While artificial-intelligence infrastructure stocks have received much of investors’ attention in recent years, it might be the software sector’s turn to rack up big gains in 2026.
In particular, companies making infrastructure software, which manage hardware and data, present a compelling opportunity for investors, according to analysts at Goldman Sachs and Deutsche Bank. Microsoft, Oracle and ServiceNow fall into this category.
Over the past year, semiconductor stocks have dramatically outperformed software stocks, with the iShares Semiconductor ETF gaining 52% and the iShares Expanded Tech-Software Sector ETF rising just 8%.
However, the next three years will present revenue acceleration opportunities for the above companies, Goldman Sachs analyst Gabriela Borges wrote in a Sunday note. AI could expand the total addressable market for software by roughly 30% through 2037, she added.
Goldman Sachs named Microsoft its top pick among AI infrastructure names thanks to its multitude of opportunities to monetize its Azure cloud business. Borges expects Azure to drive revenue estimates higher in 2026 as it capitalizes on a “fungible” capacity, meaning that its data centers can be easily utilized for a variety of different workloads and customers.
Beyond its massive cloud infrastructure, Microsoft is cementing its leadership in the software-application layer with the introduction of Agent 365 and Foundry. These products allow customers to build custom AI agents with governance and security controls, making Microsoft a critical player in the “orchestration layer” to manage chatbots, Borges wrote.
Analysts at Goldman Sachs and Deutsche Bank both called Oracle a top pick. Deutsche Bank’s Brad Zelnick noted that Oracle’s future deal pipeline has swelled to over $500 billion, making it an industry leader in deploying AI infrastructure at scale.
Oracle Cloud Infrastructure, the company’s cloud platform, has hit an “inflection point” thanks to robust demand for AI workloads, Borges wrote. She anticipates that revenue will accelerate in 2026 as the company converts these deals into recognized revenue.
Oracle’s unique database architecture, which was built to promote speed and efficiency between servers, is especially well-suited to AI workloads with high volumes of data, according to Goldman Sachs.
Beyond AI workloads, Oracle is still seeing growth from companies migrating their data to the cloud, Zelnick said. Investors could be underestimating the growth of Oracle’s non-OCI revenue, he added.
For ServiceNow, Borges believes the company will be able to grow revenues organically at around 20% annually through 2029. The workflow-management company has developed what it calls an AI Control Tower to provide governance and oversight for AI agents across a company.
“We believe outsized value will accrue in the agent orchestration layer that controls and governs agents across multiple vendors,” putting ServiceNow in a good spot, Borges wrote.
ServiceNow has also ramped up its acquisition strategy recently, announcing plans to buy cybersecurity firm Armis at the end of 2025 to expand its cybersecurity offerings. ServiceNow aims to integrate Armis’ real-time threat intelligence with its automated workflows to proactively reduce security risks. The acquisition could help create “a meaningful new product cycle for ServiceNow” over the next five years, according to Borges.
