AMD's Data-Center Pivot Is Winning the Argument, Not Yet the Valuation

Lisa Su used a Thursday interview with Yahoo Finance to leave no doubt about where AMD is headed: data centers come first. The strategic logic is straightforward enough. What investors now have to weigh is whether a business tilting ever more heavily toward a single customer base can convert surging demand into shippable product — and whether the price of the stock already assumes it will.

The stakes are visible in the numbers AMD has already put on the table. Data-center revenue hit $8.5 billion in the third quarter of 2026, up 97% year over year. By 2027, management expects data centers to account for more than 80% of total revenue, alongside a commitment to more than double the size of its AI data-center business. That is not a marginal shift in mix; it is a redrawing of the company's center of gravity.

A growth engine with a narrowing base

Concentration cuts both ways. On one hand, AMD is steering investment toward the part of the market where demand is strongest — Su said plainly that demand is outpacing supply, which argues for further expansion provided the capacity build-out with supply-chain partners can actually absorb additional orders. On the other, the more revenue depends on data-center projects moving forward, the more the investment case rests on customers following through on their capital-spending plans.

The offsetting side of the ledger is getting heavier. Gaming revenue fell 40% to $773 million in the third quarter, dragged down by lower semi-custom chip sales and softer graphics demand. For the closing quarter, AMD expects the combined client and gaming segment to decline by more than 20% sequentially. Management is also planning for a percentage decline in the double digits in PC shipments for 2027 — a tougher PC market is baked into the guidance, not an afterthought.

That makes the data-center push do double duty: it has to generate growth and simultaneously mask weakness elsewhere. It works only if the build-out outruns the drag from consumer-facing products.

Meta turns the thesis into something testable

The clearest concrete checkpoint came on October 27, when AMD announced an expanded strategic partnership with Meta covering its superintelligence infrastructure. Helios systems had been installed at rack level at Meta and were being tested, according to the company, and AMD said it would take part in Meta's share placement to the tune of $1 billion.

That combination — systems in the field plus financial skin in the game — gives investors something firmer than a growth narrative to work with. The partnership gains economic weight only if the trials open the door to broader deployment, so the question is twofold: does the technology prove out, and does the tighter relationship translate into durable follow-on business?

The quarter behind the forecast

The operational foundation looks solid. AMD reported record third-quarter 2026 revenue of $13.4 billion, a 45% increase year over year. For the fourth quarter of 2026, the company guided to roughly $17 billion, with a possible variance of 3% either way. Su also said AMD expects total revenue growth of more than 80% for 2027.

Those targets raise the bar considerably. Hitting them would validate the strategy; missing them would leave the concentration problem exposed.

UBS lifts estimates, trims the target

How differently the outlook and the valuation can read is illustrated by UBS. The bank raised its adjusted earnings estimates and kept its buy rating — yet cut its twelve-month price target to $700 from $730. More confidence in the operating business, paired with a lower target, is a useful yardstick: it says the growth story is credible while the multiple required to own it remains demanding.

For anyone assessing the stock, the measure should be execution rather than ambition — additional delivery capacity, growth promises met, and a data-center business that reliably carries the weaker segments. That is the standard Su's strategy will ultimately be judged against.