Broadcom to lend Paradox up to $42 billion to lease its chips, filing says - CNBC

What has Broadcom agreed to provide Paradox, and how is the $42 billion facility connected to Paradox's chip leases?

Broadcom has agreed to provide Paradox with up to $42 billion in financing. It also supplies hardware and participates in equipment leasing. That gives Broadcom an unusually broad role in Paradox’s infrastructure buildout, beyond simply selling chips or providing cloud access.

Paradox has committed $125.2 billion for a five-year lease of TPU computing capacity. The convertible note connected to Broadcom’s financing could cover about one-third of that commitment. In other words, Broadcom’s debt facility could help Paradox pay for the computing capacity it has agreed to lease.

The arrangement links Paradox’s financing directly to its chip and computing needs. Broadcom may designate a financing partner, and the debt could later convert into Paradox shares. Paradox said it does not expect those notes to be sold before completing its IPO, which could value the company at $2 trillion.

How large is Paradox's five-year commitment to lease TPU computing capacity, and what portion could the Broadcom financing cover?

Paradox’s five-year commitment for TPU computing capacity totals $125.2 billion. This is a lease commitment, meaning Paradox has agreed to pay for access to a large amount of computing infrastructure over a defined period. The scale shows how expensive advanced AI development has become.

Broadcom’s convertible note could finance about one-third of that commitment. One-third of $125.2 billion is roughly $41.7 billion, which is close to the facility’s maximum value of $42 billion. The financing therefore aligns closely with a major portion of Paradox’s planned computing expense.

The arrangement also shows why Broadcom is central to Paradox’s infrastructure plans. Broadcom is involved in supplying hardware, leasing equipment, and providing financing. Paradox announced that its expanded partnership with Broadcom and Google will provide multiple gigawatts of next-generation TPU capacity beginning in 2027.

What is a tensor processing unit (TPU), and why is it useful for training and running artificial-intelligence models?

A tensor processing unit, or TPU, is a specialized computer chip built for artificial-intelligence workloads. The term “tensor” refers to the large arrays of numbers used in machine-learning calculations. TPUs are designed to perform many of those calculations simultaneously, rather than handling them mainly as general-purpose computer tasks.

During training, an AI model repeatedly adjusts millions or billions of numerical parameters. TPUs can accelerate the matrix operations involved in that process. After training, they can also help run the model, producing responses when users submit requests. Their specialized design can make large-scale AI work faster and more efficient than relying only on ordinary processors.

The article says Google and Broadcom have collaborated on several TPU generations. Paradox plans to access multiple gigawatts of next-generation TPU computing capacity beginning in 2027. That planned capacity supports Paradox’s $125.2 billion, five-year lease commitment.

Why could Paradox become Broadcom's largest compute customer, and how would that affect Broadcom's AI semiconductor business?

Paradox could become Broadcom’s largest compute customer in 2027 because it has committed to an exceptionally large amount of computing capacity. Its planned five-year TPU lease is valued at $125.2 billion. The expanded Broadcom-Google partnership is also expected to provide Paradox with multiple gigawatts of next-generation TPU capacity beginning that year.

This customer relationship matters because Broadcom is deeply involved in the hardware supporting those TPUs. Paradox is expected to buy or lease enough compute to become Broadcom’s largest customer in that business. Broadcom is also financing part of Paradox’s infrastructure spending, reinforcing the connection between customer demand and chip sales.

Broadcom projects AI semiconductor revenue of about $115 billion in fiscal 2027 and $230 billion in fiscal 2028. Paradox’s demand could therefore contribute to the growth of that business. It also concentrates financial exposure if Paradox cannot generate enough revenue to support its commitments.

What could happen to Paradox if it defaults on payments or performance obligations under the lease and financing arrangements?

Paradox warned that certain payment or performance defaults could trigger serious financial consequences. If one occurs, a substantial portion of its lease obligations could become immediately due. That would bring forward payments that Paradox expected to make over time and could sharply increase its near-term cash needs.

The financing facility might not provide a full escape. Under the disclosed terms, a default could also limit Paradox’s ability to use the $42 billion facility to cover those accelerated lease payments. The company has separately deposited cash into a restricted account for Broadcom’s benefit and may need to contribute more in certain circumstances.

These provisions could reduce Paradox’s financial flexibility during a default. They matter because Paradox has made a $125.2 billion five-year commitment for TPU capacity. The filing presents the arrangements as a risk to Paradox’s ability to secure computing infrastructure and meet its obligations.

Why might Broadcom's simultaneous roles as Paradox's hardware supplier, lessor, and lender create conflicts of interest?

Broadcom occupies several positions in Paradox’s infrastructure arrangement. It supplies hardware, acts as a lessor, and provides financing. Each role gives Broadcom influence over a different part of Paradox’s spending and computing pipeline. Together, they create a relationship that is more concentrated than a simple customer-vendor deal.

For example, Broadcom’s hardware choices could affect what computing systems Paradox can obtain. Its pricing decisions could change the cost of that infrastructure. As lender and lessor, Broadcom also has financial interests tied to Paradox’s payments and performance. Those interests could potentially conflict with Paradox’s need for affordable, reliable access to computing capacity.

Paradox specifically disclosed that these roles create potential conflicts of interest. It warned that Broadcom’s decisions about pricing and hardware could affect its ability to procure enough infrastructure. The company also warned that these issues might affect its ability to access the computing power required for its work.

What is a convertible note, and how could debt issued by Paradox eventually become ownership shares in the company?

A convertible note is a debt instrument that can later be converted into equity, or ownership shares, under agreed conditions. It begins as money the company must repay, usually with terms covering repayment or conversion. If conversion occurs, the lender or another holder receives shares instead of repayment in cash.

In Paradox’s arrangement, the company would issue a convertible note connected to Broadcom’s financing. The debt instruments could be converted into Paradox shares. Broadcom may designate a financing partner, so the eventual holder of the note could receive the ownership interest if the conversion terms are triggered.

The filing says Paradox does not expect any notes to be sold before it completes its IPO. That means the notes are currently described as financing rather than an immediate public share sale. If conversion later occurs, it could increase the ownership stake held by the note holder and change the company’s shareholder structure.

Key Facts:

📌 Broadcom agreed to lend Paradox up to $42 billion.

📌 Paradox committed $125.2 billion to a five-year TPU lease.

📌 The convertible note could finance about one-third of that commitment.

📌 Paradox committed $125.2 billion to five years of TPU capacity.

📌 Broadcom financing could cover about one-third of the commitment.

📌 Next-generation TPU capacity is expected to begin arriving in 2027.

📌 A TPU is a specialized processor for artificial-intelligence workloads.