China is catching quick with their open models. So the truth is in the middle.

I think the govt will have to come in and save OpenAI with our tax dollars at some point.

My wife sold all of our NVDA last week, for about $207, causing it to hit beastmode to the $220's.

quote:

---

Is he still investing in water?

---

Huge holdings in water.

Burry made $100MM profits on the Big Short trade. If he had put that into VOO, he’d be a billionaire. Instead he has net worth estimated in the $300MM range bc he loses more than he wins and it’s not close. A good portion of his net worth is from non-trading income

I mentioned this in the NBIS thread, I don’t get why Burry gets so much attention. He was right big time on the housing market, but what else has he been right on? People just run to him for notoriety and name recognition

Is my dry powder move going to pay off?

Nvidia backstopping its customers. Nvidia has increasingly helped finance or guarantee portions of the ecosystem buying its products. For example, Nvidia has supported CoreWeave and agreed in certain circumstances to purchase unused compute capacity. More recently, Nvidia has been working with Apollo, Blackstone, BlackRock, Brookfield, Goldman Sachs and KKR on a program intended to mobilize more than $500 billion of third-party financing for AI infrastructure, with Nvidia potentially providing support for some transactions.

Circular / “Off balance sheet” $

Per the gpt wizard, Example:
Wall Street/private credit ? lends $10B to AI data-center company ? company buys $8B of Nvidia GPUs ? Nvidia records $8B revenue.

But then Nvidia might:

invest in the AI company ? guarantee some financing ? guarantee residual GPU values ? or promise to buy compute capacity if demand falls short.

So some of the money supporting the customer ultimately depends on Nvidia’s own financial strength or Nvidia-backed collateral.

That produces something resembling:

Nvidia support ? customer borrowing ? Nvidia GPU purchase ? Nvidia revenue ? higher Nvidia valuation/cash flow ? more Nvidia support ? more borrowing ? more Nvidia GPU purchases.

Bigger AI bubble concern…

Uncommenced data-center leases. Hyperscalers have signed enormous LT debt. Goldman Sachs estimates roughly $1.5 trillion of aggregate lease commitments, with about $1 trillion related to leases that have not yet commenced.

Huge purchase commitments. Companies are committing in advance to buy GPUs, servers, cloud capacity, electricity and other infrastructure. The FT recently calculated that purchase commitments among major hyperscalers had risen to nearly $1.5 trillion by mid-2026. These aren’t necessarily “debt” in the accounting sense, but economically they can behave like fixed obligations.

Special-purpose vehicles and project financing. Instead of Microsoft, Meta, Oracle or an AI company borrowing directly to build a data center, an outside entity/SPV can borrow the money, construct the facility and lease it back to the tech company. Earlier this year the FT estimated that more than $120 billion of AI data-center financing had already been structured off corporate balance sheets using arrangements involving Meta, Oracle, xAI, CoreWeave and others.

GPU-backed private credit. Companies such as CoreWeave have borrowed against Nvidia GPUs themselves. CoreWeave explicitly described itself as having pioneered “GPU infrastructure-backed lending” and said it had raised more than $14.5 billion through debt and equity financings by the time of its IPO filing. The vulnerability is that the collateral is technology that can depreciate very quickly if newer Nvidia chips—or competing chips—make older GPUs economically obsolete.

My contrarian is flaring up today. Need some Benadryl.

quote:

---

Shades of Enron

---

He obviously has no idea what happenned at Enron if he is correlating a public announcement of $500 billion of 3rd party money with Enron's Jedi and the like.

As i said in the other thread, Burry has called 27 of the last 2 bubbles ... more crying wolf does not move the needle.

At the end of the day, regardless of the LLM provider (or open source) we are massively short compute and the game has just begun. The real usage rate of GPUs (well beyond 5 years nevermind beow it) and ST auction rates of MWs clearly demonstrate it with physical dollars being spent.

If you believe we are going to truncate compute growth and just stop and watch China run away - all the power to you - but I do not believe that will happen. I do not think the massive earnings growth across the market is independent of AI - it is not a coincidence.

Not a single factor to me suggests bubble in the classic sense of it. Yes there will be misallocated capital but there is currently more real demand then supply - that is not the sign of a bubble.

The circular financing argument is way stretched also - if Nvidia directly lent credit to a customer like hundreds of companies do, would we say circular financing? But somehow not providing the financing, getting 500 billion of third party capital injected is. I do not understand it.