SRESempra
AI adoption · Q1 2026 earnings call
UtilitiesExploring
6
extracted from this call
4 / 5
quantified with specifics
Not disclosed
no breakout in this call
AI was discussed exclusively as a demand-side driver of load growth in Texas, specifically through data-center construction requiring Oncor's transmission infrastructure buildout. Management cited 271 gigawatts of data-center-related load in Oncor's queue and framed the U.S. as engaged in an 'arms race' around AI infrastructure. No AI products, internal AI use, or AI-specific investments by Sempra itself were discussed.
Beneficiary
See full leaderboard →11/ 100
28
stage: exploring · max spec: 4
0
no quantified disclosure
0
no adoption scopes
6 AI mentions from this call.
Extracted verbatim from the SRE Q1 2026 earnings call transcript. Speaker, section, and specificity tier surfaced for each mention.
- T4Q&A· CEO· Customer demand signalDo you view the quality as comparable to the prior 39 kind of high-confidence number? And do you envision any rule changes that could move that number lower?
“For Oncor's 2026 RTP filing, we submitted 102.22 gigawatts of load—that is load 75 megawatts or higher. We also submitted 5.2 gigawatts of medium-sized load, which is load greater than 25 megawatts but lower than 75 megawatts. Together, that constitutes what is known as substantiated load pursuant to the ERCOT compliance plan. To your question about the quality variance between the 38 and the 127, they are effectively the same.”
— E. Allen Nye, SRE earnings callERCOT - T4Q&A· CEO· Customer demand signalDo you view the quality as comparable to the prior 39 kind of high-confidence number? And do you envision any rule changes that could move that number lower?
“our total queue right now for large load is at 289 gigawatts, of which 271 gigawatts are data-center related.”
— E. Allen Nye, SRE earnings callERCOT - T3Q&A· CEO· Customer demand signalI am curious about your confidence level—what you are seeing on the ground, physical progress, and any limiting factors that you see realistically for the load growth outlook?
“There is 164 gigawatts installed—nameplate—versus an 85.5 gigawatt peak. But we have, as a state, about 450 gigawatts or so of generation in the queue—somewhere within that 164 gigawatts trying to connect to us in some manner right now.”
— E. Allen Nye, SRE earnings callERCOT - T3Q&A· CEO· Customer demand signalDo you view the quality as comparable to the prior 39 kind of high-confidence number?
“we keep using this term of 127 gigawatts of large-load customers—the bottom line is it is going to lead to higher levels of capital spending in Texas.”
— Jeffrey Walker Martin, SRE earnings call - T2Q&A· CEO· Customer demand signalDo you view the quality as comparable to the prior 39 kind of high-confidence number? And do you envision any rule changes that could move that number lower? Just maybe talking about how you envision the timeline for converting that into CapEx.
“the terminology I have heard a lot in our industry is the United States really is in the middle of an arms race. And it is a race to build the infrastructure that we are talking about, artificial intelligence, and continuing to improve America's standing as a technology leader in the world.”
— Jeffrey Walker Martin, SRE earnings call - T2Q&A· Other· Customer demand signalwe heard earlier today from an IPP in Texas who was more cautious on the physical ability for all that data-center capacity to come online. I am curious about your confidence level—what you are seeing on the ground, physical progress, and any limiting factors that you see realistically for the load growth outlook?
“we heard earlier today from an IPP in Texas who was more cautious on the physical ability for all that data-center capacity to come online.”
— David Keith Arcaro, SRE earnings call
What management wouldn’t quantify.
Analyst questions where management declined to share a specific number. The pattern of refusals is often as informative as the disclosures.
- Management did not quantify the direct revenue or earnings impact attributable to AI-driven data-center load specifically versus other large-load customers.
- No timeline was given for when AI data-center load would begin converting to actual CapEx within the five-year plan versus beyond it, despite a direct analyst question on this point.
- No discussion of any internal AI adoption, AI tools, or AI-driven productivity initiatives at Sempra or its subsidiaries.
Compare with peers.
Other companies in the same sector and at the same AI adoption stage.
Same GICS sector, all stages
Independent research, direct to your inbox.
Live data tracking and analysis. Deep research that cuts through consensus. Evidence-backed insights.