BE: Received a 2.6 billion order last week, but are institutions hedging with mid-term bearish bets?

On May 20th, $Bloom Energy(BE.US) signed a $2.6 billion energy agreement with Nebius, and on May 24th, it secured a 328-megawatt fuel contract for an AI data center. The stock price hit a 12-month high on May 21st. This kind of continuous positive momentum should theoretically keep pushing higher, but on Friday, institutions placed two completely opposite orders on BE simultaneously—one spent $820k betting it would break through $320 before May 29th, and the other spent $1.48M (combined for two trades) betting it would fall back to $245 or even $225 by early July. For the same stock on the same day, capital behavior is completely split. This is rare and worth discussing in detail.

First, what is BE: Bloom Energy, founded in 2001 and listed on NASDAQ in 2018, produces solid oxide fuel cells. In plain terms, it's a "box that generates electricity directly on-site for customers using natural gas." It doesn't burn fuel; it converts energy through electrochemical reactions, with efficiency over 30% higher than traditional power plants. It can be installed at the base of data centers, supermarkets, and hotels, allowing customers to buy power without connecting to the public grid. Companies like Walmart, Costco, Apple's California data center, and AT&T are its long-term clients, purchasing a three-in-one solution of "independent power supply + backup generation + emission reduction compliance."

Why did this company suddenly explode this year? Because the core bottleneck of the AI era isn't chips; it's electricity.

A single H100 GPU at full load consumes about 700 watts. Nvidia's next-gen B200 at full load approaches 1200 watts. An AI data center easily requires tens of megawatts, with large ones needing hundreds of megawatts. The old U.S. power grid simply can't handle this level of demand—just applying for new power connections from PG&E involves a 5-7 year queue. AI cloud companies like Nebius, CoreWeave, and xAI can't wait that long, so they have to find local power generation solutions, and BE's SOFC is currently the most mature option. This is the same script as the shortage of sleepers when building railroads in the 19th century, the shortage of rocket engines during the space race in the 1960s, and the shortage of fiber optics during the internet boom in 2003—the main industry takes off, but it's the infrastructure link that becomes the bottleneck, and the company standing at that bottleneck reaps all the benefits.

The timeline is clear: At the end of April, BE released its Q1 earnings, with revenue up 30% year-over-year, catching sellers' attention. On May 19-20, the NBIS agreement was finalized, with the $2.6 billion figure being BE's largest single order since its IPO. The stock price hit a 12-month high on May 21st. It opened at a historical high of $322 on May 22nd before being pushed back down to $302. On May 24th, it signed another 328MW contract. This over-a-month trend is a textbook example of a "theme stock + blockbuster order + multiple catalysts" combo, with almost all research reports raising their target prices.

But those three institutional orders on 5/22 offer another perspective.

That $820k short-term Call at $320, I don't think is bullish—it's institutions knowing that details of the NBIS agreement would be released pre-market on 5/22, betting ahead at 12:31 ET that "it could break the $320 psychological level that day." As a result, BE gapped up and surged past $320 that day but was pushed down due to heavy profit-taking. This Call will likely expire worthless soon (expires 5/29, current price $302 is still 6% away from $320, basically hopeless within a week). In other words, it's a probability bet that has already played out, not a directional bet.

The real directional bets are the next two trades. At 12:49 ET, a $636k Put was bought at the $245 strike price. At 13:56 ET, an additional $848k Put was bought at the $225 strike price, forming a medium-term bearish calendar spread for 6/5 + 6/26, with DTE14 + DTE35, totaling $1.48M. This is institutions betting that "after the intensive realization of the AI theme by the end of May, there will be a pullback in June." The logic isn't complicated either—BE is already in the $300+ range, while its 12-month average is around $200-250. The $245 and $225 strike prices correspond to "returning to the monthly moving average + falling back to the previous high-volume trading zone."

This is a typical "theme stock top game" pattern. In the short term, momentum from the May 24th 328MW contract, the quantum theme, and the AI power shortage narrative may push BE higher for another wave. But mid-June is a verification window—if no new orders are signed in Q2, or if new order sizes shrink, the institutional Put path becomes viable once sentiment cools.

(1) If you are already long on BE, do not add more now. $300+ is a 12-month high, and the stock price has already priced in the NBIS mega-order and AI power shortage expectations. Institutions putting $1.48M into Puts is real money betting on a pullback; retail investors shouldn't ignore the size of this counterparty.

(2) If you want to open a new position, waiting for a pullback to the $260-280 range is much safer than chasing highs. This is one level above the institutional Put strike price of $245, and it's also the high-volume trading zone before May 19th. Support should be relatively solid if fundamentals haven't changed.

(3) For pure short-term trading, observe whether $320 can hold before May 29th. If it fails to hold, that confirms "the good news is fully priced in," and you can follow with a small Put position for a short-term pullback play, but strictly control the position size to below 2%—trading against AI theme stocks is a double-edged sword; don't get caught halfway by a new order announcement.

I think BE's story is real, and SOFC's position in the AI power shortage narrative is irreplaceable; it's a good business in the long run. But a good business ≠ a good stock; price determines everything. The institutions' net short of $664k on 5/22 isn't bearish on BE; it's bearish on "the $300+ price level."