Chris Camillo’s Bloom and Amazon trade: thesis and risk

The story in three evidence layers

First, the interview establishes what Camillo said. At 01:09 to 01:40 he says he added to Bloom and Amazon and describes a near-eight-figure up day. At 02:42 to 02:58 he contrasts his own earlier, unlevered position with the account he gives of Situational Awareness. Those are attributed claims, not independently verified account records.

Second, company releases establish the business numbers. Bloom reported its 2025 results and 2026 outlook. Amazon reported Q1 and Q2 cloud growth and infrastructure spending. Third, the Situational Awareness 13F establishes that a different manager disclosed a large Bloom position at March 31. It does not establish Camillo’s holdings or the later financing story.

The Bloom thesis is power at the data center

Bloom makes solid-oxide fuel-cell systems that provide onsite electricity, including for data centers. The thesis is simple: AI computing needs power, grid connections can take time, and equipment that supplies electricity at the site may fill that gap. This is a demand thesis about physical infrastructure, not a prediction based on a chart.

The reported numbers make that idea testable. Bloom recorded $2.024 billion of revenue in 2025. Management’s 2026 outlook is $3.1 billion to $3.3 billion of revenue and about 32% non-GAAP gross margin. The same release names the risks: high upfront cost, installation and interconnection delays, supply constraints, financing, and any slowdown in AI data-center adoption.

Camillo’s one-day result does not settle any of those risks. The clean follow-up is whether Bloom converts demand and backlog into delivered systems, revenue, cash, and durable margin.

The Amazon thesis is AWS growth against the bill for growth

AWS means Amazon Web Services, Amazon’s cloud-computing business. In Q2 2026, Amazon reported total sales of $200.6 billion, up 20%, and AWS sales of $42.2 billion, up 37%. AWS operating income was $16.6 billion. Amazon called it the fastest AWS growth in 18 quarters and said both its AI and chip businesses had passed $25 billion annual run rates.

The bill is capex, short for capital expenditure: cash spent on long-lived assets such as data centers, servers, networking equipment, and chips. Amazon’s Q1 filing recorded $43.2 billion of cash capital expenditure, primarily for technology infrastructure and mostly to support AWS growth. It also disclosed about $364 billion of performance obligations, mainly future AWS services under long-term contracts.

More spending can widen capacity and future revenue, but it consumes cash first. Amazon’s Q2 release showed a trailing free-cash-flow outflow of $7.6 billion, driven mainly by the increase in property and equipment purchases tied to AI. That is the trade-off the thesis has to survive, not a footnote to it.

Leverage can end the trade before the thesis is tested

Leverage means using borrowed money or contracts to control more exposure than the cash committed. It enlarges a gain and a loss. It also gives another party a say in how long the position can remain open.

A margin call happens when the lender decides the account no longer has enough collateral. The lender can demand more cash or sell assets. Camillo discusses that mechanism at 04:35 and 06:06 while talking about Situational Awareness, but Reuters said it was unclear whether margin calls preceded the Citadel deal. His exact four-times estimate is commentary, not a verified term of the fund’s financing.

For a concentrated position, the practical question is not only whether the thesis is right. It is whether the position can survive being early. If an ordinary drawdown reaches the lender’s threshold or the investor’s loss limit, the timeline belongs to somebody else.

How to inspect the trade without copying it

Start with the business question, then set the risk. A reader can test power demand, Bloom’s deliveries and margins, AWS growth, Amazon’s infrastructure spending, and free cash flow without knowing Camillo’s entry price or following his position.

• Write the Bloom and Amazon claims as separate theses. One company does not confirm the other.
• Use company releases for operating facts and label management outlook as outlook.
• Choose the maximum loss first, then let that limit decide the position size.
• Treat every personal profit, position, and leverage figure as self-reported until a separate record proves it.
• Do not use a one-day move as evidence that a multi-year infrastructure thesis is finished.