AMD or NVDA? Price vs Value After the 2026 Rally

2026 told two different stories on the tape. AMD shares more than tripled and crossed the $1T mark, a rally that rewarded holders handsomely. Nvidia shares added about 29 percent yet stayed the largest listed company near $5.8T. Stretch the window to five years and the picture flips: Nvidia up more than 1000 percent, AMD up near 500 percent. One won the sprint, the other the marathon; the investment question is which price embeds less optimism today.

The size gap is stark. Nvidia books roughly seven times the annual sales of AMD and more than ten times the cash generation. Per the comparison in the discussion, Nvidia keeps about 65 cents of each revenue dollar while AMD keeps near 16 cents. Balance sheets are both comfortable, with Nvidia able to cover debt in months and AMD in a little over a year. The lesson for holders: small market value does not mean cheap; profit per share and the multiple on it decide.

The same wind lifts both boats: AI data centers. AMD data-center sales more than doubled and now form over half of revenue. Nvidia grew the same unit near 117 percent and it forms over 90 percent of sales, with Nvidia data-center scale more than ten times AMD. Picture one wheat field with two harvesters, a giant machine cutting nearly the whole field and a smaller one growing fast. The pool is shared but the margins are not.

Markets pay for the future, not the past. AMD guides near $13B next quarter while Nvidia guides near $108B for a single quarter, assuming zero China sales. Street projections see AMD running toward $90B with profits more than doubling, and Nvidia approaching $680B. AMD may grow faster in percent, yet Nvidia could add more new sales in one year than AMD produced in five. What matters is incremental profit per share against the price paid.

AMD played offense on the customer side. Reuters reports a Meta framework reaching toward $60B of processor supply over five years, alongside a similar OpenAI framework. The shareholder detail, laid out in Motley Fool analysis, is two warrants up to 160 million shares each, 320 million in total with near 20 percent dilution risk. SEC filings show a one-cent exercise tied to gigawatt deliveries and a $600 price threshold. Dilute today to lock demand tomorrow; the test is whether revenue outruns the extra count.

AMD is also patching its software gap with Anthropic, whose team helps tune compatibility on AMD products. It set about $8B for the Fei-Fei Li startup venture, paid mostly in its own shares, and kept taking server share from Intel toward 35 percent. The customer list strengthens the story while share-based payment pressures per-share value. That trade sits at the center of any fair multiple.

Nvidia does the opposite: it retires shares instead of printing them. It added $150B to the buyback program and the share count keeps shrinking, so each remaining share claims more profit. CNBC reports Nvidia is also buying Hugging Face for $12.9B, with Nvidia pledging the platform stays open. Fewer shares flatter per-share profit and soften the multiple; buying dear is less powerful than buying cheap, but the direction is clear.

On technology Nvidia pushes the new stack into the field. Nvidia Investor updates say the Vera Rubin platform moved into full production with shipments starting in the fall across major server makers. The company also fronts a financing architecture mobilizing near $500B of outside capital for data-center build. Today's revenue funds tomorrow's capacity; the open question is utilization and the burden that capacity puts on the multiple.

The AMD bull case is simple: being the second source can be enough. OpenAI, Meta and Microsoft deploy AMD products, and the Nvidia pie is so large that a thin slice could double AMD profit with rising margins. The bear case is equally simple: margins may never approach Nvidia levels, and if the full 320 million shares vest, profit per share stays pressed. A small slice can feed you, but not if you hand over a fifth of the register for it.

The Nvidia bull case is completeness: silicon, networking and software under one roof, with CUDA loyalty holding customers and a fresh series each year. The bear case is concentration: over 90 percent of sales from one unit, a few large buyers behind much of it, and those same buyers building their own silicon. China is shut today, pure upside if it reopens. Strong balance sheet, single-leg structure.

Side-by-side numbers favor the giant. On trailing figures AMD screens near 125 times cash flow and 160 times earnings against Nvidia near 40 to 45 times cash flow and 30 times earnings, with far stronger capital returns at Nvidia. Margins tell the same story, Nvidia in the 50 to 60 percent band in recent years, AMD in the 10 to 15 band. On next-year earnings AMD costs near 60 times against Nvidia near 20 times, three times more per profit dollar. The larger name looks cheaper on nearly every yardstick.

Street paths diverge. Nvidia revenue may nearly triple in a few years while earnings per share roughly double, with some margin fade assumed. AMD revenue may rise fivefold with earnings per share up sixfold, fast growth off a small base. The ten-year exercise in the video lands near $326 for AMD and $540 for Nvidia in the mid case, with AMD returns from negative to low single digits at the current price and Nvidia in the mid to high teens. AMD is priced for perfection; Nvidia sits closer to its middle path.

The moral is price versus value. A fine story bought at the wrong price becomes a poor investment, which is why a margin of safety matters. The calendar is set: AMD near November 3, Nvidia near November 18, when cash, margin and orders matter more than the headline. I judge the two giants the same way: match the market price against next-year profit per share and pick the one whose price earns its keep.