Morgan Stanley "AI Power Supply Summit" Highlights: U.S. Data Centers Prefer Going Off-Grid, Energy Storage Becomes Standard
```
Morgan Stanley's Second AI Power Summit reveals the severe challenges facing US AI infrastructure development.
According to Hard AI, Morgan Stanley's second “AI Power Summit” was recently held in New York, with attendance surging fourfold compared to last year, reaching nearly 400 people, highlighting the sharp increase in market attention to AI infrastructure power supply issues. The conference relayed three core signals:
First, US data center developers face serious power shortage risks, with supply-demand gaps expected to reach 10-20% by 2027-2028;
Second, due to political and execution risks, demand for off-grid solutions has surged, with natural gas generator sets and energy storage systems becoming standard;
Third, while transaction conditions for electricity and data center development have improved, 2026 is seen as the “year of execution,” and project execution capability will determine the performance of related stocks. Investment opportunities are focused on “fast power supply” solutions, bitcoin-to-data center projects, and the natural gas industry chain.
Morgan Stanley emphasizes in its report that investments should target the “bottleneck segments of AI deployment” as these constraints will become key determinants of stock performance in 2026.
Impending Electricity Shortage Crisis
Morgan Stanley’s analyst team confirms that US data center developers will face a 10-20% shortfall in power supply over the next few years, and summit feedback further reinforces the possibility and severity of this forecast.
The critical time nodes for this issue are concentrated in 2027 and 2028, and most current new data center project agreements are set for delivery between 2028 and 2030, which means that 2026 and 2027 may see a supply vacuum period.
A Google executive revealed the urgency of demand at the summit:
Our 7th and 8th generation TPU utilization rates are at 100%, showing how high the demand is. While everyone wants to use the latest products, they are willing to accept any product available, which shows extraordinary demand. Worryingly, I don’t think the speed of supply catching up with demand will be as fast as we hope.
Surging Demand for Off-Grid Solutions, Energy Storage Becomes Data Center Standard
With the surge in data center power demand, what was once a simple supply and demand issue has evolved into a sensitive political game. Morgan Stanley points out that developers face increasingly severe challenges in grid access and fear political backlash due to driving up residential electricity prices. As a result, a fundamental shift is happening in market logic — no longer relying solely on the grid, but seeking self-sufficiency with off-grid solutions becoming the new trend.
In consideration of avoiding political and execution risks, the market’s preference for off-grid solutions is rising. Although many issues (such as water use, land occupation, and noise) can be solved through clever project structuring, political concerns remain. This is prompting developers to turn to off-grid solutions, mainly focused on gas turbines, reciprocating engines, fuel cells, and supporting battery storage.
Specifically, popular off-grid solutions are centered on natural gas turbines, reciprocating engines, and fuel cells, combined with battery energy storage systems. The conference revealed that onsite power generation products are transitioning from temporary solutions to permanent power supplies, as microgrid contract terms are being extended, with recent negotiations reaching 10-15 years or more.
The Morgan Stanley summit also revealed an important trend: in data centers, battery storage may become the new “built-in power” standard by 2026.
Morgan Stanley believes that having battery storage can be seen as a new “political access threshold,” helping data center developers alleviate concerns that their projects may drive up electricity prices during peak periods.
Improved Transaction Conditions but Persistent Execution Risks, 2026 as "Year of Execution"
Morgan Stanley states that because demand for these products significantly outweighs supply — and this mismatch is widening — power and data center developers may secure better transaction terms, such as project risk allocation. This is a positive development given major project execution risks, from the supply chain to workforce challenges. 2026 will be the “year of execution,” and stock performance of fast power supply solutions will increasingly be driven by the success of power and data center development projects.
Against the backdrop of nonlinear growth in AI computing power, whoever solves power supply fastest will earn high premiums. The report predicts the US may face data center power shortages as early as 2026 or 2027. Therefore, the market’s core trading logic is now focused on the “time to power” metric.
2026 will be seen as the ‘year of execution,’ and stock performance will mainly depend on the success of developing power and data center projects. Even if we manage to deploy enough data center infrastructure (which is considered highly challenging in the US), demand for computing power will still exceed supply.
Morgan Stanley continues to be optimistic about the mass reconfiguration trades from bitcoin to data centers in 2025, 2026, and beyond. These trades offer significant time advantages and lower — though far from zero — execution risks. Natural gas may play a key role in meeting growing power demand, further supporting an already bullish market outlook for the coming years.
Morgan Stanley advises investors to “own the bottleneck” (investment opportunities in AI deployment constraints), focusing on bitcoin mining conversion (offering fast power access), the natural gas value chain (especially companies providing “wellhead to socket” bundled solutions, such as EQT and WMB), and off-grid solution providers. In this AI arms race, electricity is not just energy, but a qualification for survival.
This article comes from WeChat Official Account "Hard AI", for more cutting-edge AI news, please visit here

Risk Disclosure and DisclaimerThe market has risks and investment needs caution. This article does not constitute personal investment advice and does not take into account individual users' specific investment goals, financial conditions, or needs. Users should consider whether any views, opinions, or conclusions in this article fit their specific circumstances. Investing based on this is at your own risk. ```