A Century-Long Bet on Artificial Intelligence
Alphabet, the parent company of Google, has executed one of the most unusual capital markets transactions in recent technology industry history: a $5 billion offering of 100-year bonds, the proceeds of which will be directed almost entirely toward funding the company's rapidly expanding artificial intelligence infrastructure. The century bonds, which will not mature until 2126, represent a dramatic statement of confidence in both Alphabet's long-term viability and the enduring economic value of AI technology.
The offering was priced at a yield of 5.2 percent, reflecting the extreme duration risk that investors accept when purchasing bonds with a century-long maturity. For context, Alphabet's conventional 30-year bonds trade at yields of approximately 4.4 percent. The additional 80 basis points of yield compensate investors for the vastly greater uncertainty associated with lending money for 100 years, a time horizon that spans multiple economic cycles, technological paradigm shifts, and geopolitical transformations.
Why Century Bonds?
Century bonds are exceedingly rare in corporate finance. Only a handful of companies have successfully issued them, including Disney, Coca-Cola, and Norfolk Southern, typically during periods of low interest rates when investor appetite for yield drives demand toward unusual maturities. Alphabet's decision to issue 100-year bonds in the current interest rate environment, which is significantly higher than the near-zero rates of the early 2020s, signals that the company views its AI capital requirements as so large and so long-term that unconventional financing is warranted.
Alphabet's chief financial officer explained the rationale in a call with analysts, noting that the company's AI infrastructure buildout will require sustained capital investment over the next decade and beyond. By locking in long-term financing now, Alphabet avoids the refinancing risk associated with shorter-maturity debt and gains certainty about a significant portion of its future funding costs.
The Numbers Behind the AI Buildout
Alphabet has disclosed that it plans to invest approximately $75 billion in capital expenditure during the current fiscal year, the vast majority of which will be directed toward AI-related infrastructure. This includes the following categories:
- Data centers: Alphabet is constructing or expanding data center facilities in more than a dozen locations worldwide, including new campuses in Texas, Mississippi, Malaysia, and Finland. These facilities will house the GPU clusters and custom TPU accelerators that power Google's AI services.
- Custom silicon: The company is investing heavily in the development and production of its next-generation Tensor Processing Units, custom AI chips designed in-house that provide cost and performance advantages over commercially available GPUs for certain workloads.
- Networking infrastructure: AI training workloads require massive data transfer bandwidth between servers. Alphabet is deploying next-generation optical networking equipment and expanding its global fiber backbone to support the communication demands of distributed AI training.
- Energy infrastructure: AI data centers consume enormous amounts of electricity. Alphabet has signed long-term power purchase agreements totaling more than 3 gigawatts of renewable energy capacity and is exploring nuclear energy partnerships to secure reliable baseload power for its most demanding facilities.

