AMD Falls 2.33% as Broadcom Holds Up Better in AI Selloff
AMD fell 2.33% and Nvidia 4.58% in Friday's close, while Broadcom gave up only 0.74% — a split that shows how the market separates AI infrastructure from an AI share-gain bet.

AMD closed 2.33% lower at 465.58 on Friday, Aug. 28, 2026, while Broadcom slipped just 0.74% to 368.79 and Nvidia dropped 4.58% to 217.55, illustrating how differently the market prices risk across two AI chip stories.
Two chip companies are selling into the same artificial-intelligence build-out, and the market treats them as if they were in different businesses. Friday's closing tape made the point cleanly: Advanced Micro Devices (AMD) finished at 465.58, down 2.33% from its prior close of 476.67, while Broadcom (AVGO) gave up only 0.74% to end at 368.79. Nvidia (NVDA) fell hardest of the three, down 4.58% to 217.55. All figures are as of the last trade at 20:00 GMT on Friday, Aug. 28, 2026; the market is closed.
The gap is not noise. On a day when the Nasdaq 100 tracker (QQQ) fell 0.65% to $716.43 and the S&P 500 tracker (SPY) slipped 0.23% to $769.35, AMD's decline was roughly three and a half times the Nasdaq 100's and about three times Broadcom's. That is the market's shorthand for beta: the same AI demand story, priced through two very different risk lenses.
Why the tape separates infrastructure from insurgency
Broadcom's role in AI is closest to plumbing. Its custom accelerator and networking franchise is embedded in hyperscaler data-center designs, and the revenue that comes with that position tends to arrive on multi-year programs rather than in quarterly share-shift battles. That is the profile Motley Fool describes as a steady AI infrastructure bet — and steadiness is exactly what the tape paid for on Friday.
AMD is the opposite trade. Its AI case rests on taking accelerator share from Nvidia, which means the stock is priced not on installed position but on a contested outcome. Higher risk, higher reward, in the words of the same framing. When sentiment toward AI hardware wobbles, a share-gain story reprices faster than an infrastructure story, because the thing being questioned is the premise, not the run rate.
There is a second mechanism at work. AMD's fortunes are correlated with Nvidia's narrative in both directions. A soft session for the category leader pulls the challenger down with it, since the same doubts about accelerator demand hit the company with less of the market and less of the pricing power. Nvidia's 4.58% drop from 227.98 and AMD's 2.33% slide in the same session are consistent with that linkage; Broadcom's near-flat finish is consistent with being priced off a different variable.
What the intraday ranges show about conviction
Closing percentages understate how the session actually traded. AMD ranged from 465.29 to 478.75 and closed near the bottom of that band — about 2.75% below its intraday high, on an illustrative calculation from the day's range. Nvidia ranged from 216.81 to 229.26 and finished roughly 5.11% under its high. Broadcom, which traded between 365.35 and 376.59, closed about 2.07% off its high.
All three sold down into the close, but the depth of the fade tracked the riskiness of the story. That ordering — Nvidia widest, AMD next, Broadcom tightest — is what a market looks like when it is trimming exposure to the most narrative-dependent parts of a theme rather than exiting the theme itself. The benchmarks support that reading: the Dow 30 tracker (DIA) was essentially unchanged at $535.06, down 0.03%, while the tech-heavy Nasdaq 100 took the bigger hit. Money moved within equities, not out of them.
The valuation question the price action implies
An investor comparing the two names is really choosing between two different sources of return. With Broadcom, the return comes from the durability of design wins already in flight and the cash those programs generate. The risk is that expectations are already high enough that merely meeting the schedule is not enough to move the stock.
An investor comparing the two names is really choosing between two different sources of return.
With AMD, the return depends on an event: winning enough accelerator sockets to change its revenue mix materially. The risk is symmetrical and larger. If the share shift happens, the upside is not incremental. If it stalls, there is far less underlying business to catch the fall, and the multiple built on the expectation unwinds quickly. Friday's asymmetry — a modest broad-market decline producing a 2.33% drop in AMD and a 0.74% drop in Broadcom — is a live measurement of that difference in fragility.
