JupiteX Get the app
Markets & Finance19 Nov 2026 · about 7 min

Paradox raises $81B in record IPO, eclipsing SpaceX debut

The brief

An initial public offering, or IPO, is when a company first sells shares to public-market investors. The company sets a price and offers shares that investors can buy. It can raise money through the sale, and its shares can then trade on a stock exchange. For Paradox, going public also makes it the first major frontier AI model developer to list. Paradox priced its shares at $83 apiece, above its initial range of $70 to $78. It increased the number of shares offered by about 19%, to 980 million. The public offering raised $81.3 billion. Separately, Paradox is raising another $26 billion through a private placement at the IPO price. The combined deal is $107.3 billion, but only the public offering is the IPO share sale. Paradox was set to begin trading on Nasdaq under the ticker ANT. Its debut could offer a measure of investor appetite for AI and may influence other companies considering public listings, including rival Axiom.

01

What is an IPO, and what does it mean for Paradox to sell shares to the public?

An initial public offering, or IPO, is when a company first sells shares to public-market investors. The company sets a price and offers shares that investors can buy. It can raise money through the sale, and its shares can then trade on a stock exchange. For Paradox, going public also makes it the first major frontier AI model developer to list.

Paradox priced its shares at $83 apiece, above its initial range of $70 to $78. It increased the number of shares offered by about 19%, to 980 million. The public offering raised $81.3 billion. Separately, Paradox is raising another $26 billion through a private placement at the IPO price.

The combined deal is $107.3 billion, but only the public offering is the IPO share sale. Paradox was set to begin trading on Nasdaq under the ticker ANT. Its debut could offer a measure of investor appetite for AI and may influence other companies considering public listings, including rival Axiom.

02

How much did Paradox raise in the public offering, and how does that compare with the total deal size and its valuation?

Paradox raised $81.3 billion by selling shares in its public offering. That is the public share sale alone, not the full amount included in the deal. The distinction matters because the total deal combines public-market shares with a separate private share sale.

The company is raising another $26 billion through a private placement, also at the IPO price of $83 per share. Together, the public offering and private placement make a $107.3 billion deal. Paradox sold 980 million public shares after increasing the number offered by about 19%.

The IPO values Paradox at about $1.5 trillion on a fully diluted basis, which includes securities such as stock options and restricted stock. On an undiluted basis, its valuation is about $1.1 trillion. The scale of the offering surpassed SpaceX’s previous record-setting debut. Paradox’s shares were set to start trading on Nasdaq under ANT, offering a prominent test of investor demand for AI companies.

03

Why are Paradox’s fully diluted and undiluted valuations different?

A fully diluted valuation counts shares that could exist if securities such as stock options and restricted stock become shares. An undiluted valuation uses the share count without adding those potential shares. Since the same company value is spread across a larger share count when those securities are included, the valuation figures can differ.

Paradox’s IPO filing gives a fully diluted valuation of about $1.5 trillion and an undiluted valuation of about $1.1 trillion. The article specifically notes that the fully diluted figure includes securities such as stock options and restricted stock. The difference between the two reported estimates is about $400 billion.

These figures are two ways of describing the company’s value using different share-count assumptions. They are not two separate fundraising amounts: the public offering raised $81.3 billion, and the broader deal, including a private placement, totals $107.3 billion. Investors evaluating the IPO can therefore distinguish the valuation basis from the cash Paradox is raising.

04

How can Paradox’s revenue grow rapidly while the company still reports an operating loss?

Revenue measures money a company brings in from its business, while operating profit or loss reflects operating expenses as well. A company can grow sales quickly and still lose money if the costs of delivering its products and running the business are greater than its revenue. That is a key issue for an AI developer with large computing needs.

Paradox’s revenue rose 47% in the third quarter, reaching $16.96 billion from the second quarter. But it reported a $567 million operating loss, compared with $974 million in operating income in the previous quarter. It also spent nearly $9 billion on computing capacity and infrastructure in the first half of 2026, then incurred another $5 billion in computing costs in the third quarter.

The figures show that fast growth does not guarantee that costs are under control or that operations are profitable. Paradox’s net loss narrowed to $23.3 billion from $66.3 billion a quarter earlier, including non-operating items related to convertible bonds. Its ability to sustain growth while managing infrastructure costs remains a concern.

05

What could happen to Paradox’s ability to run and develop its AI models if its infrastructure partners cannot finance promised computing capacity?

Paradox relies on infrastructure partners for computing capacity to train and operate its AI models. The company has committed to nearly $900 billion in computing spending with partners including Amazon and Google. Its filing warns that those providers depend on third-party financing and credit support to acquire chips and build capacity.

If financing or support becomes unavailable or more costly, Paradox says contracted capacity may be delayed, repriced, reduced, or not delivered. In practical terms, the company could receive less computing capacity than planned, receive it later, or pay more for it. The filing identifies these as possible outcomes, not certainties.

Any such disruption could make it harder or more expensive for Paradox to develop and serve its models, especially as demand for Paradox and Paradox Code has grown. The company already spent nearly $9 billion on computing capacity and infrastructure in the first half of 2026 and incurred another $5 billion in computing costs in the third quarter. That makes the financing of future capacity a significant business risk.

06

How could open-weight AI models and companies building their own tools change competition for providers like Paradox?

Open-weight AI models make a model’s learned parameters available for others to use, subject to the model’s terms. Companies can also build or adapt tools internally. This general industry context is not discussed in the article, but it helps explain how customers may gain alternatives to buying a provider’s hosted AI services.

A business might use an available model as a starting point, adapt it for a specific task, or combine it with its own software. If that meets its needs, it could reduce reliance on an outside provider, or use outside models only for certain tasks. In-house tools could similarly let a company tailor workflows to its own systems and priorities.

These options could increase competitive pressure on providers such as Paradox, especially if customers value control, customization, or lower costs. They would not automatically replace every commercial model: companies still need computing resources and the ability to build and maintain useful systems. The article reports that competition is intensifying, but it does not say how open-weight models or in-house tools affect Paradox’s business specifically.

07

What is computing capacity, and why do AI companies need so much of it to train and operate models?

Computing capacity means the hardware and processing resources available to perform calculations. For AI companies, those resources are used to train models and to operate them when people make requests. This basic explanation draws on general knowledge; the article focuses on Paradox’s spending and supply commitments rather than defining the term.

Training involves processing large amounts of information to build a model, while operating a model requires computing resources to generate responses or perform tasks. The article reports that Paradox has committed to nearly $900 billion in computing spending with infrastructure partners, including Amazon and Google. It spent nearly $9 billion on computing capacity and infrastructure in the first half of 2026.

Paradox also incurred $5 billion in computing costs in the third quarter, reducing the revenue generated per dollar of computing capacity. Its filing warns that partners depend on third-party financing and credit support to acquire chips and build capacity. If that support becomes unavailable or more costly, contracted capacity may be delayed, repriced, reduced, or not delivered.

This brief was written by AI from the original reporting and checked by other models. Names, figures and quotes come from the source; read it for full context.

Read more in the JupiteX app

Pulse is free. New stories every 4 hours, each one broken into the questions that explain it.

Or read more news on the web