1. Two contracts, 1,602 megawatts of new capacity
Bloom Energy (NYSE: BE) announced the two largest contracts in its history on November 5, 2026, and its shares fell about 16% the next trading day, November 6. The orders from NTT DATA and Google totaled 1,602 megawatts of solid oxide fuel cell capacity, yet a broad selloff in AI power stocks overwhelmed the good news for the Bloom Energy stock price.
The first agreement, with the data center operator NTT DATA, covers 900 megawatts of Bloom's solid oxide fuel cell systems. The equipment is intended to serve as the main onsite power source for several NTT data center projects around the world. Bloom described it as the biggest international deal it has signed, with first shipments expected in 2027.
The contract is structured so it can grow to gigawatt scale across the network of NTT Global Data Centers. For Bloom, that scaling option matters as much as the headline figure, because it suggests the relationship could widen well beyond the initial 900 megawatts if the first projects perform as planned.
The second deal came from Alphabet's Google, which unveiled its first data center in Wyoming, a project valued at $17 billion. Bloom will build fuel cells for a 702-megawatt plant to power the site. The plant is also planned to include carbon capture technology, which is meant to trap part of the emissions produced by the onsite generation.
Added together, the two agreements bring 1,602 megawatts of new capacity to Bloom's order book in a single news cycle. For a manufacturer of power equipment, that is a very large volume of demand to land at once, and it came from two customers in the data center industry that are building at global scale.
A megawatt is a unit of power, and the scale here is easier to grasp in context. A single large data center campus can require tens or hundreds of megawatts of continuous supply, so a 900-megawatt commitment corresponds to several major facilities rather than one. The 702-megawatt Wyoming plant is similarly sized to serve a very large site on its own.
Announcements of this kind also tell the market something about how the technology industry is solving its electricity problem. Building power plants and transmission lines takes years, while data center operators want capacity as soon as their servers arrive. Contracts for onsite generation are one answer to that mismatch, and Bloom is positioned to supply it.
2. Why did Bloom Energy stock fall 16% on good news?
Bloom shares dropped roughly 16% on November 6, 2026, as investors sold AI power and energy infrastructure stocks across the board. The selloff followed the results of the United States midterm elections, which left traders worried that changes in policy could slow the construction of data centers inside the country.
That concern lands hard on Bloom because most of its revenue comes from the United States. A company whose sales depend on domestic data center construction is exposed to anything that might delay permitting, approvals or financing for new facilities, even when its own contracts are growing.
The timing created an awkward contrast. The NTT DATA agreement is exactly the kind of international business that could reduce Bloom's reliance on the American market. But shipments under that contract do not begin until 2027, so it offers little immediate cushion against a sudden change in sentiment about domestic projects.
The episode shows how a share price can move on expectations about an entire sector rather than on one company's announcements. When investors reassess the outlook for AI infrastructure as a group, strong individual contracts can be sold along with everything else, at least in the short run.
Sector-wide selloffs of this type are common in industries tied to a single investment theme. When many investors hold the same group of stocks for the same reason, a shift in outlook can trigger selling in all of them at once, regardless of how each company's own order book looks on a given day.
The distinction between contracted demand and realized revenue also helps explain the reaction. An order announced today may take quarters or years to be manufactured and delivered, whereas a change in policy expectations affects how investors value those future earnings immediately. A stock can therefore fall sharply on the same day its backlog grows.
3. How large is Bloom's data center backlog?
The NTT DATA and Google contracts are the latest in a series of large data center agreements for Bloom, which has benefited as artificial intelligence workloads push electricity demand higher. Operators that cannot wait years for a grid connection have looked at onsite generation as a way to bring capacity online faster.
An expanded collaboration with Oracle could reach as much as 2.8 gigawatts, which is more than the NTT DATA and Google deals combined. If that arrangement reaches its full size, it would be the biggest single commitment in the company's pipeline.
To deliver on commitments of this scale, Bloom has set a goal of 2 gigawatts of annual production capacity by the end of 2026. Reaching that figure is a manufacturing challenge as much as a commercial one, since announced contracts only become revenue once the systems are built, shipped and installed.
A backlog is the total value or capacity of signed orders that a company has not yet delivered. It is useful to investors as an indicator of future sales, but it carries execution risk: delivery schedules can slip, customers can alter the size of projects, and the cost of producing equipment can change between signing and shipping.
For a manufacturer, scaling output to meet such orders usually requires investment in factories, supply chains and skilled labor. That is why the 2 gigawatt production goal is central to the story: the company's announced agreements now exceed what it could readily build without a significant expansion of capacity.
4. What is a solid oxide fuel cell and why do data centers want one?
A solid oxide fuel cell generates electricity through an electrochemical reaction rather than by burning fuel, which is what separates it from a conventional turbine or engine. Bloom's systems can be installed at the site of a facility, so a data center can produce its own power instead of relying only on the public grid.
Bloom points to several advantages for data center operators: rapid deployment, high reliability, low emissions, minimal water use and quiet operation compared with traditional power generation. Water use is a particular concern for large computing campuses, many of which already draw scrutiny in the communities where they are built.
The carbon capture element planned for the Wyoming plant adds another layer. It is intended to trap some of the emissions from the onsite generation, which speaks to the environmental questions that large power projects raise when they serve technology companies with public sustainability commitments.
5. What does the selloff mean for Bloom and its investors?
The decline exposes a gap between Bloom's operating momentum and the political risk attached to its main market. The order book is growing quickly, but the share price remains tied to broader sentiment about the AI infrastructure trade, and that sentiment took a hit after the election.
The NTT DATA agreement stands out as the most important piece for geographic diversification. With gigawatt-scale potential and a customer operating facilities worldwide, it offers a route to revenue that does not depend on decisions made in Washington or in individual American states.
Investors weighing the company face a basic question about conversion. Contracts of this size are only valuable if they become delivered systems, and the pace of that conversion depends on factory output, installation schedules and the continued willingness of customers to fund large projects.
Exposure to political risk is not unique to Bloom. Any company whose growth depends on the pace of data center construction, from chip suppliers to builders of cooling systems and power equipment, is sensitive to expectations about regulation, permitting and the availability of local approvals.
For long-term holders, the central issue is whether the demand driving these contracts is durable. Technology companies have committed enormous sums to computing infrastructure, and the contracts with NTT DATA and Google are evidence that those plans include onsite power. Short-term price swings, by contrast, reflect how investors feel about the risks around that spending at a given moment.
6. What to watch next
The nearest test is whether Bloom reaches its target of 2 gigawatts of annual production capacity by the end of 2026. Hitting or missing that mark will show how well the company can turn a rapidly expanding backlog into physical output.
The second marker is 2027, when first shipments to NTT DATA are expected. Those deliveries will indicate whether the international strategy can begin to offset the concentration of Bloom's sales in the United States.
Finally, the market will keep watching how policy toward data centers develops after the midterms. If concerns about domestic construction fade, the selloff in AI power stocks may prove to have been a reaction to uncertainty rather than a change in demand. If they harden into restrictions, companies such as Bloom that supply the sector's power equipment will feel the effect first.
Readers following the story should also look for updates on the Oracle collaboration, which could reach 2.8 gigawatts, and for any further contracts that add to the order book. Each new agreement raises the importance of manufacturing capacity, and each delay in delivery would draw attention to the same constraint.