So, Advanced Micro Devices Inc. (AMD) just cut a deal with Meta Platforms Inc. (META) to power its AI supercomputers. On paper, it's huge: up to 6 gigawatts of graphics processing unit (GPU) capacity, starting in the second half of 2026, with revenue potential that one analyst pegs at $90 to $100 billion over four years. That's the kind of number that makes you sit up straight. But Wall Street's reaction? Let's just say it's more of a thoughtful nod than a standing ovation. The analysts are intrigued, but they're also squinting at the fine print, especially the part where Meta could end up owning about 10% of AMD.

Here's the Wall Street scorecard up front. JPMorgan analyst Harlan Sur is keeping a Neutral rating on AMD. Goldman Sachs analyst James Schneider also reiterated a Neutral, though he did raise his price target to $240 from $210. On the more optimistic side, BofA Securities analyst Vivek Arya maintained his Buy rating with a $280 price target. So, we've got a classic case of "show me" versus "let's go."

Starting with the cautious crew, JPMorgan's Sur sees the big picture. The Meta deal, combined with AMD's earlier agreement with OpenAI, makes him more confident that AMD can land similar multi-gigawatt deals with other cloud giants and AI labs. He estimates the Meta deployment alone could generate $23 to $25 billion in annual revenue from 2027 to 2030. That's enough to blow past current 2027 expectations for AMD's data center GPU business, even before counting potential sales to other customers like Oracle Corp (ORCL).

But then there's the catch. The deal includes a warrant structure—up to 160 million AMD shares tied to deployment and performance milestones. Sur calculates this could represent $30 to $100 billion in equity value for Meta. More importantly for AMD investors, he thinks it could reduce gross margins on the deal's revenue by 200 to 400 basis points because of pricing adjustments. It's a classic trade-off: you get the massive revenue, but you have to share more of the pie. On the bright side, Sur highlights AMD's ability to customize its chiplet-based platforms as a key advantage and expects the company to also expand its server CPU business with Meta, which is a lead customer for its upcoming "Venice" processors. He believes AMD is well-positioned to keep taking server CPU market share from Intel Corp (INTC).

Over at Goldman Sachs, Schneider's view is similar. He says the five-year partnership with Meta strengthens AMD's position with a top-tier customer and improves visibility into future market share gains. He notes the financial structure mirrors the OpenAI deal, with those performance-based warrants translating to roughly 10% ownership. This announcement, he says, "modestly pressures" Nvidia Corp (NVDA) and Broadcom Inc (AVGO) as competition from big customers heats up. Yet, he's sticking with a Neutral rating. Why? He cites AMD's "significant exposure" to OpenAI and elevated operating expenses. He'd become more constructive if he had clearer visibility into exactly when Meta and OpenAI will deploy all this capacity through 2027. In other words, the promise is there, but the timeline is fuzzy.

Then we have the bull. BofA's Arya sees this deal as a positive step toward AMD's long-term goal of generating more than $20 in earnings per share by 2030. He acknowledges the equity component raises dilution questions, but argues that full dilution would require AMD's stock to approach $600—roughly triple its current price. At that point, he figures shareholders would still be looking at substantial accretion. His math is compelling: he estimates each gigawatt of deployment represents a $15 to $20 billion revenue opportunity and about $6 billion in net income potential. He also makes an important point about the broader AI landscape: this deal reinforces his view that central processing units (CPUs) are becoming increasingly crucial for AI inference workloads, not just the GPUs that get all the headlines.

Arya also provides some context for the competitive dynamics. He notes Meta could spend over $150 billion annually on capital expenditures. If AMD captures about $25 billion of that per year, there's still plenty of room for Nvidia and possibly Broadcom (through networking gear) to get their share. It's not necessarily a zero-sum game—yet.

As for the market's immediate reaction? AMD shares were down 1.39% during regular trading and down another 1.91% after hours on Wednesday, last trading around $206.84. Sometimes the market needs a minute to digest a deal this big and this complex.

So, what's the takeaway? AMD just landed one of the biggest customer endorsements in the AI chip race. The revenue potential is staggering, and it solidifies the company as a genuine alternative to Nvidia in the data center. But the path to cashing in on that potential is paved with warrants, margin questions, and execution risk. It's a high-stakes, high-reward move that has Wall Street analysts carefully weighing the billion-dollar upside against the 10% ownership catch. The AI chip war isn't just about who has the best technology anymore; it's also about who can structure the smartest deals.