Of course, there's the energy question. Uh, can the energy supply required meet the demand. Our next guest is out with a new report projecting up to a 20% shortage of US power for data centers through 2028.Joining us this morning, Morgan Stanley's global head of thematic research, Steven Bird, is with us. Stephen, welcome back. It's good to see you again.>> Thanks so much, Carl. Thanks, Sarah. >> So, is this going to get worse before it gets better.My concern is you can run into so many different snags when ...

Investment & Expansion - Anthropic plans to invest $50 billion in US data centers in partnership with Fluid Stack, focusing on owning, building, and operating the facilities [1][2] - OpenAI's Stargate project is projected to cost $500 billion, with plans for overseas expansion [3] - A lab within Matter will spend $600 billion on domestic data centers [4] Revenue & Profitability - Anthropic projects revenues of $70 million by fiscal year 2028 [5] - OpenAI's annual revenue run rate is approximately $13 billion for 2023 and projected to be $20 billion for 2025 [5][6] - There is a growing consensus that Anthropic may achieve profitability faster than OpenAI due to its capital-light model relying on partners like Amazon, AWS, and Google [4][6] Competitive Landscape - Anthropic's $50 billion investment is smaller compared to OpenAI's $500 billion Stargate project and Matter's $600 billion domestic data center spending [3][4] - Anthropic relies more on hyperscale partners like Amazon, AWS, and Google for data center resources compared to OpenAI [4]

Financial Performance & Shareholder Returns - Chevron aims to grow free cash flow at a 10% compound annual growth rate [3] - At a $70 oil price, Chevron could return 45% of its market cap to shareholders over the next 5 years through dividends and share repurchases [4] - Earnings per share growth is expected to be better than 10% annually if Brent stays above $70 through 2030 [4] - Break-even point to cover capital spending and dividends is below $50 [7] - Free cash flow is expected to triple from 2024 to 2026 at a $60 oil price [7] Production & Capital Expenditure - Production is growing at a 2% to 3% compound annual growth rate [9] - Capital expenditure is being reduced to a range of $18 billion to $21 billion per year through 2030 [8] - An additional $1 billion in cost cuts has been announced [9] - Synergies on the Hess transaction have been increased by 50% from $1 billion to $15 billion [9] Market Outlook & Strategy - The International Energy Agency's updated report shows demand for oil and gas growing to 2050 [13][14] - Chevron is in discussions to build data centers powered by natural gas, targeting large customers and off-grid power generation [19][20][21][22]

Data center is now our largest segment. So almost half of our business 47% of our business approximate there. We are now let's call it roughly about 40% revenue share of the market in the most attractive uh segments of the market.So large hyperscaler you know very much standardizing uh you know AMD technology through our five generations in addition to you know growing enterprise footprint and it's growing it's growing much faster than the other pieces. uh one because the TAM is growing and the other becaus ...

Risk Management - The company is continuously working on managing third-party risk, especially concerning supply chain constraints [1] - Procedures are being put in place to handle potential backlog doubling and further supply tightening [1] Client Diversification - The company has significantly diversified its client base [2] - At the beginning of the year, 85% of revenue came from a single client [2] - Currently, no single client represents more than 35% of the company's backlog, a decrease from 50% last quarter [2] - Tremendous progress is being made on the contracting side [2] Supply Chain Diversification - No single data center provider represents more than 20% of the company's 29 gigawatts of power to be delivered over the next 24 months [3] - The company is focused on diversifying both clients and suppliers [3]

AI & Technology - Hyperscalers resemble the operating systems of the dotcom boom and have the best chance of creating software packages that bring AI to the people [4] - The rapid depreciation of AI chips, with new chips appearing every 18 months that are ten times more powerful, poses a risk to independent companies [7] - The focus is on who will create the software package that can be monetized, with hyperscalers having the best chance [7] Market & Investment Trends - The world is different from the dotcom boom era because most companies understand that many of those companies never generated earnings or revenues [1] - Only Microsoft, Apple, and Amazon recovered from the dotcom boom and became substantially larger companies [2] - The ETF business has changed the investment landscape from mutual funds, offering tax efficiency and lower running costs [9][10] - Private debt relative to GDP is down 2% annually, indicating the private economy is saving money [16] - The dollar's stability suggests that the anticipated economic problems may not be as severe as expected [17] Value Investing - The most important engine of a value company is generating free cash flow and collapsing capitalization [13] - Value companies in mature industries should focus on generating cash flow rather than investing heavily in CapEx [13]

