AMD Is Up 158% While Nvidia Slips. The Case Runs Deeper Than Rival Trouble

AMD's stock is up 158.17% year to date, and over the same stretch Nvidia is down 6.63%. That divergence frames the argument a 24/7 Wall St. columnist laid out on November 12, 2026: the AI buildout needs a second chip engine, and AMD has become that engine. His case, notably, does not rest on Nvidia's problems alone. It rests on AMD's own numbers, four structural advantages, and a forward schedule of products already in customer hands.

The backdrop matters for anyone tracking the AI hardware market. Hyperscalers and neoclouds are spending hundreds of billions on accelerated computing, and they want more than one supplier for the silicon underneath it. That preference is now showing up in AMD's financials, its product roadmap, and its customer list.

What are the headline numbers?

The financial evidence comes from AMD's latest earnings report, and it is concentrated in one segment. Data Center revenue reached $8.546 billion, up 97% year over year. Segment operating margin climbed to 34%, from 25% a year earlier. Data center now accounts for 64% of AMD's total revenue, up from 47% a year ago.

Profits are growing faster than sales. Non-GAAP EPS reached $2.09, up 74%, while revenue grew 45%. The columnist notes that this marked AMD's fourth consecutive EPS beat. The CFO expects 2027 operating expenses to grow at less than half the top-line revenue growth rate — a signal that operating leverage should keep compounding if demand holds.

The guidance extends the trend. Fourth-quarter revenue is guided to roughly $17 billion, about 66% growth at the midpoint. AMD expects 2027 revenue to grow more than 80%. Early demand for Venice, the newest EPYC server chip generation, is, in the company's words, "the strongest we have seen for any new EPYC generation."

Why do hyperscalers keep paying AMD?

The columnist's core argument rests on four structural advantages, each aimed at a cost or design problem that large AI buyers actually face.

• Cost. CEO Lisa Su frames Instinct accelerator sales around "tokens-per-dollar" advantages — a pitch that speaks directly to buyers who measure AI spend by output, not by chip count.
• Memory. AMD says its Helios rack-scale system offers "an even larger lead in memory bandwidth and higher scale-up bandwidth than the competition." Memory bandwidth is a binding constraint on large AI models, which makes the claim commercially pointed.
• Openness. AMD contributed its Helios rack design to the Open Compute Project. Buyers can adopt the architecture without locking themselves into a single supplier.
• Breadth. AMD sells both the accelerator and the CPU, and the rise of AI agents, in the columnist's telling, increases demand for both.

The customer evidence is concrete. Five lead customers, including Anthropic and OpenAI, already run workloads on Helios racks, according to the column. That detail matters because it converts the structural argument into deployments: the second-engine thesis is not hypothetical.

How much trouble is Nvidia actually in?

Nvidia remains, by the columnist's own description, an elite business, with a 75.0% gross margin. The concerns are about its balance sheet and its momentum. Supply obligations rose to $279.00 billion. Partner guarantees run up to $108.5 billion. Customers now take 60 days to pay, compared with 45 days before — a receivables shift that suggests financing strain inside the AI supply chain.

The market has noticed. Nvidia shares trade at about 35 times earnings and are down 9.84% over the past year. None of this, the columnist concedes, breaks the business. It does, however, sharpen the question investors are already asking: whether Nvidia's turmoil produces lasting wins for AMD. His answer is yes, and for reasons that would hold even if Nvidia's troubles eased.

Why does the argument skip Intel?

Intel is the other logical second supplier, and the column addresses it directly. Intel grew its data center unit 78% — a genuinely strong number. But its foundry business still lost $1.85 billion last quarter. Intel guides to a 38.0% non-GAAP gross margin, while AMD guides to about 55%. And Intel carries more debt: borrowings stand at 0.41 times equity, compared with AMD's 0.071. On profitability and leverage, the gap is wide enough that the columnist treats AMD, not Intel, as the credible alternative to Nvidia.

What could hurt the AMD position?

The column is explicit about the risks, and they're worth listing in full.

• No dividend. Income-focused investors get nothing while they wait.
• Rich valuation. The stock trades at 135 times free cash flow.
• Shrinking cash generation. Free cash flow fell 70.86% to $554 million as inventory rose to $11.463 billion.
• A warrant charge. A $1,097 million customer warrant charge reduced GAAP EPS to $1.46.
• Customer concentration. Sales depend heavily on a few large buyers.
• Export controls. The columnist flags them as a live threat, one that policy shifts could tighten at any time.

Any of these could cause real damage if demand cooled, he writes. So far, it hasn't. Su says "demand is higher than supply," and AMD built that $11.463 billion inventory deliberately — to support fourth-quarter data center growth of about 50% sequentially. The risk picture, in other words, depends heavily on whether that demand assessment holds.

What comes next on the roadmap?

The forward case has three moving parts. Helios production is increasing. Venice is shipping. And AMD is already working with customers on MI500 and MI600 — future Instinct generations that extend the accelerator line beyond what's deployed today. The combination of rising rack production, a new server CPU in volume, and a next-generation accelerator pipeline is what the columnist points to when explaining why he keeps adding to the position from here.

The structural driver behind all of it is simple: hyperscalers want a second supplier for AI chips, and AMD is the one with the product breadth, the open rack design, and the customer deployments to fill that role.

The columnist also widens the lens beyond chipmakers. The whole AI expansion needs power, cooling, and networking behind it, which is why his outlet assembled a separate report on seven non-chipmaker suppliers in the AI infrastructure trade. For readers, the so-what is straightforward: AMD's earnings give the clearest quarterly scoreboard for the second-supplier thesis, and the fourth-quarter print — with roughly $17 billion in guided revenue and about 50% sequential data center growth — is the next test of whether the demand that Su describes stays ahead of supply.