Neither profile is inherently better. They suit different portfolios. A reader who wants exposure to AI capital spending without underwriting a competitive outcome owns the infrastructure name. A reader willing to accept drawdowns of the kind seen Friday, repeatedly, in exchange for a shot at a re-rating owns the challenger. Owning both does not diversify the theme; it diversifies the risk expression within it.
What to watch from here
Three things will determine whether this pricing gap widens or closes. First, hyperscaler capital-spending commentary: it feeds Broadcom's program visibility directly and AMD's addressable opportunity indirectly. Second, any disclosure that quantifies AMD's accelerator traction — the stock's premium is a bet on a number that has to eventually be reported. Third, Nvidia's own trajectory, because the challenger trade cannot decouple from the incumbent's narrative while the incumbent still defines the category.
Also worth watching is whether the pattern repeats. One session in which AMD falls three times as far as Broadcom is a data point. If it holds across several down days for the Nasdaq 100, it tells you the market has settled on a stable risk premium between the two — and that premium, not the AI story itself, becomes the thing an investor is actually trading.
For now, the tape says the AI boom is intact and the market has simply become more particular about which balance sheet it wants to hold while the boom plays out.
Key facts
- AMD last close: 465.58, -2.33% (as of 20:00 GMT, Fri, Aug 28, 2026)
- AVGO last close: 368.79, -0.74% (as of 20:00 GMT, Fri, Aug 28, 2026)
- NVDA last close: 217.55, -4.58% (as of 20:00 GMT, Fri, Aug 28, 2026)
- Benchmarks: QQQ $716.43 (-0.65%), SPY $769.35 (-0.23%), DIA $535.06 (-0.03%)
Frequently asked questions
How much did AMD and Broadcom move in the latest session?
As of the last trade at 20:00 GMT on Friday, Aug. 28, 2026, AMD closed at 465.58, down 2.33% from a prior close of 476.67. Broadcom closed at 368.79, down 0.74% from 371.54. Nvidia fell furthest of the three, ending at 217.55, a decline of 4.58% from 227.98. The market was closed at the time of writing.
Why is Broadcom described as a steadier AI bet than AMD?
Broadcom's AI exposure sits in custom accelerators and data-center networking that are designed into hyperscaler systems on multi-year programs. That revenue is tied to schedules already in motion rather than to winning new competitive battles. AMD's AI case depends on taking accelerator share from Nvidia, an outcome that is not yet settled, so its stock reprices faster on sentiment shifts.
Did the broader market fall on the same day?
Yes, but far less. The S&P 500 tracker SPY closed at $769.35, down 0.23%, and the Nasdaq 100 tracker QQQ closed at $716.43, down 0.65%. The Dow 30 tracker DIA was almost unchanged at $535.06, down 0.03%. AMD's 2.33% decline was several times the Nasdaq 100's move, showing the selling was concentrated in AI hardware names.
How did the three chip stocks close relative to their intraday highs?
On an illustrative calculation from the supplied day ranges, AMD finished about 2.75% below its intraday high of 478.75, Nvidia about 5.11% below its high of 229.26, and Broadcom about 2.07% below its high of 376.59. All three faded into the close, with the deepest fades in the most narrative-dependent names.
Does AMD's price move depend on Nvidia?
Partly. Because AMD's AI thesis is built on taking accelerator share from Nvidia, doubts about Nvidia's demand outlook tend to hit AMD too — it has less of the market and less pricing power to absorb them. The two names fell together in the latest session, 4.58% and 2.33% respectively, while Broadcom held near flat.
Should an investor choose one or hold both?
They express different risks rather than different themes. Broadcom offers exposure to AI capital spending without underwriting a competitive outcome; AMD offers a larger potential re-rating in exchange for larger drawdowns. Holding both diversifies how the AI risk is expressed but not the underlying dependence on continued data-center spending. This is analysis, not investment advice.
Sources
- Broadcom or AMD? Here's Why the Market Is Pricing the Same AI Boom So Differently Across 2 Chip Stocks. — Motley Fool
Photo: Nicolas Foster · Pexels Licence — source


