Feel free to share your thoughts and preferences
Hey all,
Mods removed my last post because they claimed I posted saying 'i had XXX amount of money what should I do with it'. Despite not mentioning capital amounts and not saying I am looking to be persuaded otherwise. As per their rules there are no memes. I wrote it myself so no AI and despite their claim of no research I chose those specific stocks after having done research.
However, it is fair because I didn't explain why I chose those stocks soI will include my reasoning underneath each stock just so they can see that I have in fact done my research and yes I wrote it myself. I am not a fan of LLM's they make constant mistakes. Doesn't mean that you can't ride the AI investing wave though! I'm just interested in what others who are more serious in the space are doing. I am simply just looking for an open discussion regarding choices you might have made.
I know that everyone is sick to the teeth of hearing about Al infrastructure picks etc. As a very small play I have created my own basket (mostly for fun, it's just to scratch the itch). For those of you invested in this space, please feel free to share your picks and if you disagree with any of mine.
As I said this is a mostly speculative/fun play. The majority of my funds are in an all world ETF and a bit of Google. My reasoning for this is that the absolute crazy hype around this entire sector might make a lot of people money but with excessive hype comes exceptional volatility. This is just my way of scratching that itch in a small way. Also final add on, I chose mostly mid cap companies, not all but most. Thank you and please be polite, this is not a super serious make or break investment basket just a speculative one.
Compute/cloud: NBIS
GPU cloud business built from the ground up for AI, comparison might be AWS but focused entirely on the compute that AI companies actually need. Hyperscaler capacity is still massively constrained so demand is running way ahead of supply, which is showing up in triple digit revenue growth. GPU cloud is a market heading toward $500B by 2030 and i reckon this is a good way to get into it (despite run up).
Connectivity silicon: Astera Labs (ALAB)
Makes chips that stop AI servers from creating bottlenecks when they're talking to each other. They're already qualified into Nvidia's Blackwell platform which means they're in on the next wave of data center builds. The market they're in is growing at around 30% a year and they're one of very few companies actually shipping product into it at scale.
Optical manufacturing: Fabrinet (FN)
Manufactures the optical components that let data move around inside AI data centers fast enough to keep up with modern GPU clusters. They make parts for some of the biggest names in photonics. I don't see as many people talking about them but pull them out and the whole thing slows down. With hyperscalers committing hundreds of billions to buildouts over the next few years, even if they nab a small part of the pie it's worth a lot.
Cooling equipment: Modine (MOD)
As we all know chips run hotter than your mum so Modine makes the industrial cooling systems, chillers and thermal management equipment that keeps data centers from melting. Their Airedale division has landed major contracts with hyperscalers and their data center revenue grew 31% last quarter with an order book stretching five years out. Cooling is obviously pretty essential when you're running hundreds of thousands of GPUs.
Data center MEP/HVAC: Comfort Systems (FIX)
Comfort Systems handles the HVAC, plumbing and power systems that make them actually operational. Their technology sector backlog has gone through the roof on the back of AI capex and this is contracted revenue not speculation. Every major data center that gets built needs exactly what they do.
Site prep + fab construction + electrical: Sterling Infrastructure (STRL)
Basically prep foundation work, and heavy civil construction required to build these massive data centers. Through acquisitions like the CEC Facilities Group, STRL provides and installs the mechanical and electrical systems needed for high-density servers and massive cooling systems required for AI chips. They also dip more than their toes into construction for semiconductor manufacturing sites.
Power generation/backup: Kodiak Gas Services (KGS)
Okay this one I wasn't as sure on. I was thinking maybe BE instead. Having said that they have moved into power supply through their $675 million DPS acquisition which is a big bottleneck for data centres atm. Data centers running AI workloads need reliable power at scale and KGS is positioning right at that chokepoint. Gross margins sitting at 69% with long term contracts already in place suggests the business quality is there regardless of the narrative.
Chip inspection: Onto Innovation (ONTO) - I am playing around with swapping for VRT or APH
Makes the inspection equipment that checks AI chips for defects during manufacturing. Basically quality control. Their main platform is built specifically around the packaging technology inside Nvidia's GPUs so they're deeply embedded in the supply chain. Revenue hit $1B in 2025, they carry no debt, and as chip complexity keeps increasing their equipment becomes more necessary not less.