AI Enters Energy Sector as Power Limits Growth
OpenAI's Gulf partnerships show frontier models moving into heavy industry, with electricity supply now the sector's main bottleneck.
Updated on Nov. 2, 2026 in Artificial Intelligence
Live Poll
Do you think AI will leave you better or worse off in your job?
OpenAI Chief Executive Sam Altman said advanced AI models are now mostly smarter than top scientists, speaking at the opening of the Abu Dhabi International Petroleum Exhibition and Conference in the United Arab Emirates. OpenAI also announced partnerships with G42 and Abu Dhabi National Oil Company; the extraction does not indicate whether these are shipped deployments or announcements.
Why it matters
The partnerships mark the first global deployment of OpenAI's frontier models by an energy company, signaling AI moving from software vendors into heavy industry. Altman's comments, framed against a claim that AI was mostly not as smart as top scientists at the start of 2026, point to a rapidly shifting capability baseline.
The cited figures are token consumption measures, where a token is the basic unit of text a model processes, reported by the company rather than an independent audit. The reported scale is roughly 100,000 tokens per person per month globally, against near zero six years earlier, and about 500 billion tokens monthly for the largest institutional users.
The players
OpenAI
A frontier AI lab whose GPT model family powers ChatGPT, now extending its frontier models into enterprise and energy-sector deployments.
Sam Altman
OpenAI's chief executive, who has led the company's push to scale frontier models and its compute infrastructure.
G42
An Abu Dhabi-based AI and cloud company that builds large-scale AI infrastructure and models in the United Arab Emirates.
Abu Dhabi National Oil Company
The state-owned oil producer of Abu Dhabi, now partnering with OpenAI to bring frontier AI into energy operations.
The details
The partnerships, announced during the Abu Dhabi event, bring OpenAI's frontier models, its most capable AI systems, into an energy company's operations. Altman described physical power supply as the single biggest bottleneck for the sector, meaning the electricity and infrastructure needed to run large AI systems constrains growth more than model design does. The extraction does not describe the specific technical integration or deployment scope.
Timeline
Start of 2026: AI was mostly not as smart as top scientists, according to Altman's account.
Monday, November 2, 2026: Altman spoke at the Abu Dhabi petroleum conference.
The Tech Race
The partnerships extend OpenAI's competition with rival frontier labs into sovereign and state-linked infrastructure, where Gulf states have been investing heavily in AI compute. The extraction does not name specific competing programs in this announcement.
Readers in the UAE are already using ChatGPT at high rates, and the energy-sector deployment would change how an oil producer's operations use frontier AI. The extraction gives no availability dates, pricing, or rollout timeline for the partnerships.
The takeaway
Altman's claim that power supply is now the sector's biggest bottleneck suggests energy capacity will set the pace of AI deployment. Watch for published details on the G42 and ADNOC deployment scope and any independent measurement of the token-consumption figures.
Further reading
For more on the models and companies driving this shift, explore our Artificial Intelligence coverage.
Source note: This article includes information reported by Anadolu Ajansı.
Live Poll
Do you think AI will leave you better or worse off in your job?
