JP Insights

Currently building a position in cash. Not sure what % I will be targeting yet.

Free posts can be just as good as paid ones if you find accounts that help you cut through the noise. I added to my $CRDO position at around $163 late last week. It’s now at $175, but still nowhere near a price where I’d even consider trimming. https://t.co/pMrAlFNHzj

Ciena’s new targets caught my attention, and I think anyone following $NOK should spend some time on them too. $CIEN is targeting roughly 30% annual revenue growth from FY2026 through FY2029, with adjusted gross margin around 50% and operating margin of 32–35% in FY2029. Management also described continuing supply constraints. Those are ambitious targets, but the combination of growth and profitability is very interesting. Nokia’s optical business, including Infinera, puts it in the same market for spending on optical networks. Ciena’s outlook makes me more interested in the opportunity, although the two companies compete for customers. There was a concrete Nokia development this week as well. On September 15, Nokia announced that Telxius is deploying its 800G coherent pluggable optics across terrestrial networks in Europe, the US and Latin America. The stated purpose includes supporting growing AI and datacenter connectivity demand while reducing power consumption and expansion costs. This brings me back to the question I raised in my Nokia article: volume versus pricing power. I like seeing customer deployments. But I want to see how much profit Nokia retains as those deployments grow. Ciena’s margin ambitions give us a useful benchmark for that discussion, although Nokia’s broader business mix means the group margins won’t be directly comparable. For me, Ciena strengthens the demand case for Nokia’s optical business. Nokia’s own margins and cash generation will determine how much that demand is worth to shareholders.

$ETN gave some pretty wild numbers yesterday. Eaton acquired Fibrebond in 2025 for $1.43B. At the time, the business was doing around $375M in annual revenue. Just over a year later, Fibrebond has generated more than $600M over the last twelve months. And Eaton is preparing for much more. They are expanding the existing facility and building a new factory in Arkansas, with each project expected to double capacity. The broader data center business isn't slowing either. Revenue is running around +65% and orders +85%. Eaton started 2026 expecting 8% organic growth. They raised that to 10%, and CEO Paulo Ruiz now says they could reach 12%. Hard to look at those numbers and argue that the physical AI infrastructure buildout is losing momentum.

$VRT CEO Giordano Albertazzi pushed back on the idea that slowing frontier AI development would slow the infrastructure buildout. His point is simple. There is still a huge amount of adoption left for what already exists today, while compute demand continues to run ahead of available supply. Like I stated earlier, I spend time listening to the people actually running these businesses to gather information. $VRT remains one of my highest conviction holdings.

Conviction as strong as ever in $CRDO. The lower it goes the more I want to add to it. I don't mean that buying the dip in every company is a good way to make money. It is not. But Credo is quality and will outperform over time. Timeframe 2-5 years. https://t.co/OrTbLDXrfD

$ASML says Samsung and $SKHY plan to use High-NA EUV in production from 2028. $MU has already ordered the tools as well. $TSM plans to use High-NA in high-volume production from 2030. What I like here is the timeframe. Memory manufacturers are spending today on equipment meant for production several generations from now. Hard to square that with the idea that the HBM buildout is some short cycle that peaks next year.

My Best Ideas 2026 basket is up 33.72% so far this year. I shared it for free at the start of January so anyone could follow along. The aim was to find quality companies in the AI buildout that I could hold through the year with very few changes. I'm in my early 40s now, and I honestly think I'd have a lot less money today if I'd spent my investing years constantly trading in and out of names. Please read the disclaimer in the picture. This basket is not a copy of my family's portfolio. So let's go through the names. $ALAB, $ANET, $VRT and $POWL have been the strongest performers so far. $AVGO has been more frustrating, almost flat for the year. My conviction remains very high and it's staying in the basket. I still expect it to be one of my best performers from here. $CDNS has also disappointed, though looking back, the valuation left very little room for disappointment. I was too willing to accept that starting valuation. $CDNS has also disappointed. Looking back, I was too comfortable with the starting valuation and how much growth it already priced in. It's the name I'm most likely to replace, with $LITE, $NBIS and $IREN among the contenders. I've made just two changes: replacing $GOOGL with $ETN, then swapping $HUBB for $MU in June. I enjoy the weekly portfolio updates from @pequityresearch and @TacticzH, among others. Think I'll start doing something similar here, including what went wrong and why I make changes. The basket and my updates will be free to follow.

The calls to slow AI development have me thinking about $CRDO’s H2 ramp. I’ve put considerable weight on Credo preparing inventory and capacity ahead of it. That shows me management believes the customer forecasts. It doesn’t guarantee customers take delivery on schedule. Credo generally lacks long-term minimum purchase commitments. Some orders can be delayed or cancelled with little notice and no penalty, while Credo has commitments of its own to suppliers. That risk was already there. A coordinated slowdown could make it more relevant. Slower model development doesn’t automatically mean customers stop building clusters. I’m still bullish, and I haven’t seen evidence that Credo’s H2 projects have slipped. But I can’t use the inventory build as proof that the revenue will arrive when expected. The question I’d like management to answer: how much of H2 can customers push back without compensating Credo?

I’ve added to my $CRDO because I think the next two quarters will make the H2 opportunity much harder to ignore. Why? Credo’s outlook for more than 85% revenue growth this fiscal year implies revenue above roughly $2.47B. Q1 delivered $479M, and Q2 guidance is $525–535M. Using the Q2 midpoint, that gives us about $1.01B in the first half. To reach 85% growth, Credo then needs roughly $1.46B in H2. About 45% more revenue than H1. That acceleration is already embedded in management’s outlook. For a sense of the quarterly progression, approximately $655M in Q3 and $806M in Q4 would get them there. These are my illustrative numbers, not quarterly guidance from Credo. And if we hang on to the Q4 figure. We’re looking at a business that just reported $479M in quarterly revenue, with an annual outlook that could support an exit quarter above $800M. That’s why I struggle with putting so much weight on whether Q1 beat expectations by enough. My investment case has been built around the larger revenue ramp still ahead, and management raised its full-year growth outlook after reporting Q1. The profitability is another reason I’m bullish. Q1 non-GAAP gross margin was 68%, with Q2 guided to 67–69%. At that level, an additional $100M in quarterly revenue produces roughly $68M in gross profit before operating expenses, assuming the margin holds. So I’m thinking about what earnings could look like as those larger revenue quarters arrive. There is considerably more earning power in an $800M quarter than a $479M quarter, even without assuming any gross margin expansion. Customer timing is the main execution risk I’m watching given how much revenue sits in H2. But my expectation is that the ramp happens, and I see room for Credo to exceed the full-year outlook. I’ve added because I want to own more of the business before those quarters arrive. By then, we’ll be debating what comes after an entirely different revenue base.