AI Stocks: Wall Street’s Price Targets Reveal Where the Biggest Upside May Be

Wall Street remains broadly bullish on the AI trade, but the latest price targets show that analysts are not treating every AI stock equally.

Across a group of major AI beneficiaries — from memory and semiconductors to AI infrastructure and space-based computing — the average analyst targets imply meaningful upside, with several names carrying expectations of 70% or more.

The dispersion between low, average and high targets is equally important. It shows not only where Wall Street sees potential upside, but also where expectations — and therefore risk — are especially elevated.

🧠 Memory: Micron and SK Hynix Lead the Bull Case

$Micron Technology(MU)$ stands out with one of the most aggressive analyst outlooks.

• Low target: $1,100

• Average target: $1,600

• High target: $2,200

• Average upside: 78%

The bullish case is closely tied to the AI-driven memory cycle. Demand for DRAM and high-bandwidth memory remains strong, while supply constraints are supporting pricing power. Recent industry commentary continues to point to tight HBM and NAND conditions, reinforcing the earnings-growth argument for memory producers.

But the expectations are already extremely high. With Micron having gained roughly 200% year to date, the key question is no longer whether AI is driving memory demand — it is whether earnings can continue accelerating fast enough to justify the valuation and expectations.

$SK hynix(SKHY)$ offers a similar AI-memory thesis:

• Low target: $200

• Average target: $245

• High target: $320

• Average upside: 78%

Recent analyst initiations have emphasized SK Hynix's leadership in the AI memory market, with several firms assigning targets around $240–$260. Its valuation is also viewed as relatively attractive compared with Micron, adding another layer to the bullish argument.

The takeaway is clear: Wall Street sees AI memory as one of the strongest structural beneficiaries of the next phase of AI infrastructure spending.

☁️ AI Infrastructure: Nebius and IREN Offer Higher-Risk, Higher-Reward Exposure

$NEBIUS(NBIS)$ sits in a different category.

• Low target: $144

• Average target: $241

• High target: $286

• Average upside: 28%

The average target is less aggressive than those of the memory names, but the range remains extremely wide.

That reflects the market's uncertainty around the speed at which AI cloud infrastructure companies can convert massive demand into revenue, capacity and ultimately free cash flow.

For NBIS, execution is everything. Investors will be watching contracted capacity, GPU deployment, customer commitments and the pace of revenue growth. The bull case is powerful, but the market will demand evidence that infrastructure expansion is translating into sustainable economics.

$IREN Ltd(IREN)$ represents an even more aggressive version of the AI infrastructure trade:

• Low target: $46

• Average target: $75

• High target: $100

• Average upside: 82%

The 82% average upside places IREN among the most optimistic setups in this group.

The opportunity comes from the transformation of power and data-center infrastructure into AI compute capacity. But that opportunity also creates significant execution risk. Capital requirements, financing, power availability, GPU deployment and customer concentration can all have an outsized impact on the investment case.

In other words, the upside is substantial, but so is the sensitivity to execution.

🏆 Nvidia and Broadcom: The AI Leaders Still Have Room to Run

Despite their enormous size, Wall Street continues to see upside in the industry's two most important semiconductor infrastructure names.

$NVIDIA(NVDA)$ :

• Low target: $250

• Average target: $309

• High target: $500

• Average upside: 38%

The $500 high target highlights how aggressively some analysts continue to view Nvidia's long-term AI opportunity.

Nvidia's advantage is that it sits at the center of the AI compute stack. The debate is increasingly shifting away from whether demand exists and toward how long hyperscaler and enterprise AI spending can remain at elevated levels.

That makes forward guidance, networking demand, next-generation architectures and hyperscaler CapEx increasingly important variables.

$Broadcom(AVGO)$ :

• Low target: $390

• Average target: $516

• High target: $630

• Average upside: 21%

Broadcom has a lower implied upside than Nvidia, but that does not necessarily mean the opportunity is smaller.

Its AI exposure extends beyond accelerators into custom AI chips and networking infrastructure, giving investors another way to participate in the expansion of AI data centers.

The relatively tighter range between the low and high targets also suggests that Wall Street expectations are somewhat less extreme than they are for some of the more speculative AI infrastructure names.

🚀 SpaceX: The Widest Range of Them All

Perhaps the most striking numbers belong to $SpaceX(SPCX)$ :

• Low target: $75

• Average target: $230

• High target: $800

• Average upside: 72%

The enormous spread between $75 and $800 tells investors something important: SpaceX is exceptionally difficult to value using traditional frameworks.

The investment thesis extends beyond today's launch business into Starlink, satellite connectivity, reusable launch technology and potentially massive future AI infrastructure and compute applications.

That creates an enormous long-term opportunity — but also extraordinary uncertainty.

A $230 average target implies substantial upside, while the $800 bull-case target represents a completely different valuation framework based on much more aggressive assumptions about SpaceX's future growth and monetization.

📊 What the Targets Tell Us

Looking across the group, the most interesting observation is that Wall Street's biggest upside expectations are not necessarily concentrated in the largest AI companies.

The pattern is revealing.

Memory and AI infrastructure currently carry some of the largest implied upside, while established AI leaders such as Nvidia and Broadcom have more moderate — but potentially more defensible — expectations.

That creates two very different investment approaches.

One is to own the companies already generating enormous AI revenue and profits.

The other is to target the next layer of the AI infrastructure buildout, where earnings are smaller today but the potential growth rates are much higher.

🎯 The Bigger AI Trade

The common thread across all seven names is that AI spending is expanding beyond GPUs.

The next phase of the cycle increasingly involves:

Memory → Networking → Power → Data Centers → AI Cloud → Compute → Custom Silicon

That is why companies such as Micron, SK Hynix, Broadcom, Nebius and IREN are increasingly appearing alongside Nvidia in AI investment discussions.

But there is an important distinction between Wall Street's target price and an investment thesis.

A high price target does not guarantee high returns.

In fact, the widest target ranges often belong to the companies with the greatest uncertainty.

The numbers therefore should be read as a map of expectations, not a promise of future performance.

For investors, the more important question is whether the companies can continue to beat the assumptions already embedded in these targets.

Because in the AI trade, the biggest risk may no longer be missing the opportunity.

It may be paying for too much of the future before that future arrives.