Artificial intelligence remains one of the market’s most significant long-term investment themes. But as spending on data centers and computing capacity continues to expand, a different set of constraints is becoming increasingly important.
The challenge is no longer just securing advanced chips. It’s also securing the electricity, grid connections, transformers, switchgear, and financing required to operate those chips.
Recent regulatory developments, equipment shortages, and growing scrutiny of AI infrastructure spending suggest that the next 30 to 90 days could bring greater differentiation between companies supplying critical infrastructure and those dependent on completing large, capital-intensive projects.
Beyond AI, developments in Chinese technology export policy, government defense spending, and pharmaceutical regulation could create additional opportunities and risks.
My central question for the coming months: Which companies have the infrastructure, financial strength, and execution capacity to benefit from these structural changes—and which may struggle with rising costs and delays?
The Macro Picture: Four Themes to Watch
| Theme | Key Development | Potential Market Implication |
|---|---|---|
| Data Center Power | Rising electricity demand and grid constraints | Greater focus on power availability, generation contracts, and equipment suppliers |
| Technology Trade | China’s proposed AI and solar technology export controls | More uncertainty around cross-border technology access and supply chains |
| Defense & Space | Continued investment in satellite and communications infrastructure | Opportunities for established contractors and specialized suppliers |
| Healthcare | Potential restrictions on compounded GLP-1 drugs | Competitive and regulatory risks for telehealth providers and pharmaceutical companies |
These themes have different drivers and timelines. Some are already affecting capital allocation, while others depend on regulatory decisions or future contract execution.
1. Data Center Power: From Computing Capacity to Electricity Constraints
The rapid expansion of AI data centers is creating substantial demand for reliable electricity.
A large AI facility requires more than servers and GPUs. It also needs access to sufficient power generation, transmission capacity, substations, transformers, and specialized electrical equipment.
These requirements can create long project timelines even when financing and computing hardware are available.
FERC’s PJM Decision Highlights the Challenge
On October 12, the Federal Energy Regulatory Commission approved PJM Interconnection’s proposal to establish a registry for large data-center loads.
However, FERC rejected a separate proposal that would have allowed certain large data centers to face temporary curtailment during grid emergencies.
The decision reflects the difficulty of accommodating rapidly growing electricity demand while maintaining grid reliability and fair treatment of different customers.
It does not eliminate the underlying capacity problem. PJM still faces the challenge of ensuring sufficient generation and transmission resources to meet expected demand.
Why This Matters for Investors
Electricity availability is becoming a more important factor in determining where data centers can be built and how quickly they can become operational.
This creates several possible investment implications:
Power generators: Companies with existing nuclear, natural gas, and other reliable generation assets may have opportunities to negotiate long-term supply agreements.
Electrical equipment suppliers: Manufacturers of transformers, switchgear, substations, and power-management equipment may benefit from sustained infrastructure spending.
Data center developers: Companies unable to secure timely power connections or equipment deliveries could face project delays, higher costs, and slower revenue realization.
My read: Power availability may increasingly influence which AI infrastructure projects advance first. That could shift investor attention toward companies controlling scarce physical assets rather than those simply announcing large expansion plans.
2. Grid Equipment: A Supply-Chain Constraint Worth Watching
Electricity generation is only one part of the challenge. Moving power from generation facilities to data centers requires specialized equipment that cannot always be manufactured or delivered quickly.
Large power transformers, generator step-up units, transformer bushings, and medium-voltage switchgear are particularly important.
Industry reports indicate that lead times for some large electrical components can extend well beyond a year, with certain specialized transformer orders taking multiple years.
These delays reflect a combination of manufacturing capacity constraints, skilled labor requirements, specialized materials, and competition for equipment from utilities, industrial facilities, and data-center developers.
The Potential Investment Implications
Equipment manufacturers with established production capacity and strong order backlogs may benefit from sustained demand.
However, higher demand does not automatically translate into higher profits.
The important questions are whether manufacturers can expand production, manage input costs, fulfill orders on schedule, and maintain attractive margins.
For data-center developers, equipment shortages may create a different problem: capital can be committed to projects long before those facilities begin generating revenue.
My read: I would pay close attention to order backlogs, delivery schedules, manufacturing capacity, and margin trends. These measures may reveal more about the investment opportunity than announcements of future AI spending alone.
3. China’s Technology Export Proposals Add Another Layer of Uncertainty
On October 16, China proposed revisions to its technology export-control catalogue covering 40 technologies, including 23 related to artificial intelligence.
The proposals address technologies involving AI models, intelligent agents, computing clusters, and other strategically important capabilities.
They also extend into biotechnology and photovoltaic technologies.
The public consultation period is scheduled to continue through November 15.
Why This Matters
Technology export controls can influence where companies conduct research, manufacture equipment, establish partnerships, and deploy advanced systems.
For multinational technology companies, additional restrictions may increase compliance requirements and complicate cross-border expansion.
The inclusion of photovoltaic technologies also raises questions about potential effects on solar manufacturing and energy infrastructure supply chains.
However, the proposals are not equivalent to a blanket ban on Chinese AI technology exports. Their ultimate impact will depend on the final rules, licensing provisions, and how companies adapt.
My read: The broader trend is toward greater government oversight of strategically important technologies.
