I just wrapped up attending TexAI, the annual public conference of the Texas Lyceum, which covered, “Where Artificial Intelligence Meets Policy”. The one day conference in Austin had a packed agenda spanning infrastructure, policy, and practical presentations from a number of artificial intelligence (AI) companies operating in Texas today. AI is coming, and from how I see things the decision in front of us as Texans is to determine where and if we want to participate in the AI economy. Texas has the energy, water, and economic resources to be the leader in AI infrastructure. We also have the current workforce plus the educational programs to position Texas as a major player in the evolution of AI applications. It is up to us to decide if we embrace the AI economy, or if from my perspective, put the “Texas Miracle” at risk over the span of the next twenty years.

First, on energy infrastructure, being the home to the data centers which will power a generation of AI models presents an opportunity for hundreds of billions of dollars to be invested in the state. The five-hundred billion dollar project Stargate is already underway with the first billion dollar facility being built in Texas. While some have doubted the size of the project, in an interview at Davos with Satya Nadella, the Microsoft CEO said clearly, “all I know is I’m good for my eighty billion” which is their commitment out of the overall capital for project Stargate. Adding this base load to our energy grid under agreements where the facilities will bring their backup generation online during gridwide shortages has the potential to make ERCOT the most resilient grid in the world.

Next, tied to electricity for a data center is water, in order to keep the systems cooled to proper temperatures. Data center water usage is measured by an industry term, WUE, Water Usage Effectiveness. For the average facility, the usage is 1.8 liters per kWh per day (L/kWh), however, large scale facilities can be an order of magnitude more efficient, with Amazon reporting 0.19 L/kWh. Texas can 10x the data center computing capacity, while maintaining similar water usage, through the process of average facilities being upgraded or consolidated into hyperscale state of the art operations. If we do not permit and construct the new facilities, our businesses, state agencies, schools, and municipal governments will be stuck using inefficient, expensive, and outdated platforms.

Once the infrastructure is running, an AI operating system is installed on the computers, which enables the training and usage of AI models by data scientists and software developers. Our universities are offering cutting edge degrees, such as the Artificial Intelligence Multidisciplinary Studies Degree at UTSA. The jobs building AI models and end-user applications are separate from the infrastructure opportunities. These scientists and software engineers can work on computing systems based anywhere on earth, so long as public policy allows them access to do so. One key public policy area is how copyright law applies to building and using an AI model. Due to the nature of copyright law, Texas could decide to use its influence on Federal legislation, or we could choose to sit on the sidelines and see what happens – my perspective is we should look to enable a common sense framework which sets up Texas and the United States to be the clear AI leader.

After the model is online, applications are built to run on it, which is how most of us will experience AI. Examples of common AI general purpose applications are ChatGPT, Claude, Gemini, Copilot, and Grok. There’s also tons of narrow AI and vertical specific AI applications focused on solving specific problems or enhancing specific workflows. Research is happening at Good Systems, an AI research center that spans multiple colleges at UT Austin plus other industry and academic partners, focused on six core projects about the ethical implications of AI. While there are risks related to AI implementation, nations, states, and cities have laws in place today regarding discrimination. There are organizations looking to enhance the ethics laws specifically to cover AI, which will put a new burden on AI infrastructure operators, model builders, and application developers looking to provide AI powered services to consumers or businesses. Because of the global nature of the internet, and the ability of AI processing to provide quick responses from anywhere, if a locale wants AI built and running in their jurisdiction by scientists and engineers living there, my recommendation is to work with the builders in a collaborative effort, or they will choose to move themselves and to build their platforms elsewhere.

In closing, Texas has a once in a generation opportunity to lead in AI by ensuring a collaborative and clear legal framework, by enabling the construction of AI infrastructure through an efficient permitting and construction process, by enabling colleges and universities to develop the degree programs and to do the research needed so the workforce is educated and ready, and lastly by stewarding a fair balance between the stakeholders in the AI ecosystem in a manner the key parties choose Texas as the place for AI. Without all four pillars, Texas risks losing AI investment to states or countries that move faster and more decisively. So, while we have many assets creating an advantage today, it isn’t guaranteed that Texas will be the national or global leader in AI by the year of our tricentennial in 2036.