Nvidia Stock Hits Record as $500B AI Financing Push Takes Shape

A record rally built on more than chip sales

Nvidia shares closed at $239.24 on Tuesday, October 6, just below an all-time high set earlier that afternoon. The company's market value is now approaching $6 trillion.1 One record close in a strong year is not the story here. The story is how the rally is being paid for. Over the past week the stock has been pushed up by three things at once. Wall Street is putting together a financing system around Nvidia's GPUs. Nvidia keeps writing large checks into AI startups. And the board approved the largest buyback authorization in corporate history. Together they turn a chip company into something closer to the AI industry's financier.

The buyback came first. Nvidia's board added $150 billion to its repurchase program, which leaves $235 billion authorized through fiscal 2028.3 On October 2 the stock rose almost 3% to an intraday record near $237.75, its first new high since May. That same day Morgan Stanley put Nvidia back at the top of its semiconductor picks after meeting with CEO Jensen Huang.3 A few trading days later, shares gained another 1.2%. Traders tied that move to Nvidia's financing partnership with major Wall Street firms, which is meant to cut the upfront cost of AI infrastructure for customers.2

The company's results back up the move. Nvidia reported $96.2 billion in revenue in its August quarter and guided to between $105.8 billion and $110.1 billion for the current one.1 Data center sales grew 117% from a year earlier.4

The $500 billion machine, and what it isn't

The main financing development dates to August 10. That day Nvidia signed memorandums of understanding with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR. The goal is to build independent compute-financing platforms that would raise more than $500 billion in outside capital for AI infrastructure.2228 Nvidia can guarantee up to 25% of residual value on qualifying deals. Spread across the full target, that works out to a maximum exposure of about $125 billion.2224

This is where the coverage disagrees most. Some reports describe Nvidia as "channeling" up to $500 billion into AI infrastructure itself, as if the chipmaker were putting up the money.29 Morningstar's credit analysts draw the line more carefully. They note that the platform is not a fully committed fund and that Nvidia is not a direct investor. Most of the money is expected to come from banks, insurers, asset managers and private credit lenders.24 Other reporting also points out that MOUs are agreements in principle, so the $500 billion figure is a fundraising goal, not money already raised.23 Morningstar's reading is the accurate one. Nvidia is offering its name and a partial guarantee, not $500 billion of its own cash.

That difference matters because lenders have already pushed back. By October 1, some Wall Street investors had questioned whether fast-depreciating chips make good long-term collateral and were asking for stronger guarantees on future deals.21 Recent debt pricing shows how much a guarantee is worth. An Apollo-led $35 billion package backed by Broadcom's AI chips was priced at 5.75% with Broadcom's guarantee and 8.5% without it, a gap of 275 basis points.26 Lenders who want a bigger cushion from Nvidia have a strong argument, and Nvidia has every reason to give in, because each extra point of guarantee makes deals easier to close.

The market did not cheer at first. Nvidia fell 2.86% on the August announcement day, while Apollo and KKR each rose more than 6%.26 Investors initially saw the deal as a fee opportunity for asset managers. Only later did they treat it as support for Nvidia's own demand.

Nvidia's venture book keeps growing

The Wall Street platform sits alongside a large and fast-growing pile of direct equity bets. By late July 2026, Nvidia's equity investments had reached about $99 billion.23 One estimate puts its startup pledges above $70 billion over three years, plus roughly $300 billion in financial support to customers. Another puts its total guarantees and commitments to the AI ecosystem above $530 billion.5 These totals come from different methods and should be read as rough sizes, not audited figures. They all point the same way, though.

