AI Chips Cool as Power Stocks Heat Up: Pelosi Household Adds Bloom Energy
Chips determine how fast AI can run, but electricity determines whether a data center can operate at all. A new Bloom Energy position disclosed by the Pelosi household and Aggreko’s proposed IPO are putting the AI power shortage back in focus.
As semiconductor stocks pull back, companies that can solve the data-center electricity bottleneck are attracting fresh attention.
The latest catalyst is a congressional filing showing that Nancy Pelosi’s spouse, Paul Pelosi, established and increased a position in Bloom Energy (BE).
According to the disclosure, the Pelosi household purchased:
-
A total of 15,000 BE shares
-
A total of 200 BE call options
-
A $100 strike price
-
An expiration date of June 17, 2027
The transactions took place on July 24 and July 28. Their disclosed value ranges total approximately $3 million to $12 million, although the exact purchase prices and option premiums were not reported. Official House disclosure
The filing became public on August 21, almost a month after the trades. Congressional disclosures are more useful for identifying investment themes than for copying real-time entries.
Why Bloom Energy?
Bloom Energy produces solid-oxide fuel-cell systems that can generate electricity directly at or near a customer’s facility.
Traditional data centers often have to wait for grid connections, transmission upgrades and new substations. In power-constrained regions, that process can take years.
Bloom’s systems can use natural gas, biogas or hydrogen to generate electricity onsite, reducing dependence on the timing of public-grid expansion.
The company is addressing one of the most important issues in AI infrastructure:
Time to power.
For a cloud company, a one-year delay means expensive GPUs, servers, buildings and land cannot begin generating revenue on schedule. Faster onsite power can therefore have real economic value, even when its electricity costs are higher.
BE is beginning to convert the story into financial results
Bloom Energy reported a sharp acceleration in its second-quarter results:
-
Revenue reached a record $1.07 billion, up 165.5%
-
Product revenue rose 215.4% to $935 million
-
Gross margin improved to 33.4%, up 668 basis points
-
Operating income reached $182 million, compared with a $3.5 million loss a year earlier
-
Full-year revenue guidance was raised to $3.9 billion–$4.2 billion
At the midpoint, Bloom expects full-year revenue to grow approximately 100%. Bloom Energy results
That separates BE from many early-stage energy concepts. Revenue, margins and operating profit are now improving simultaneously.
The project pipeline is also expanding
Bloom and Brookfield recently expanded their AI-infrastructure financing framework from $5 billion to $25 billion.
The capital is intended to accelerate fuel-cell deployment for AI data centers globally. Bloom has already worked with companies including American Electric Power, Equinix and Oracle. Reuters
Bloom also expanded its MiTAC partnership in August, adding a fuel-cell microgrid at an AI-server manufacturing campus.
The company said its AI infrastructure segment now covers nearly two dozen customers and approximately 250 MW of contracted capacity. Bloom Energy announcement
The AI power theme is beginning to move from long-term forecasts into contracts and deployed capacity.
Another signal: a temporary-power company is going public
Aggreko has filed for a U.S. IPO and plans to list on the NYSE under the proposed ticker AGKO.
The company provides temporary power generation and cooling equipment to data centers, utilities, mining operations and manufacturers.
Its data-center revenue nearly doubled to $391 million for the year ended January 3. First-half 2026 revenue increased 28% to $1.92 billion, while the company reported approximately $6 billion of secured net revenue. Reuters
A temporary-power provider returning to public markets during the AI boom highlights what has become scarce: power that can be deployed quickly.
Five ways to invest in the AI power theme
1. Onsite generation: BE
Bloom’s main advantage is deployment speed. Its fuel cells can reduce the need to wait for a full grid connection.
The risks include natural-gas costs, execution on large orders and the ability to turn rapid revenue growth into sustainable cash flow. Its systems are lower-emission than some conventional generation methods, but they are not automatically carbon-free.
BE also carries a much higher valuation and greater volatility than traditional utilities.
2. Nuclear and independent power producers: $Constellation Energy Corp(CEG)$CEG, $Vistra Energy Corp.(VST)$VST, $NRG Energy Inc(NRG)$NRG
-
CEG: Large nuclear fleet and exposure to long-term baseload demand
-
VST: Nuclear, natural-gas and retail electricity operations
-
NRG: Greater exposure to gas-fired generation and retail power markets
These companies already own generating assets and produce cash flow. Their main risks include regulation, electricity prices, fuel costs and contract structures.
3. Gas turbines and generation equipment: $GE Vernova Inc.(GEV)$GEV
GE Vernova provides large gas turbines, grid equipment and power-generation technology.
Microsoft, Chevron and Engine No. 1 are working on natural-gas power projects designed to supply AI data centers, with GE Vernova expected to provide the turbines. Reuters
GEV is exposed to large power-plant construction and equipment backlogs, but its deployment cycle is longer than Bloom’s onsite approach.
4. Power distribution and cooling: $Eaton Corp PLC(ETN)$ETN, $Vertiv Holdings LLC(VRT)$VRT
Electricity still needs to be transformed, distributed, backed up and cooled after it is generated.
