IBM’s mainframe business is experiencing a remarkable resurgence, demonstrating resilience in a landscape dominated by cloud computing. According to the company’s recent fourth-quarter earnings report, this sector achieved its highest revenue in over two decades, with a staggering 61% increase year-over-year after adjusting for currency fluctuations. This surge also contributed to a significant 17% uptick in IBM’s overall infrastructure segment.
Mainframes continue to be the backbone for many organizations that require robust reliability and security. Market Reports World highlights that approximately 71% of Fortune 500 companies rely on mainframes, alongside 92% of large financial institutions and 63% of governmental organizations. Notably, IBM serves as the provider for over 90% of installed mainframe systems worldwide, processing an astonishing 87% of global credit card transactions through these systems.
The secret to the ongoing relevance of mainframes lies in IBM’s consistent innovation and adaptation to client requirements. The latest z17 mainframe systems, which have fueled these record financial results, are specifically designed for the artificial intelligence (AI) era. This system supports more than 250 AI use cases encompassing diverse applications such as loan risk management, medical imaging, and retail security, and is capable of executing up to 450 billion AI inferencing operations daily—50% more than its predecessor—while achieving rapid response times averaging just one millisecond.
In addition to the capabilities of the z17, IBM markets the Spyre AI accelerator, which enhances the performance of AI models on these mainframes. Looking ahead, CEO Arvind Krishna forecasts a significant shift in AI workload deployment over the next three to five years, predicting that 50% of enterprise AI utilizations will move to private clouds or in-house data centers, with the other half relying on public models. This strategic pivot could lead to substantial cost efficiencies for businesses as cloud-based AI operations tend to be more expensive.
The z17, which was launched in mid-2025, is expected to remain a key fixture in IBM’s offerings until 2028, and with the increasing emphasis on AI, the next generation of mainframes is anticipated to further enhance AI inference capabilities.
IBM’s efforts extend beyond mainframes; the tech giant also provides AI consulting services and the watsonx AI software platform. With a comprehensive enterprise client base and significant mainframe installations, IBM holds a competitive edge in delivering AI products and services to its customers. The company’s revenue is projected to grow at least 5% by 2026, with software revenue anticipated to rise by 10%, and free cash flow improving by approximately $1 billion from its 2025 figures.
Although the mainframe represents a smaller segment of IBM’s overall business compared to two decades ago, it continues to be an essential driver of profit and cash flow for the company. By successfully adapting the mainframe for the AI age, IBM is not only achieving record sales but is also positioning itself as a formidable leader in the rapidly evolving AI landscape.