The frontier labs account for the biggest checks. Nvidia and SoftBank are reportedly putting in a final $20 billion to close OpenAI's pre-IPO round, which had been reported at about $30 billion at a $1.4 trillion valuation.15 Nvidia also holds a stake of roughly $10 billion in Anthropic, which has reportedly prepared a confidential IPO filing.15 In July, Nvidia invested a reported $5 billion in Ilya Sutskever's Safe Superintelligence at a $32 billion valuation, even though the lab has no product or revenue.11

Below the frontier labs, Nvidia shows up in a long list of rounds:

• Fireworks AI raised a $1.5 billion Series D at a $17.5 billion valuation, with Nvidia participating.11
• Together AI raised $800 million at an $8.3 billion valuation, with Nvidia among the investors.11
• GMI Cloud raised $223 million in equity plus a $445 million credit line, and CScale raised $145 million for optical interconnects. Nvidia backed both.18
• Nous Research reached unicorn status with a $90 million Series B at a $1.5 billion valuation, with Nvidia, Microsoft's M12 and Samsung participating.17
• Mecka raised a $60 million Series B from Sequoia, Nvidia, M12, Qualcomm Ventures and Samsung Next.13

Nvidia has also reportedly discussed another $1 billion investment in humanoid robotics company Figure AI. Figure is said to be raising at about a $38 billion pre-money valuation. No deal has been confirmed.1920

One newsletter-based report says Nvidia bought Hugging Face for about $13 billion and paid roughly $6 billion to acquire Poolside's team and license its technology. It also cites a roughly $2 billion investment in Reflection AI at a $25 billion pre-money valuation.12 These figures come from a single analyst dispatch and should be treated as reported, not confirmed. If accurate, they would make Nvidia a buyer of the main distribution layer for open-weight models, not just an investor in them.

Other companies are also making acquisitions. AMD bought World Labs, its second-largest deal after Xilinx. AMD and Nvidia had both invested in World Labs' $1 billion round in February.18

The circularity question gets sharper

Reporting on the risks largely agrees. Nvidia's critics call the pattern circular investing: Nvidia funds customers, who then buy more Nvidia chips.1 The OpenAI round shows this most clearly. Nvidia is supplying the chips and also helping fund their purchase. One analysis notes that OpenAI's annualized revenue has stayed around $25 billion since February while its valuation kept rising.15

The financing platform adds a second loop. One report notes that Nvidia makes money by releasing faster chips that make older ones less attractive, while this financing model depends on older chips holding their value.21 If borrowers default, repossessed GPUs could flood the resale market and push down the value of the collateral Nvidia guaranteed. Nvidia reportedly held talks with insurers in late September about covering defaults on loans backed by GPU collateral.5 Michael Burry has argued that Big Tech depreciation schedules are too generous, and Nvidia has publicly disputed that.9

Morningstar's assessment is balanced. It calls the partnerships credit-positive overall, but warns they could deepen Nvidia's ecosystem lock-in and set up overinvestment if new capacity outruns AI demand that can actually be monetized.24 Companies outside Nvidia's ecosystem may also find capital harder and more expensive to get.24

Why it matters

The constraint on AI growth is moving. Morgan Stanley argues that the main bottleneck is shifting from chip production to land, power and financing.3 The firm also estimates that AI infrastructure will need $1.5 trillion in outside financing by 2028.30 Borrowing is already large. SpaceX is reportedly seeking $40 billion, led by Apollo, mostly to buy Nvidia chips.30 Banks such as Citigroup, JPMorgan and Barclays are now underwriting loans secured directly against GPUs in Asia, risk that used to sit mostly with private credit funds.27 Some AI cloud providers have started hiring traders to buy and sell GPU contracts.8

Nvidia has seen this shift coming and is positioning itself on the financing side as well as the hardware side. That is the strongest reading of the record stock price. In the near term, the setup is a strong positive. It widens the customer base, makes revenue more predictable and makes CUDA harder to replace. The risk is a slow build: more of Nvidia's reported demand now depends on credit conditions and on GPUs holding their resale value, with much of the exposure sitting off the balance sheet.5 Nvidia has about $99 billion in cash and liquidity as a cushion.5 The things to watch are whether the MOUs turn into binding deals, how far lenders push the 25% backstop up, and what used GPUs actually sell for. Those will show whether this is a durable new asset class or leverage that only works while demand keeps rising.