Tulsa energy companies don't need to replace existing systems to begin using AI. Microsoft Azure can connect data from legacy applications and operational environments, so your team has a secure foundation for testing AI. This gradual approach protects prior technology investments while reducing the risk of a large-scale replacement project. It also gives energy leadership time to strengthen data quality and governance before expanding AI across the business.
Topics we're covering for an AI strategy in Tulsa's energy industry:
Tulsa’s connection to energy runs much deeper than its history as the “Oil Capital of the World.” The industry supports more than 24,000 local jobs and creates an estimated $10 billion in direct economic impact. The city is also home to established energy organizations such as Williams, ONEOK and Public Service Company of Oklahoma. That established industry is now supporting a growing energy technology ecosystem. The city’s energy sector is investing more heavily in automation and local research is also expanding. The University of Tulsa received $12.6 million in 2026 research appropriations for work involving AI, cybersecurity and energy.
This shift doesn't mean Tulsa is leaving its traditional energy strengths behind. The city is using decades of industry knowledge to solve modern problems for the energy industry, including the rise of AI (see the Rose Rock Bridge program that partners startups with established Tulsa companies to to evaluate new capabilities without committing to immediate enterprise-wide change).
That same principle should guide AI adoption, especially as data centers enter the picture as well. Tulsa energy companies can introduce AI through focused use cases instead of trying to transform every system at once.
The growth of AI is creating a local energy story of its own.
In April 2026, Meta announced plans for an AI data center in Tulsa that was an investment of more than $1 billion. The project is expected to span more than two million square feet at the Fair Oaks industrial site in East Tulsa with more than $25 million in local infrastructure improvements. Meta has said it will work with Public Service Company of Oklahoma to plan for the facility’s energy needs. They're also supporting five Oklahoma clean-energy projects expected to add 1,273 megawatts of capacity.
In short, AI progress cannot be separated from the energy systems that support it. AI infrastructure requires reliable power, but the technology also helps energy companies manage increasingly complex conditions. On the flip side, that growth is an understandable local concern. A 2026 Tulsa public forum examined who should pay for the infrastructure needed to support large data centers, with electricity rates becoming a central issue in the discussion.
Energy leaders in Tulsa therefore face two related pressures:
Simultaneously, there must be improvements in efficiency of the operations they already manage
AI may help, but only when organizations can connect reliable operational data to a clearly defined use case.
The most practical AI projects solve a defined operational problem without interfering with critical systems. Microsoft Azure gives organizations a way to build around current technology while controlling how information moves into an AI workload.
Some strong starting points include:
Each use case should begin with a measurable business problem. A pilot focused on equipment knowledge, for example, could track whether employees locate approved information faster. An asset-monitoring project could measure whether earlier warnings reduce unplanned downtime.
These focused outcomes make it easier to decide whether an AI project deserves your investment.
Energy companies operate a mix of newer business applications and highly specialized SCADA and EAM/CMMS platforms that have been in place for years. Replacing these systems all at once can create more risk than value. A platform may look outdated from an IT perspective but still support a process the business simply cannot interrupt.
Microsoft Azure can provide an integration layer between those systems and new AI capabilities.
This approach changes the first AI question from “What should we replace?” to “Which existing information will help us make a better decision?”
AI applications are only as useful as the information supporting it. For those in the energy sector, the real challenge is preparing data that lives across separate environments.
Microsoft’s current architecture guidance for Azure AI workloads starts with a data layer that gathers raw information before cleaning and organizing it for the AI application. The architecture then retrieves trusted context when a user asks a question, which helps the application produce responses based on approved business information. For an energy company, that context might come from maintenance history or operational manuals. The original source system does not necessarily have to disappear. Azure can support a controlled architecture around the information that the AI workload actually needs.
Microsoft Fabric can further support real-time energy use cases. Microsoft’s energy grid reference architecture shows how organizations can stream information from generation facilities and smart meters, then combine those events with asset data from an ERP system. Machine learning helps forecast potential energy deficiencies or demand spikes without requiring every source system to be replaced first.
The practical value comes from unifying usable data while allowing existing systems to continue performing the jobs they were designed to do.
A rip-and-replace project often forces the organization to move critical workloads before value has been proven. A hybrid Azure strategy creates another option.
Microsoft Azure Arc can extend management and governance to resources that remain outside Azure. Azure IoT Operations can help connect operational environments at the edge, where energy data is often generated. Microsoft highlights Chevron’s use of Azure IoT Operations to monitor operations remotely and make dynamic adjustments, reducing the need for routine physical inspections.
This type of architecture is relevant for Tulsa energy companies because many operational workloads cannot move directly to the public cloud. Equipment may need to continue working when connectivity is limited. Some applications may have demanding latency requirements while other systems may remain on-premises because replacing them would interrupt the business.
A hybrid approach lets the company modernize around those realities rather than ignoring them.
Poor data doesn't become more trustworthy simply because an AI system can access it. Tulsa energy companies should first understand which information supports the use case and whether that information is accurate enough to guide a business decision.
The initial assessment should answer a short set of practical questions:
Security controls need to follow the information as it moves into the AI architecture. Start with visibility into who can access source data and how AI output is used. Monitoring also matters after deployment because model behavior and business data can change over time.
These requirements aren't reasons to delay every AI initiative, but rather reasons to begin with a contained project where risk can be measured and managed.
Tulsa isn't approaching energy innovation as a choice between its industrial past and a technology-focused future. The city’s current momentum shows that both can support each other.
The University of Tulsa continues to expand research in energy and subsurface engineering. Its 2026 National Academy of Engineering event brought together industry leaders and national laboratories to examine emerging energy technologies and unresolved technical challenges. The city’s workforce is also beginning to support the physical infrastructure behind AI. Crusoe expanded its Tulsa manufacturing operations in 2026 to produce electrical equipment for data centers, adding more than 300 local manufacturing jobs across its two sites.
For local energy companies, the practical lesson is clear. AI adoption does not have to begin with a disruptive replacement program.
Tulsa’s energy industry already knows how to adapt without abandoning the expertise that made it successful. The strongest AI strategies will follow the same path.
Want to get started? We're here to help.
Yes. A hybrid architecture can keep selected workloads on-premises or at the operational edge while using Microsoft Azure for integration and AI services. This can be useful when an application has strict availability requirements or cannot be replaced without disrupting operations.
A strong first use case solves a narrow problem with reliable data and a measurable outcome. Searching approved maintenance knowledge is often more practical than immediately trying to automate an operational decision. The organization can measure time saved while keeping employees responsible for the final action.
Not necessarily. A legacy system may continue operating while Azure receives selected data through an integration layer. Whether a system should eventually be replaced depends on its security and reliability. AI modernization should not be used to hide a system that has become unsafe or unsupported.