Citi analysts highlight that OpenAI (stock code: OPAI.PVT) plans to partner with chip manufacturers such as NVIDIA (NVDA), Broadcom (AVGO), and Advanced Micro Devices (AMD) to deploy its promised massive computing power; to fulfill this commitment, the company needs to invest over $1 trillion in the next five years. OpenAI's recent collaboration agreements with these three companies include an ambitious pledge to utilize their chips to provide 26 gigawatts (GW) of computing capacity. This scale of computing power is close to the total electrical demand of New York State during summer peak usage. Citi estimates that to achieve 1 gigawatt of operational computing capacity, an investment of $50 billion is required in computing hardware, energy infrastructure, and data center construction. Based on this assumption, Citi analyst Chris Danely stated in a report sent to clients this week that OpenAI's capital expenditures could reach $1.3 trillion by 2030. It has been reported that OpenAI CEO Sam Altman has presented even bolder targets internally. According to a report from The Information at the end of September, Altman indicated that the company plans to deploy 250 gigawatts of computing power by 2033, implying related costs could reach as high as $12.5 trillion. However, it remains uncertain whether OpenAI can secure adequate funding to support these goals. Citi predicts that although OpenAI's costs will soar above $1 trillion, by 2030, its revenue may only reach a fraction of this amount—$163 billion. This severe imbalance between costs and revenue has heightened Wall Street's concerns about a market bubble. The stock market's surge to historic highs this year has been largely attributed to investors' optimistic expectations surrounding the artificial intelligence (AI) sector. Prior to the latest collaborations with chip manufacturers, OpenAI had already committed significant investments towards the global AI infrastructure. In September, the company announced a $300 billion partnership with Oracle (ORCL) as part of its U.S. "Stargate" AI infrastructure project, which aims to provide 10 gigawatts of computing power. Additionally, OpenAI has partnered with NVIDIA to launch more "Stargate" overseas infrastructure projects in the UAE and Norway. The company has also pledged an investment of $22 billion to CoreWeave (CRWV), an AI data center service provider, to acquire data center computing capacity. The intricacies of investment networks among leading industry players have sparked market concerns over whether AI demand is overestimated. Bernstein analyst Stacy Rasgon wrote in a report on October 6, "Sam Altman (CEO of OpenAI) could either plunge the global economy into a decade-long recession or lead us toward an ideal future, and currently, we have no way to determine which outcome it will be." Aside from funding issues, there are concerns over whether U.S. power infrastructure can timely scale up to meet the energy demands of the latest AI projects—if power supply cannot keep pace, OpenAI's investments will struggle to yield actual returns. Nonetheless, if OpenAI can achieve its goals, chip manufacturers may reap substantial profits. Bank of America analyst Vivek Arya predicts that if the partnership agreements are fully realized, NVIDIA could generate revenues as high as $500 billion from OpenAI. Bernstein's Rasgon estimates that Broadcom could potentially earn over $100 billion through its collaboration with the ChatGPT developer.