SMCI被定位为主权AI领域快速、可配置的硬件供应商，拥有创纪录的600亿美元订单储备以及2027财年650亿至720亿美元的营收指引，但面临经营现金流为负和出口管制审查带来的风险。

The first one I want to talk about is Super Micro Computer, ticker SMCI. So, SMCI, they essentially design and build AI server systems. Those are the the racks of hardware that go inside data centers to run AI workloads.

They're built around chips from the likes of Nvidia, AMD, and others. and their data center building block solutions architecture essentially lets customers configure servers modularly rather than having to buy these very rigid prefixed systems.

Now, as governments and enterprises are pushing to build AI infrastructure domestically rather than renting compute from some of these hyperscalers around the world, someone has to physically build and ship the hardware for those data centers. and Super Micro has positioned itself as one of the fastest and more configurable options for doing uh exactly that.

You know, they reported their Q4 and fullyear fiscal 2026 numbers just uh about a month ago in mid August. So 2026 revenue reached just shy of $40 billion compared to 22 billion in the prior fiscal year.

So essentially doubling and their net income for the 12-month period was $2.2 billion. They also have favorable margins. just their Q4 gross margin was around 18%.

Their backlog is also a really important thing to watch here. So they reported $60 billion in new orders during the fourth quarter of fiscal 2026 alone. So that's essentially pushed their backlog to a record high entering fiscal 2027.

They're guiding for revenue of 65 to 72 billion in fiscal 2027. That's a number that came in well above what analysts had been uh modeling for.

You know, kind of taking a 30,000 foot view here, Super Micro has really positioned itself as a launch partner for Nvidia's newest Vera Rubin and Rubin platforms with expanded manufacturing capacity specifically built around advanced liquid cooling.

And this is really important. You know, we are seeing uh that projections that liquid cooling penetration and AI chips could rise from about a third of deployments as of 2025 to over half of deployments by the end of 2026 and about 60% of deployments by 2027.

And Super Micro's uh data center building block solutions approach essentially lets it pre-engineer and ship these rack scale liquid cool clusters much faster than rivals building those more custom one-off configurations.

So it's an interesting company to watch. One risk all note, you know, they are obviously investing heavily in the buildout and in providing these solutions to their clients. Their uh fullear fiscal 2026 operating cash flow was actually negative as they're scaling up inventory and manufacturing capacity to meet demand.

Um and we're also seeing uh the board had disclosed that they're conducting an independent review of certain transactions related to export control uh issues. So, their final results for their Q4 report were preliminary and unudited at the time of the update.

We should be looking to see more updates from management in the coming weeks.

Yeah, Ra, I mean, you broke this one down perfectly. This is one that obviously you have that bearish case of all these issues that keep coming around for this company or sometimes it's not even the company, it's just like ex managers or ex employees and unfortunately does get tied back to super micro.

Now, if I if I was to take one bullish case from what you shared was that growth of $60 billion in backlog in one quarter. Like that is wild. That is a a mixture of all types of AI solutions, but it does kind of give us more of that sentiment that sovereign AI is picking up.

Uh because it's not just your your your consumer-based companies that are building these solutions. We're starting to see nations build more and more AI's tools. Uh so pretty excited to see this one.

Obviously another one that's very controversial uh because some of the financial stuff that it has been put through uh in the past year or so.

We take we took a closer look at Super Micro, a player that kind of that pigs and shovels play um in in this space. People building AI infrastructure are most likely going to give a call to Super Micro one way or another.