energy

AI investment is shifting focus from broad enthusiasm to essential data center infrastructure. Investors are prioritizing companies with substantial computing power and facilities, as highlighted by Goldman Sachs’ “flight to quality.” AI workloads will significantly boost data center capacity and energy demand, with infrastructure limitations now shaping AI strategy and geographical site selection. This pragmatic phase emphasizes foundational hardware and energy solutions over experimental software.

Meta is partnering with nuclear power providers Vistra, TerraPower, and Oklo to secure significant energy for its AI ambitions, including the “Prometheus” supercluster. These collaborations aim to supply 6.6 gigawatts by 2035, supporting data centers and AI development. This move highlights the growing trend of tech giants relying on nuclear energy to power their expanding AI infrastructure and maintain global leadership in the field.

Despite U.S. restrictions on advanced chip exports, China is making strides in AI development by leveraging domestically produced chips and strategic advantages. Huawei’s cluster approach links multiple chips to rival Nvidia’s performance. China’s access to affordable energy, driven by investments in renewables and nuclear, supports the high power consumption of these clusters. Government subsidies further incentivize the use of domestic hardware. The long-term challenge remains bridging the performance gap as Nvidia and TSMC innovate, given ongoing technological restrictions.

ConnectM Technology Solutions acquired Geo Impex & Logistics, gaining control of 76 acres near key transportation hubs in Odisha, India. This strategic location, approved for industrial development, will host a multimodal logistics park and an AI-enabled Energy Intelligence Data Center (EIDC). ConnectM aims to integrate its AI energy technology, including Amperics battery technology and Keen Labs’ software, to create a sustainable platform optimizing logistics, power, and data convergence. CEO Bhaskar Panigrahi sees this as a blueprint for sustainable industrial growth.

Dynamix Corporation III raised an upsized $175 million in its IPO, potentially reaching $201.25 million with additional shares. The SPAC targets acquisitions in AI, energy, or digital assets, aiming for companies valued at $1 billion or more. Advisors from Prologis and Nvidia will aid in identifying suitable targets. This debut occurs during a cautious recovery phase for the SPAC market after a period of decreased activity and scrutiny, requiring Dynamix III to demonstrate strong valuation and synergies to investors.

China’s electricity consumption surpassed 1 trillion kilowatt-hours in July, a new record, driven by extreme weather and economic recovery. This figure equals Japan’s annual consumption and rivals the US’s Q1 consumption. Overall electricity consumption from January to July rose 4.5% year-on-year to 5.8633 trillion kilowatt-hours. Primary, secondary, and tertiary industries saw increases of 10.8%, 2.8%, and 7.8% respectively. Residential consumption also increased by 7.6%. China’s power supply system remains stable despite peak demand.

CNPC launched the Kunlun large language model, a 300-billion-parameter AI, marking a significant advancement for the energy giant. Developed with partners including China Mobile and Huawei, it’s the first of its kind approved for China’s energy sector. The model supports 100 application scenarios and includes a visual AI model built with Huawei, leveraging Ascend hardware. The Kunlun AI Middle Platform, operating on a China-made infrastructure, supports training and deployment, with computing power reaching 1950P.