Over the next 30 to 90 days, I’ll be watching whether the proposals become more restrictive, how licensing arrangements develop, and whether companies identify material effects on their operations.
4. Defense and Space: Government Spending as a Separate Growth Driver
Another area worth monitoring is the expansion of military space and satellite communications infrastructure.
Government investment in resilient communications, satellite networks, and space-based surveillance systems supports demand for specialized electronics, communications equipment, and satellite manufacturing capabilities.
Unlike commercial AI infrastructure, these programs are driven primarily by national security priorities and government procurement decisions.
That can create a different demand profile, although contract timing, funding, and execution risks remain important.
Companies and Capabilities to Watch
Established defense contractors may benefit from their relationships with government customers, technical expertise, and ability to manage complex programs.
Smaller companies with specialized satellite, launch, or communications capabilities may also participate.
However, contract announcements should be evaluated carefully. A contract’s headline value does not necessarily represent immediate revenue, guaranteed profitability, or funding that has already been obligated.
My read: Government-backed demand may provide opportunities that are less directly tied to consumer spending or commercial technology investment cycles.
For individual companies, I would focus on funded backlog, contract execution, margins, and cash flow rather than projected market size alone.
5. Healthcare: GLP-1 Regulation Could Reshape Competition
Outside technology and infrastructure, the GLP-1 weight-loss and diabetes treatment market faces another potential regulatory catalyst.
The FDA has proposed excluding semaglutide, tirzepatide, and liraglutide from the list of bulk drug substances that qualifying outsourcing facilities may use for compounding under Section 503B.
The proposal could affect certain compounded-drug business models, although the precise impact depends on the final decision and the applicable legal requirements.
Potential Beneficiaries and Risks
Established pharmaceutical manufacturers with approved GLP-1 treatments could benefit if tighter compounding restrictions reduce competition from certain compounded alternatives.
Telehealth platforms that rely on compounded GLP-1 offerings may face additional regulatory or operational challenges.
However, the effects will vary depending on each company’s product mix, partnerships, compliance practices, and ability to adapt.
My read: Regulatory developments could become an important source of company-specific volatility.
I would distinguish between businesses with diversified revenue streams and those whose growth depends heavily on a narrow set of products or regulatory arrangements.
Companies on My Research Watchlist
The following companies represent different exposures to the themes discussed above. They are research candidates rather than buy or sell recommendations.
| Company | Ticker | Exposure | What I’ll Be Watching |
|---|---|---|---|
| Constellation Energy | CEG | Nuclear generation and power contracts | Contract economics, generation capacity, regulatory developments |
| Vistra | VST | Nuclear and natural gas generation | Power pricing, long-term agreements, margins |
| Powell Industries | POWL | Electrical distribution and control equipment | Backlog, production capacity, gross margins |
| Digital Realty | DLR | Data centers and digital infrastructure | Power availability, development timelines, leasing, FFO |
| Rocket Lab | RKLB | Space systems and launch services | Government contracts, execution, cash requirements |
| Eli Lilly | LLY | Approved GLP-1 treatments | Regulatory developments, demand, manufacturing capacity |
| Hims & Hers Health | HIMS | Telehealth and weight management | GLP-1 regulatory exposure, revenue mix, business-model adaptability |
These companies should not be treated as interchangeable investments.
For example, a power producer with operating generation assets faces a different set of risks from a data-center developer building new facilities. Likewise, an electrical equipment manufacturer may benefit from industry demand while still experiencing margin pressure from labor, materials, or execution costs.
The objective is to identify where economic value may accumulate—not simply which companies are associated with a popular theme.
What I’m Watching Over the Next 30–90 Days
| Time Frame | Development | Key Question |
|---|---|---|
| Next 30 days | China’s technology export-control consultation | Will the final framework materially restrict technology transfers? |
| Next 30–60 days | NERC computational-load standards process | How will proposed reliability requirements affect large data-center loads? |
| Next 30–90 days | PJM regulatory and capacity developments | Will new rules improve clarity around power connections and reliability obligations? |
| Next 30–90 days | Hyperscaler earnings and capital spending updates | Are AI infrastructure budgets continuing to expand, and how are projects being financed? |
| Next 30–90 days | Electrical equipment earnings and backlog updates | Are manufacturers converting strong demand into revenue and cash flow? |
| Q4 2026–Q1 2027 | FDA GLP-1 regulatory developments | How might final decisions affect pharmaceutical and telehealth business models? |
| Ongoing | Defense and satellite contract announcements | Which companies are securing funded work and demonstrating execution? |
The Bigger Picture: Where Could the Value Shift?
The common thread across these developments is that access to critical resources is becoming increasingly important.
In AI, those resources include electricity, grid connections, specialized equipment, and financing.
In advanced technology, they include intellectual property, manufacturing capabilities, and regulatory permissions.
In defense, they include technical expertise, government relationships, and the ability to execute complex programs.
And in healthcare, they include approved products, manufacturing capacity, and regulatory compliance.
These constraints can create opportunities for some companies while increasing costs or limiting growth for others.
But identifying a structural tailwind is only the first step. An attractive industry does not automatically make every company in that industry an attractive investment.
Valuation, balance-sheet strength, execution, and the timing of expected returns remain essential.