AI Industry Profitability & Investment - Anthropic anticipates breaking even by 2028, contrasting with OpenAI's projected $74 billion operating losses that year [1] - SoftBank's Masayoshi Son may regret selling Nvidia stock, using proceeds to invest in OpenAI [3][8] - The AI trade needs a wider aperture beyond large language models to include sectors like energy, biopharma, and robotics for better investment returns [9][10] AI Applications & Technological Advancements - AI is transforming the Gulf region in both the digital and physical realms [2] - AI for the physical world, exemplified by the Nvidia deal, focuses on creating new battery chemistry catalysts, essential for 80% of manufacturing and the energy industry [4][5] - AI can improve energy efficiency by enabling new battery chemistries to store the 40% of electrons wasted at night in the US [5][6] - Nvidia is rapidly innovating its chips, with new specs and systems like NVQ link being announced frequently [11] - A hybridized scenario involving GPUs, CPUs, and QPUs (quantum processing units) is emerging, with connections between GPUs and quantum computers being established [11][12] Nvidia's Market Position - Nvidia's GPUs and associated software will continue to see strong demand due to AI's expansion into various sectors [10]

Grid Investment and Infrastructure - The electrical grid is the most critical engine in America, with companies investing over $1 trillion (万亿) in the next 5 years to ensure reliable and affordable power for the growing economy [2][3] - Permitting costs can account for 20% of electricity bills, highlighting the need for permitting reform to reduce costs and expedite power delivery [7] - Data centers want to be connected to the grid for reliability and a diverse energy mix, potentially lowering costs and stabilizing the grid [12][13][14] Energy Sources and Technologies - There are currently zero nuclear power plants under construction in the United States, with optimism surrounding small modular reactors, though none are currently operating [5] - The industry needs as much power and as many electrons on the grid as possible, emphasizing the need to accelerate power generation [6] - While government incentives exist for nuclear power development in national labs, proving the technology and building large-scale nuclear plants is crucial [9][10] Regulatory and Policy Issues - Permitting reform is needed, as building a power plant or transmission line in the US can take over a decade, compared to less than three years in China [6] - Bureaucratic red tape, including NEPA, water permits, and litigation, adds time and costs to infrastructure projects, which is unacceptable for customers [7][8] - Congress is called upon to pass permitting reform to accelerate power delivery and lower costs [6]

Financial Performance & Strategy - The company has lowered rates three times in the last 15 months and expects to lower them again in 2026 while growing earnings at over 9% per year [3] - Infrastructure investment is offset by operating maintenance cost reductions, leading to significant cost savings [1] - Improving credit metrics results in lower cost financing, allowing the company to lower rates for customers [2] Infrastructure & Technology - The company has buried its 1,000th mile of power lines, at a million dollars a mile less than a couple of years ago [4] - Technology underpins situational awareness and responsiveness, leading to a 35% reduction in ignitions this year despite similar fire conditions [5] - The company is implementing layers of protection, starting with hardening the system to mitigate fire risk [4] Load Growth & Rate Design - The company is experiencing rate-reducing load growth due to increased compute demand, particularly from AI applications [7] - Residential customers do not subsidize large load; large load pays its full freight [7] - The company has rate designs in place so that residential customers only benefit from new and growing load [9] - Compute data centers are located throughout the service area, supporting companies using AI in their daily business [10]

RT Michelle Ma (@HiMichelleMa)Two of the world’s biggest data center developers have projects in Nvidia's hometown that may sit empty for years because the local utility isn’t ready to supply electricity.We called the developers and utility in question to find out why: https://t.co/kWMbWBr2l7 ...