The Legitimate Challenger: Why AMD’s 6-Gigawatt Meta Deal Changes Everything

Written by Romeo Kuok

Citi’s double upgrade to Buy with a $575 price target validates AMD as a credible AI GPU challenger. The 6-gigawatt Meta partnership and custom MI450 architecture mark the first time a hyperscaler has bet its core AI infrastructure on AMD over Nvidia.

Advanced Micro Devices (AMD) surged 4.6% on Friday, June 12, to close at $515.92—pushing its year-to-date gains to an astonishing 131%. The catalyst was a massive double upgrade from Citi analyst Atif Malik, who elevated the stock from Neutral to Buy and aggressively hiked his price target from $460 to $575. Malik’s core thesis is simple but profound: “The market has yet to fully recognize AMD as a legit second source in the GPU market.”

For the past three years, the artificial intelligence hardware narrative has been entirely dominated by Nvidia. But beneath the surface, a structural shift is occurring. AMD’s custom MI450 chips are not just catching up—they are winning massive, foundational deployments at the expense of the incumbent.

The turning point is AMD’s unprecedented 6-gigawatt AI infrastructure partnership with Meta Platforms. This four-year supply agreement, which includes a 160 million-share warrant from AMD to Meta, represents the first time a major hyperscaler has bet its core AI infrastructure on AMD over Nvidia. The deployment begins ramping with an initial 1-gigawatt tranche in the second half of 2026. Citi’s analysis indicates that the custom MI450 architecture provides Meta with a substantially lower total cost of ownership compared to Nvidia alternatives.

This is a watershed moment for AMD’s financial trajectory. Citi projects that AMD’s AI GPU revenue will reach $33 billion in the near term, expanding to $50.8 billion as the Meta deployment scales. In its most recent quarter, AMD reported $10.25 billion in total revenue, with the Data Center segment contributing $5.78 billion—a 57% year-over-year increase. CEO Lisa Su confirmed that customer engagement around the MI450 Series and Helios architecture is strengthening, with leading customer forecasts exceeding initial expectations.

However, the valuation debate between AMD and Nvidia remains complex. Despite AMD’s momentum, it trades at a staggering 159x trailing earnings and 68x forward earnings. In stark contrast, Nvidia—which just posted an $81.6 billion quarter with $75.25 billion in Data Center revenue—trades at a forward multiple of just 24x. Nvidia is growing at 85% with 75% gross margins, recently authorizing an $80 billion buyback and a 25x dividend hike. While AMD is the momentum challenger, Nvidia remains the cheaper stock on a forward basis.

The broader market context also provided a massive tailwind on Friday. The S&P 500 added 0.5% and the Dow surged 353 points as President Trump called off planned strikes against Iran, sparking hopes for an imminent peace deal to be signed in Switzerland as soon as Sunday. Oil plunged below $85 a barrel, and the 10-year Treasury yield dropped to 4.47%, releasing the interest rate pressure that had hammered semiconductor stocks earlier in the week. Meanwhile, SpaceX executed the largest IPO in history, closing up 19.2% at $160.95 and pushing its valuation to $2.2 trillion.

Amid this macro euphoria, the AI infrastructure trade is broadening. The market is finally pricing in a duopoly rather than a monopoly. AMD has proven it can secure the hyperscaler commitments required to sustain massive revenue growth. The risk, however, is that AMD’s premium valuation leaves little room for execution missteps. If the MI450 ramp disappoints or Meta’s deployment timeline slips, the 68x forward multiple could compress violently.

The bull case is straightforward: AMD is building a second source moat in AI GPUs that the market has not fully priced. The 6-GW Meta deal is not a one-off—it is a template for future hyperscaler engagements. As AMD’s ROCm software ecosystem matures and HBM4 collaboration with Samsung for the MI455X advances, the company’s competitive position will only strengthen. Citi’s $50.8 billion AI GPU revenue projection implies AMD could eventually capture 15-20% of the total AI accelerator market.

The bear case centers on execution risk and valuation. At 68x forward earnings, AMD is priced for perfection. Nvidia’s networking revenue alone ($14.8 billion in Q1) exceeds AMD’s entire Data Center business. The CUDA ecosystem remains the industry standard, and Nvidia’s platform sweep—from Vera Rubin CPUs to Dynamo 1.0 inference optimization—continues to widen its moat. AMD must prove that cost efficiency can overcome ecosystem lock-in.

For investors seeking AI exposure with a favorable risk-reward profile, AMD represents the highest-beta play on the AI GPU duopoly thesis. The Citi upgrade is not just a price target revision—it is a structural re-rating of AMD’s competitive position. The 6-gigawatt Meta deal provides the revenue visibility that previously only Nvidia could claim. Buy on pullbacks below $500 for a 12-month target of $575.

Analyst Verdicts and Price Targets