Rightsized CapStack Capital

Officially sold all my $VRT. 730%+ gain realized in my ROTH IRA. Time to find the next man up! Ideas?

I’ve seen ~5 almost consecutive posts from supposed popular FinTwit accounts who are like “I told you yesterday $CBRS would fall from the IPO run”. My brother in Christ, it doesn’t take a rocket scientist to predict that. Anyone with half a brain and a rudimentary understanding of finance and valuation could figure that out. You are just a FinTard. That is my Ted Talk.

@FL_DAYTR8R I think it just attributable to the recent econ data. But the changes are getting absurd now.

$NNBR can anyone explain why this just keeps going down this week? 😆 $TIC as well. Think they’re just selling small caps based on inflation data the last couple days, but sheesh…..ridiculous

@FL_DAYTR8R It's diabolical.

$NNBR trading like a bottomless pit after exceptional earnings.....terrible

@firstadopter @n0th0stl3tk0v3r Agreed. Touched on it earlier this AM x.com/capstackcapita…

CLOWN SHOW. Everyone who told you Cerebras is the next Nvidia is an idiot.

$SHIM +8% today after starting the day way lower. Really solid Q1'26 results. As revenue ramps sequentially through the remainder of the year and SHIM maintains and/or further expands margins, there will a powerful EBITDA uplift.

@SimplicitasOpes LOL

$SHIM Q1'26 Earnings are out. Backlog of ~$944mm, highest since Q1'24. Adj. EBITDA of $3mm, continues to improve. Gross Margin increased significantly, despite lower YoY revenue. 2.6x Book-to-Burn, its highest ever as a PubCo. +$289mm of new awards in Q1'26 +$174mm of additional awards pending as of May 2026 in Water & Electrical target markets. Really great development. FY26 guidance was reaffirmed: Revenue: $550mm - $600mm (+17% YoY @ mdpt) Adj. EBITDA: $15mm - $30mm (+350% YoY @ mdpt)

At the current valuation, I think there is actually heightened execution risk for $CBRS. As with so many other AI-tailwind driven semi/semicap names that are driven up, investors are pulling forward immense expectations and implying flawless execution and then some. The CBRS thesis right now is underpinned by the OpenAI contract. But that does not negate the fact your competing against $NVDA whose CUDA moat is formidable and keeps customers within the NVDA ecosystem. $AMD has spent 10 years playing catch-up to NVDA with regards to improving its software stack. CBRS revenue is essentially all tied to Saudi Arabia right now, and while that will inevitably change the current valuation is clearly not conducive to the a favorable R/R. Investors are likely bag-holders at this level for the foreseeable future. The valuation would need to (i) compress significantly or (ii) will have to grow into the current valuation some and thus likely trades sideways for a time. It will probably be a mix of both. Given the current industry dynamics, CBRS will most likely always trade at a significant premium valuation, until it gives the market a reason not to ascribe it one. But to be that much of a premium to NVDA is nonsense. The technology is interesting and unique, and hope it succeeds. But imo, there are too many risks that are large enough and not even remotely reflected in the valuation to make it worth considering at this level. CBRS dropping to low 200s or high 100s would make me reconsider.

Passing up on $CBRS at ~$100B MC. The bull case is compelling since they're a genuine architectural-paradigm story in AI compute. Highlighted by a stellar list of customers using their chips: - OpenAI: $20B+ contract - $AMZN: AWS data centers - $META: Llama API for devs - Perplexity: sonar model - $GSK: biomed models - Mistral AI: Le Chat - Notion: enterprise search - AlphaSense: enterprise search All of that diversifies them away from G42 in the UAE which have been the historic rev driver. Regardless, not too sure when I'll look to initiate a position, maybe like the $200 mark? Kinda feel there's some downside risk rn when you layer on top: - $NVDA Vera Rubin volume ramp (H2 2026) - potential Nvidia absorption of Groq - CUDA moat So current valuation looks like all the execution upside is being aggressively priced-in.

There are 2 college pitchers since 2019 that have had: Top Velo >= 98.5 BB% < 5% IP/G >= 5 K% > 35% 2023 Paul Skenes and 2026 Logan Reddemann

@ebarica @DocCPA78 @JackJuliano3 @GagHendo007 @Aksel465 @Lil__Oaty @dayansb1986 @SingleBestIdeas @addictedtostock @j16425646 @ValueHunter7 I accepted the invite.

To $SPCB investors: I created a discord group so we can share information more easily. discord.gg/CzmdGbX9Y @DocCPA78 @JackJuliano3 @GagHendo007 @Aksel465 @Lil__Oaty @dayansb1986 @SingleBestIdeas @addictedtostock @j16425646 @CapstackCapital @ValueHunter7

@Temple_Eight Props to @AtlasShrug1 on this one.

