Latest Stock Buy or Sell? Make More Informed Decisions!

We've gone through one cycle of investments, which were largely in large language models and data centres. Now we're entering another investment phase, which revolves around agentic AI. The types of hardware needed for model training are very different from the types needed for agentic AI. That means change, and that's where we're seeing a lot of opportunities.

For the model training era, the main focus was on GPUs. You were able to use one CPU to connect 16 GPUs together. Now, in the agentic AI era, we're using one GPU per CPU -- creating massive demand for CPUs. That's why stocks like AMD, INTC, and ARM perform really well. We're at an inflection point, and we're in year 1 of that cycle.

Another area she likes is networking. The number of people on earth is finite, but the number of agents is infinite. Connecting these agents leads to tremendous opportunity -- copper, connecting components, and optical components.

A third area is memory, but this one is trickier because that sector is a lot more cyclical than the other two.

Yes, she'd absolutely invest in traditional players. Concerns around software were valid, but the concerns around cybersecurity were mispriced. Cybersecurity stocks sold off together with software stocks, but it wasn't justified.

Seeing lots of opportunities. First time in history where robot traffic on the web exceeds human traffic. Interfaces will have to change to accommodate this new reality.

Very attractive here, if you look 3-5 years out. You need to have a very long-term horizon. She's looking for the point when launch costs drop from $3k per kg to $300, it'll open a massive opportunity for new industries. Not super-cheap at this price; ideally, you want it when market cap is below $2T.

Unique, as the only one right now that can use reusable rockets to achieve cost reduction.

META selling excess capacity is specific to META, not necessarily a sign that hardware is slowing. Still likes it here.

Her team is very excited about this. Use cases haven't scaled or been proven over the last 2-3 years. They've been looking at commercial applications, which are in a very experimental phase. Makes it hard to justify on risk/reward right now. Time to invest is once we see commercial acceleration.

Big players are IONQ and RGTI. Unclear which technology will be the winner.

Her team is very excited about this. Use cases haven't scaled or been proven over the last 2-3 years. They've been looking at commercial applications, which are in a very experimental phase. Makes it hard to justify on risk/reward right now. Time to invest is once we see commercial acceleration.

Big players are IONQ and RGTI. Unclear which technology will be the winner.

Her team is very excited about this. Use cases haven't scaled or been proven over the last 2-3 years. They've been looking at commercial applications, which are in a very experimental phase. Makes it hard to justify on risk/reward right now. Time to invest is once we see commercial acceleration.

Big players are IONQ and RGTI. Unclear which technology will be the winner.

Not invested right now, as it's very hard to justify upside to earnings from analysts' estimates here. Stock's cheap, so not bad from risk/reward. On the flipside, when the market sells off it has very good valuation support because the multiple's so low. Some upside, with pretty decent downside protection. 

But other stocks have much more significant earnings upside. Not the pick for outsized performance.

Good opportunity if you're looking for broad exposure to semis with both GPUs and CPUs. Whichever cycle comes out the winner, this name will perform well from risk/reward. Pullback today due to overreaction on META news, not a bad entry.

Her team prefers to play one or the other, and right now is really excited about CPUs. 

A bit late in the cycle for here to put new $$ in, though there's about 2-3 more years in the memory cycle. Prices have really spiked (up 7x). If prices stall, and even if volume continues to ramp up, people will stop being so excited. Still upside, but risky at this point.

Memory is a big area for the agentic AI wave, but there are better ways to participate.

We'll see power and energy constraints develop over the next 5-10 years. Many of the power producers have capacity sold out to 2030. To get upside, they need to either increase capacity or reprice the backlog. Decent opportunity on margins, which are still below where they could be.

Likes it here. She could be interested on a pullback.

Her firm looks at defense as part of space/defense exposure. A lot of stocks have pulled back due to Iran conflict being resolved. Defense budgets are going up significantly. She prefers other names.

Main competitor is CRWV. But META news yesterday also makes it a new competitor. The cloud-as-a-service space is getting much more crowded and competitive. Economics won't be as strong as in the prior cycle. Not necessarily the end of the trade, but good time to take some profits.

(Note the short timeframe.)  Still one of her highest-conviction ideas. Real upside in optics will become very apparent by 2027-28. Pullback is a good entry point.