Deep Dive: Understanding the AI Trade

The AI trade continues to ebb and flow, evolving with different winners and losers. The Daily Breakdown takes a deep dive into the AI trade. Before we dive in, let’s make sure you’re set to receive The Daily Breakdown each morning. To keep getting our daily insights, all you need to do is log in…

The post Deep Dive: Understanding the AI Trade appeared first on eToro.

The AI trade continues to ebb and flow, evolving with different winners and losers. The Daily Breakdown takes a deep dive into the AI trade.

Before we dive in, let’s make sure you’re set to receive The Daily Breakdown each morning. To keep getting our daily insights, all you need to do is log in to your eToro account.

Deep Dive

ChatGPT launched in November 2022 and hit 100 million users within months, kicking off a whirlwind AI trade — from Nvidia’s massive rally to more recent turbulence in software stocks

Despite the flurry of headlines this week, one item stood out: Nvidia investing $2 billion in Coherent and $2 billion in Lumentum. That’s a reminder that the AI buildout isn’t just chips — it’s also “optics and photonics,” where bandwidth demands are exploding inside data centers. COHR and LITE are up 37.5% and 77% year-to-date, respectively.

Understanding the AI Trade 

Chips were the first, most obvious winners — Nvidia, Advanced Micro Devices, Broadcom, Taiwan Semi, and the toolmakers like ASML, Lam Research, Applied Materials, and KLA Corp. But the opportunity set is broader.

The AI infrastructure trade is really a “picks-and-shovels” buildout: compute, networking, and the physical systems required to power and cool dense AI racks. Beyond silicon, that includes networking (Broadcom, Marvell Technology, and Arista Networks), optics (Applied Optoelectronics, Coherent, Lumentum), and increasingly power and thermal enablers like Eaton and Vertiv as electricity and cooling become gating factors. Datacenters are the “housing” layer for all of that gear — the physical campuses and leased capacity that determine how quickly new AI compute can actually be deployed. That’s where operators like Equinix and Digital Realty fit in. 

Finally, there’s the “software infrastructure” layer: Firms like Datadog and Snowflake help enterprises monitor and manage data and AI workloads in production, while companies like Palo Alto Networks and CrowdStrike operate on the cybersecurity side of the expansion as AI increases complexity, automation, and the attack surface.

Want to receive these insights straight to your inbox?

Sign up here

Diving Deeper — Sentiment & Risk

We’ve seen how sentiment can ebb and flow, particularly within this group. Take Nvidia for instance. Despite beating on earnings and revenue estimates and providing strong guidance, shares have moved lower since earnings. Its valuation now hovers near multi-year lows. In other words, sentiment has dampened despite a strong fundamental backdrop.  

A key dynamic in the AI infrastructure trade is bottleneck rotation. It began with scarce accelerators and advanced-node capacity, then shifted to networking and optics as moving data between chips became the limiter. Now power delivery and cooling are increasingly becoming key factors, with grid constraints and long-lead electrical and thermal gear dictating how fast new capacity can come online.

In other words, these dynamics can shift quickly, with sentiment flipping from hot to cold — and back again.

Risks 

Sentiment swings are one thing to be cognizant of, but there are other risks too. Some include: timing risk on “next leg” themes, power/permitting delays that push revenue recognition out even when demand is intact, and concentration and pricing risk in the enablers, where a small set of customers can re-source, insource, or renegotiate as volumes scale. That doesn’t even take into account bigger-picture macro risks.  

The Bottom Line

AI is still a multi-year buildout with real, investable opportunities across the stack as spending expands from compute into networking, optics, and power infrastructure. The trade won’t be linear — bottlenecks and sentiment will rotate — but the direction of travel is clear, and the winners tend to emerge where constraints are tightest and demand is most durable. Still, investors should do their own research to determine which risk-reward setup best fits their style.

Disclaimer:

Please note that due to market volatility, some of the prices may have already been reached and scenarios played out.

The post Deep Dive: Understanding the AI Trade appeared first on eToro.