usdaed

Nvidia’s strategic pivot towards data center solutions has proven remarkably successful. The inherent parallelism of GPU architecture aligns effectively with the computational demands of AI model training and inference. This has resulted in a commanding market share – currently estimated at 92% according to IoT Analytics – and a corresponding revenue stream that has significantly outpaced industry averages. Recent commentary from CEO Jensen Huang suggests a backlog exceeding $500 billion, with projections of at least $1 trillion in revenue from Blackwell and Vera Rubin chips by the end of 2027. While such figures warrant scrutiny, they underscore the current strength of demand.

I’ve been having a look, and five of them, well, they look… reasonable. Not necessarily screaming bargains, mind you. I’m not one for screaming. But worth a closer inspection, certainly. Let’s have a poke around, shall we?

And right now, these analysts are muttering about DraftKings (DKNG 4.92%). A rather peculiar name, isn’t it? Sounds like a villain from a penny dreadful. But this DraftKings, it seems, might be worth a peek. BMO Capital, a bunch of chaps in suits, have scribbled a target price of $50 on it. Double what it costs now! Even if they’re only half-right – and analysts are rarely entirely right, mind you – it could be a jolly good year for anyone who buys a few shares.

It is in such periods of quietude that a discerning eye may find opportunity. The fervor surrounding artificial intelligence has cooled, leaving behind a landscape littered with overvalued promises. Yet, beneath the surface, genuine potential persists, awaiting the patient investor. Two companies, in particular, warrant consideration, not as mere vehicles for speculative gain, but as studies in the evolving relationship between man and machine.

They call themselves a software company. SMART. Software doesn’t require factories, doesn’t require unions, doesn’t require… reality. They’re selling the BRAINS for these unmanned systems – the drones, the buzzards circling the battlefield. They’re not building the hardware, they’re providing the GHOST in the machine. And in this new world, that’s where the REAL money is. They’ve apparently logged over 100,000 missions in Ukraine. A hundred THOUSAND. That’s not testing, that’s baptism by fire. Real-world data, folks, the kind you can’t get in a sterile lab. It’s a gruesome education, but an education nonetheless.

Shares in both companies currently trade below their former heights, prompting a prudent investor to examine the particulars with a degree of circumspection. The question, then, is not merely which offers the lesser risk, but which presents the more agreeable prospect of future prosperity. Is Ford, with its broad appeal, or Ferrari, with its carefully cultivated exclusivity, the more judicious acquisition at this juncture?

This whole private credit thing… it’s not getting the attention it deserves. Everyone’s fixated on the AI hype, which, let’s be honest, feels like a very expensive magic trick. But underneath that, this quiet panic is building in a corner of the market most people don’t even know exists. It’s the kind of thing that doesn’t make headlines until it’s actively ruining your day. These firms are basically lending money to companies that are already a little… precarious. Lots of software companies, apparently. Which, given the current obsession with algorithms, feels a bit like lending money to a robot with a gambling problem.

Nebius, at its heart, is a purveyor of artificial intelligence – a cloud computing enterprise, if one insists on the prosaic. It constructs digital sanctuaries – data centers – and leases space to others, offering computational power to both ambitious corporations and solitary innovators. It is, admittedly, curious to observe entities like Meta Platforms erecting their own digital fortresses to utilize Nebius’s services. However, it speaks to a desire for an asset-light existence, even if it incurs a slight premium. A touch extravagant, perhaps, but undeniably modern.

ASML manufactures the lithography machines essential for producing advanced chips. Broadcom designs the networking silicon and accelerators that allow data centers to process the resulting flood of data. Both are currently enjoying considerable success. However, when one considers not merely their performance, but the price one pays for a share in that performance, a divergence emerges.