-
ETN: Electrical equipment, power management and distribution
-
VRT: Data-center power systems, thermal management and cooling
These companies provide the infrastructure surrounding AI computing, although their valuations already reflect significant growth expectations.
5. Next-generation nuclear: $Oklo Inc.(OKLO)$OKLO, $NuScale Power(SMR)$SMR
Oklo and NuScale represent advanced nuclear and small modular reactor technologies.
Their potential is significant, but commercialization, regulatory approvals and construction timelines remain uncertain. Their risk profiles are closer to early-stage technology companies than mature power producers.
Why power could become the second stage of the AI trade
The first stage of the AI investment cycle focused on GPUs and computing capacity.
The second stage is increasingly focused on the infrastructure limiting data-center expansion:
-
Where will the electricity come from?
-
How quickly can a project connect to the grid?
-
Can transformers and gas turbines be delivered on time?
-
How will the facility be cooled?
-
Who will finance the power project?
Texas has already paused certain new data-center grid approvals pending an audit of electricity demand.
The next phase of AI competition may depend on more than access to chips. It may also depend on which companies can secure power first.
Tiger Radar View
The Pelosi disclosure is a highly visible catalyst, but Bloom’s revenue, margins, orders and deployment record matter much more to the long-term investment case.
Four indicators are worth monitoring:
-
Conversion of contracted projects into recognized revenue
-
Whether gross margin remains above 30%
-
Whether operating cash flow keeps pace with profit growth
-
Whether customers sign long-term power agreements
CEG, VST, GEV and ETN offer more mature assets or order books. BE offers greater growth potential but also carries more valuation and execution risk. OKLO and SMR remain longer-duration, high-volatility investments.
The AI power shortage is a credible long-term theme, but these companies have very different business models. They should not all receive the same valuation simply because “AI needs electricity.”
Today’s Poll
Which AI energy route looks most attractive?
A. Bloom Energy and onsite fuel cells
B. Nuclear and power producers such as CEG and VST
C. GE Vernova and gas turbines
D. Eaton and Vertiv power infrastructure
E. Next-generation nuclear through OKLO and SMR
Disclaimer: This post is for informational purposes only and does not constitute investment advice. Congressional financial disclosures are delayed and report transaction-value ranges rather than exact prices.
Which AI energy route are you most optimistic about?(Single choice)Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.
Next-generation nuclear power is redefining clean energy through passive cooling systems that require no human intervention, external power, or massive water supplies to stay safe.
Unlike traditional plants, advanced small modular reactors use natural physics, such as gravity, convection, and specialized heat pipes to cool themselves automatically.
This design physically eliminates the risk of traditional meltdowns, making nuclear energy far safer and more versatile to deploy.
Investors can gain exposure to this technological shift through several U.S.-listed stocks.
Oklo Inc. ( $Oklo Inc.(OKLO)$ ) designs liquid-metal-cooled microreactors that can operate without water, while NuScale Power Corporation ( $NuScale Power(SMR)$ ) pioneers commercial small modular reactors with built-in passive safety features.
@rL @MHh @Shyon @koolgal @icycrystal
That said, $Bloom Energy Corp(BE)$ is definitely interesting. The combination of record revenue, improving margins and onsite power generation gives BE a compelling growth story, especially when grid connections can take years. However, its valuation and execution risk make me less comfortable chasing the stock after a strong run.
I also like nuclear and gas power producers for the longer term, but I think the infrastructure surrounding electricity could benefit across multiple power sources. For me, the key question is no longer just how many GPUs we can build, but how quickly we can power and cool them.
@TigerStars @Tiger_comments @TigerClub
AI电力这条线我最看重的不是“谁的故事最性感”，而是 谁已经有订单、现金流和明确的扩产路径。所以相比OKLO、SMR这种更远期的核能故事，我更偏好 Eaton、Vertiv 这类直接吃数据中心配电、冷却和电力管理需求的公司，兑现路径更短。
Bloom Energy也很有意思，尤其是“time-to-power”这个逻辑很真实：数据中心如果等几年并网，GPU再先进也没法赚钱。但BE现在估值和执行风险都更高，后面要继续验证 合同转收入、毛利率和经营现金流。
CEG、VST我也看好，因为AI最终需要的是稳定的大规模电力，核电和现有发电资产在这点上更有确定性。
一句话：AI第一阶段缺芯片，第二阶段缺电；我更愿意买已经在卖电力基础设施的人，而不是只买远期概念。
I would choose D: Eaton and Vertiv. The AI power bottleneck is no longer simply about generating electricity; it is about delivering, managing and cooling that power efficiently.
Bloom Energy offers exciting growth potential, while CEG and VST provide established generation assets. But Eaton and Vertiv have a key advantage: they can benefit regardless of whether data centers ultimately rely on nuclear, natural gas, fuel cells or the traditional grid.
As AI workloads become more power-intensive, every new data center needs more switchgear, power distribution, backup systems and cooling capacity. In my view, D offers the best balance between structural AI growth, diversified demand and execution visibility, making it my preferred route into the next phase of the AI infrastructure boom.
@Tiger_comments [你懂的]