AMD’s $1T marks AI chip rivalry thawing

The Trillion‑Dollar Threshold

AMD’s stock touched $613.92 intraday on Monday, lifting its market value just above $1 trillion for the first time. The milestone arrived after a five‑day winning streak that added roughly 24% to the share price, erasing much of the June dip triggered by a second‑quarter earnings beat that nonetheless rattled investors. Year‑to‑date, AMD has gained more than 180%—a return that would have been unthinkable a few years ago.

The numbers driving the rally are stark. Data‑center revenue hit $6.7 billion in Q2, up 107% from the prior year. Total revenue came in at $11.54 billion, a 50% increase year‑over‑year. Lisa Su, AMD’s chairwoman and CEO, said during the earnings call that the company expects data‑center sales to double by 2027. That guidance arrived just days after NVIDIA CEO Jensen Huang projected his own shipments would double next year. Two CEOs, one shared conviction: the appetite for AI silicon shows no sign of saturating.

Yet the trillion‑dollar mark is less a triumph than a turning point. AMD’s valuation surge reflects a structural shift in how the market prices second‑source risk. For years, NVIDIA’s dominance in AI accelerators was treated as unchallengeable. Today, hyperscalers and cloud providers are actively diversifying their supply bases, and AMD’s momentum is the clearest evidence that the oligopoly is fracturing.

The NVIDIA–AMD Gap Narrows

NVIDIA’s market capitalisation sits at roughly $5.4 trillion. AMD, at $1 trillion, still trails by a factor of five. But the gap is closing faster than raw multiples suggest. AMD’s data‑center revenue grew more than twice as fast as NVIDIA’s in Q2, and the company’s MI300 series has won design wins with every major cloud provider and several large language model developers. Customers now have a credible alternative for training and inference workloads that were previously NVIDIA‑only.

This diversification is not abstract. OpenAI, Microsoft, Google, Meta, and Amazon all run mixed fleets of NVIDIA and AMD chips. The economic logic is simple: second sourcing reduces vendor lock‑in, softens pricing pressure, and spreads performance risk. As AMD’s share of the AI accelerator market climbs, NVIDIA’s pricing power will face sustained downward pressure. That does not mean NVIDIA will lose ground—it still holds a commanding lead in software ecosystem depth and performance per watt—but it does mean the profit premium that has inflated the company’s valuation must now be contested.

The narrowing gap also changes how investors evaluate both firms. AMD’s current multiple is lower than NVIDIA’s, yet its growth rate is higher. Over time, market caps converge when growth trajectories diverge. If AMD delivers on its 2027 data‑center double‑digit target, its valuation could re‑rate sharply. If NVIDIA’s growth slows even modestly, the relative gap will shrink further.

European and Asian Suppliers Reckon with Reality

For chipmakers outside the US, AMD’s surge is a double‑edged signal. On one hand, rising AI demand expands the total addressable market for foundries, packaging services, and memory producers. On the other, the shift toward second sourcing accelerates capacity investment by multiple buyers, which can compress margins in mature segments.

Taiwan’s TSMC remains the dominant foundry for advanced AI chips, but Samsung Foundry and Intel Foundry Services are courting AMD and other designers eager to diversify beyond a single supplier. European firms like ASML and Infineon benefit indirectly from higher capital expenditure across the industry, but they also face longer order books that can delay delivery and inflate costs for smaller customers. South Korean memory giants SK Hynix and Samsung Electronics, which supply high‑bandwidth memory critical to AI accelerators, will see demand grow—but so will negotiation leverage for buyers who can switch between vendors.

The geopolitical layer complicates the picture. US export controls have already constrained China’s access to leading‑edge chips, and both AMD and NVIDIA have adjusted their product portfolios accordingly. China’s push for self‑sufficiency, through firms like Huawei’s Ascend, creates a parallel market that competes for Asian foundry capacity. Meanwhile, the EU’s Chips Act and the US CHIPS Act are funneling subsidies into domestic production, which may slow the growth of overseas supply chains that European and Asian suppliers have historically served.

What Happens Next

The most immediate consequence of AMD’s trillion‑dollar milestone is a rebalancing of buyer power. Hyperscalers that have historically accepted NVIDIA’s pricing terms will now demand competitive bids. That pressures NVIDIA to invest more aggressively in R&D and cost reduction, while rewarding AMD for execution. The result should be faster iteration cycles and more features per dollar—a healthy outcome for the industry, but a test for both companies’ engineering discipline.

Valuation convergence will continue as long as AMD maintains its growth trajectory. If data‑center sales double by 2027 as guided, AMD’s revenue could approach $15‑18 billion quarterly, up from $6.7 billion. That level of scale, combined with improved margins from higher mix of AI chips, would justify a meaningfully higher market capitalisation. NVIDIA’s own doubling of shipment volume next year, as Huang indicated, suggests the total pie is expanding fast enough to accommodate both players without immediate margin erosion.

For European and Asian suppliers, the lesson is one of adaptation rather than alarm. Demand for AI‑grade silicon, memory, and packaging will remain robust, but the competitive landscape is shifting toward a multi‑sourcing model. Firms that position themselves as flexible partners across multiple design wins will capture more value than those locked into a single customer relationship. Foundries that secure capacity commitments from both AMD and NVIDIA, while keeping options open for emerging Chinese designers, will be best placed to weather policy swings.

The broader implication is that the AI chip boom is maturing. The era of a single dominant supplier is giving way to an oligopoly with two clear leaders and a handful of challengers. AMD’s trillion‑dollar entry marks the moment the second player crossed the threshold that once seemed reserved for NVIDIA alone. In markets where dominance is measured in trillions, that crossing is the new normal.

Investors and competitors alike will now watch AMD’s ability to sustain its momentum. If the company can translate its data‑center growth into consistent margin expansion and continue winning design cycles, the valuation gap with NVIDIA will keep shrinking. If execution stumbles, the rally may prove cyclical rather than structural. Either way, the semiconductor landscape has changed—and the trillion‑dollar club just got a new member.