First off, I should explain that I retired in 2010 so a lot of things might have changed since then. But here’s the way things were, not just for Boeing where I worked, but for a lot of big companies.
My full time position for the bulk of my career was an analyst. I did not cut code. When I started at the company around 1980 Boeing was still writing their own applications. COTS had not yet really arrived, and even the off the shelf software that was available would fall far short of Boeing’s requirements. My job was to interview expert business folks from various organizations and write specifications for the coders so that they could build applications that would address the needs of the customer. In this job I wrote specs for more than one inventory system (i.e., tool inventory is just a lot different from parts, and there are differences in requirements for raw material than parts, etc.). The reason I’m relating this is to let you know that I learned a lot about Boeing’s general approach to inventory management.
Later in my career I became an enterprise architect. This job was very forward looking. We were the group that set IT standards for the entire company (more or less). One of the projects I worked on was helping to manage Boeing’s transition from Big Blue (IBM) centralized computing to a network of UNIX servers. Boeing is a big place. This was a massive infrastructure project.
Up until this project Boeing had been a dedicated IBM customer. There was even a common saying back then, “No IT director ever lost his job for buying IBM.” But this was open for bidding. It was a “green field.” In the end, HP (it was just one company then) won the contract for servers.
What I’m getting at here is that long-term vendor relationships was the way Boeing conducted business. Didn’t make any difference if it was airplane parts or IT components. Boeing was not the only company that operated like this. From people I talked to at various IT conferences it as pretty evident that this was common MO for a lot of large companies. Certainly, not all. Some companies still liked the game of pitting vendors against each other for everything from office supplies to whatever they bought that went into their products.
I will say that eventually, that changed to some extent after Boeing merged with McDonnell Douglas. With the merger, the former MacDac CEO got a seat on Boeing’s C-suite. At first he was COO and later he became CEO. His philosophy was that the bottom line was the only thing that mattered. He was focused on cost cutting. He considered QA an inhibition to productivity. I won’t go on with my Stonecipher rant, but I will say that nearly 100% of Boeing’s troubles since 1997 can be traced back to Harry Stonecipher. Boeing didn’t really have many troubles before then.
Anyway, as previously noted, I retired in 2010, so if there was a total shift in IT purchasing after the merger I wasn’t around to see it.
To @Graydrake’s point, I don’t think that is a contradiction to what I said earlier about long-term vendor relations. In fact, IMO, the story I just related is exactly consistent with his comments.
@Iamnzane probably has more recent insight than I do with respect to data center operations. I would just steer you to his assertion that “compatibility, delivery schedule, cost, and quality requirements will steer them to the chosen vendor.” SMCI is focused on these attributes. There are already a gaggle of low cost Asian server suppliers. SMCI does not even try to compete with these companies. If nothing but price matters, then SMCI will not win the bid. But, at the same time, SMCI has the lowest time to delivery, time to online (as much as a year sooner in some cases) and lowest total cost of ownership in the business. These things matter to most of the companies with huge data centers. I will add, SMCI has over 250 stock servers in their catalog, far exceeding any of their competitors. Additionally, they welcome customer engineers to bring their specific use cases to them so that they can customize their product in order to exactly address customer requirements. They do not just rely on their DLC (liquid cooling) as the only discriminator. I have no insight or experience in this, but I would speculate that if a customer wanted different configurations to address different needs, SMCI would be only too happy to accomodate them.
I can’t really add anything to @intjudo’s comment. When I was at Boeing we did bring in a lot of EMC (now part of Dell) storage. They became a dedicated vendor for storage. I have no idea if we ever had an incident like the one @intjudo related.
In any case, IMO SMCI is a very good investment at present. That is unlikely to last forever, and a timely exit will not be an easy decision. Personally, I find selling decisions much harder than buying decisions.