$IPWR our power semi micro cap play delivered a good earnings report this am. They buildsolid state bi directional switches sitting at the same stage as Wolfspeeds switches (they're not competitors). These bidirectional switches allow power to flow in both directions between the grid and facility level energy storage. This is only the beginning. Exciting little company with a bright future.

@BussinBiotech @BlackScholesMan Doesn’t mean it’s a bottleneck tho. Just because a company is in a semiconductor supply chain does not mean it’s a bottleneck.

$ELTK Most underrated bottleneck you say? x.com/JackFarley96/s…

@BussinBiotech $ELTK is a really solid business.

@WisemanCap Holy shit I never turned off my ATM order.

Who is the genius who bought at $386 $CBRS

@BryzonX Ofc!

@CapstackCapital Thanks man

I am long Clearfield, Inc. $CLFD Here’s my thesis: I’ve been researching a lot about the expanding importance of fiber optics in the new age data center as we move into the agentic ai era The recent deal between $NVDA & $GLW basically confirmed to me that high performance optical fiber is the foundation for AI infrastructure However... as more & more optical fiber start to be deployed data centers are hitting a physical "density wall" creating a spaghetti of cables that blocks airflow and traps heat, the #1 killer of GPU performance Because AI depends on light traveling through glass, physical management is now widely considered signal management In the current 800G environment, a single bend that is 1mm too sharp can leak enough light to stall an entire AI training cluster In traditional data centers, a rack might require a few dozen fiber connections but in AI clusters, they are pushing toward 10x more fiber per rack to support the massive speeds (800G to 1.6T) required for LLMs DC's can no longer expand capacity by just adding more racks. Instead, they are forced to pack more fiber into their existing footprint without causing a meltdown of tangled glass cables and trapped heat And the #1 thing DC's can't afford to have is MORE HEAT generated Here is why CLFD is uniquely positioned for this problem and why the interconnect layer where Clearfield specializes in will soon be a strategic priority for the entire AI industry For 15 years, Clearfield was a niche player protecting fiber in harsh outdoor environments Now, the AI Data Center is the harsh environment. The fiber density and heat levels inside a 2026 AI rack are so extreme that they require the rugged, modular protection Clearfield perfected for the outdoors Management realized this, which gave them an opportunity to enter the data center space In Jan 2026, they announced their new product “NOVA” which is their pivot to capture the market share of optical fiber density management The NOVA HD Panel is the only cassette based system that can fit 384 LC fiber ports into a 4U space (just 7 inches tall) While also guaranteeing a 0.2dB or less insertion loss (50% better than peers) Reducing insertion loss by 0.2dB can reduce a data center's overall power consumption by 10%, making Clearfield an infrastructure choice for an already power scarce environment For those who don't know, At 800G speeds, high insertion loss causes data errors, forcing AI systems to resend data, which wastes power and slows down training Here's the exciting part... Just a few days ago, the CEO spoke at Needham and she said that the data center community’s "biggest surprise" was recognizing Clearfield not as a newcomer, but as a superior expert in fiber termination and management that they now desperately need When NOVA was announced in Q1 26’ they didn’t expect any revenue at all until 2027 but in their recent Q2 26’ ER, orders already started flooding in Reporting a book to bill ratio of 1.3, backlog growing 39% for “adjacent markets” aka data centers at $31.6M On the call, management noted that data centers are skipping the typical multiyear testing phases because the current infrastructure is literally failing to support the new gen AI chips oh... and Clearfield is also a direct beneficiary of the move toward CPO A major challenge for CPO is that lasers are heat sensitive and fail often if they are buried inside a hot AI chip package The industry is moving toward ELS, placing the lasers at the front of the rack in a separate, hot swappable module Clearfield’s NOVA platform is designed specifically to act as the "standard socket" for these external lasers It provides the high density frame that allows technicians to pull out a failed laser cassette and snap in a new one in seconds without touching the delicate AI chip I am a buyer today because the signal from management is their confidence in their ability to pick up new customers given their expertise in the field + they had earnings already with a solid reaction which is a good signal for me I am anticipating their backlog to pickup at an even more rapid pace now fiber is being deployed at record pace This is an incredibly niche play, but I think their is plenty of upside of this company if they execute properly NFA. Please use for RESEARCHING PURPOSES ONLY.

$SHIM Q1’26 earnings are AH today. Looking forward to getting an update on the Company’s turnaround and developments on its Axia Electrical division.

@AtlasShrug1 I agree with you. SSTs will be big thing

$ENPH $SEDG Solid state transformers are going to be a massive theme heading into 2027 as I see it. We know traditional transformer demand for companies like $HPS is off the charts. The need for SSTs for data centers is critical to help alleviate a severe bottleneck. Utilities are always slow to act, but I believe they will ultimately become customers as well, and that market is friggin enormous. I highly recommend reading the White Paper Enphase recently posted to their website on their SST solution. Remember, you heard it here first! Happy trading!