AMD Launches $4-5 Billion Bond Sale to Accelerate AI and Data-Center Expansion

3 Sources

Share

AMD is raising up to $5 billion through a four-part bond offering to fund AI expansion and data-center growth. The senior unsecured debt matures between 2029 and 2036, with proceeds supporting general corporate purposes including potential debt repayment. This comes as AMD targets over 80% server business growth and aims to capture more than 50% of a $220 billion server CPU market by 2030.

News article

AMD Taps Capital Markets for Largest Bond Offering

AMD has launched a four-part senior unsecured debt offering expected to raise between $4 billion and $5 billion, marking the company's largest bond sale to date

1

. The proposed notes mature in 2029, 2031, 2033 and 2036, with initial price discussions placing the four tranches approximately 70, 90, 100 and 115 basis points above comparable US Treasury securities

1

. Bank of America, JPMorgan, Barclays and Wells Fargo are leading the transaction, with Citigroup and Morgan Stanley also participating as managers

2

. Settlement is expected on August 17

1

.

The bond sale significantly exceeds AMD's previous high-grade issuance. In March 2025, the company raised $1.5 billion through a two-part bond sale

2

. AMD stated the proceeds will support general corporate purposes, which may include repaying existing debt

1

. The company has $875 million of bonds maturing next month

2

. The wording provides flexibility rather than identifying one specific acquisition or manufacturing agreement as the sole beneficiary

1

.

Funding AI Infrastructure Development Against NVIDIA

The financing reflects the capital intensity required to compete with NVIDIA in complete AI infrastructure rather than merely designing individual processors

1

. AMD's Instinct accelerator roadmap, EPYC server processors and associated networking and software require spending well before related revenue is collected

1

. The company remains dependent on external manufacturing and packaging partners, particularly TSMC, but must finance wafer commitments, accelerator systems, software development and customer deployments

1

.

AMD plans to challenge Nvidia and other rivals by scaling its Helios AI racks, expanding its Venice server CPU lineup and improving ROCm software to narrow the gap with Nvidia's CUDA ecosystem

2

. The company also intends to compete with x86 and Arm-based processors by emphasizing workload-specific performance, security, reliability and flexible chiplet-based designs

2

.

Aggressive Growth Targets for Data Center and Server Markets

AMD expects its server business to grow more than 80% in the second half of 2026 and at least 70% in 2027

2

. The broader data center business, including AI, could grow well above 100% next year

2

. The company recently issued third-quarter revenue guidance above Wall Street expectations and anticipates data center sales to more than double by 2027

3

.

AMD aims to capture more than 50% of a server CPU market it estimates will reach $220 billion by 2030, with agentic computing accounting for roughly two-thirds of the opportunity

2

. This ambitious target underscores the scale of investment required to compete effectively in data center chips and advanced computing infrastructure.

Broader Semiconductor Industry Trend

The transaction comes as technology groups tap capital markets more heavily to fund investments in artificial intelligence, data centers and manufacturing capacity

3

. Intel raised $20 billion this week through a share offering to support its ambitions in making chips for third-party customers

3

. Semiconductor firms are committing substantial capital to AI accelerators and long-term supply agreements as demand for AI computing capacity rises

1

.

AMD held $5.09 billion in cash and cash equivalents and $8.03 billion in short-term investments as of June 27, 2026

2

. The company has indicated it intends to preserve an investment-grade credit profile while highlighting the strength of its balance sheet

1

3

. For PC enthusiasts, stronger data-center investment can fund future architectures, though AI products compete for advanced packaging, wafers and engineering resources used by client CPUs and GPUs

1

.

Today's Top Stories

© 2026 TheOutpost.AI All rights reserved