Cisco's Super Cycle Meets Broadcom's Absurd AI Math

Cisco reported what 24/7 Wall St. called a networking super cycle quarter while Broadcom posted AI semiconductor results bordering on absurd, yet both stocks now trade at nearly the same valuation multiple; Broadcom closed at 392.99, down 5.94%, and Cisco at 111.68, down 1.58%, on Aug. 14, 2026.

Two of the biggest suppliers of physical plumbing for artificial intelligence just reported, and the market's response was to sell both. Cisco (CSCO) delivered what 24/7 Wall St. described as a networking super cycle quarter. Broadcom (AVGO) put up AI semiconductor numbers the same outlet called close to absurd. On the last trading day before this was written, Cisco closed at 111.68, down 1.58%, and Broadcom closed at 392.99, down 5.94% — a decline of 24.83 a share from the prior close of 417.82.

That is the tension worth sitting with. The two companies are on visibly different growth curves, but according to the 24/7 Wall St. comparison, they now sit at nearly the same valuation multiple. When two businesses growing at very different rates carry the same price tag on earnings, one of them is being mispriced. The question is which.

Why the same multiple on different growth curves is the whole story

A valuation multiple — price divided by earnings, whether trailing or forecast — is shorthand for how much investors will pay today for a dollar of profit tomorrow. Faster, more durable growth normally commands a higher number. Slower, steadier growth commands a lower one. That is the mechanism by which the market sorts a compounding franchise from a mature one.

When the sorting breaks down, it usually means one of two things. Either the market has quietly stopped believing the faster grower can sustain its pace, and is pricing in the deceleration in advance. Or the slower grower has been re-rated on a story — in this case, that AI infrastructure spending has turned a mature enterprise networking business into a cyclical growth business again.

Both readings are plausible right now, and they lead to opposite trades. If the fast grower's rate holds, paying a mature-company multiple for it is the better deal. If the fast grower's order book is being pulled forward from future years by a hyperscaler build-out that eventually normalizes, then the slower, cheaper, more diversified business is the safer place to sit.

Broadcom's 5.94% drop against a flat tape

The single most informative number in the market data is not either company's earnings — it is the shape of the selling. On Aug. 14, 2026, the S&P 500 tracker (SPY) closed at $776.34, off just 0.20% on the day. The Nasdaq 100 tracker (QQQ) closed at $731.07, down 0.14%. The Dow tracker (DIA) closed at $536.80, down 0.21%. In other words, the broad market barely moved.

Against that backdrop, Broadcom's 5.94% fall was not a market move. It was a stock-specific verdict. Its intraday range ran from 388.50 to 412.50 — a wide band that says positioning was heavy and opinions were being violently repriced rather than gently adjusted. Cisco's 1.58% decline, with a range of 110.06 to 113.68, looks far more like ordinary post-print digestion.

This is the pattern that has recurred through the AI hardware trade for several quarters: a company reports numbers that look extraordinary on the page, and the shares fall anyway, because extraordinary was already the base case. Expectations, not results, set the price. A stock can beat and still be too expensive relative to what was assumed.

What each business is actually selling into the AI build-out

The two companies sit at different points in the same data center. Broadcom's exposure is at the silicon layer — custom accelerators and networking chips designed with and for the largest cloud operators, where order visibility is enormous while it lasts and highly concentrated among a handful of customers. Cisco's exposure is at the systems and switching layer, plus a large installed base of enterprise and service-provider customers who buy on refresh cycles rather than on a race to stand up new compute capacity.

That distinction matters for how each revenue stream behaves when the cycle cools. Concentrated hyperscaler demand can be spectacular and then abruptly lumpy. Broad enterprise networking demand tends to be slower on the way up and slower on the way down. A super cycle quarter in networking is a genuinely strong signal precisely because that business does not normally produce them.

The checks to run before choosing a side

The comparison cannot be settled on price action alone. A few things determine which multiple is the mistake:

• Customer concentration. How much of the AI semiconductor growth depends on a small number of cloud buyers, and how much of that is committed versus indicated?
• Backlog quality. Orders that convert on a known schedule are worth more than orders that can be rephased. This is where a super cycle either proves itself or unwinds.
• Margin mix. Custom silicon and high-end switching carry very different gross margins from legacy hardware and software subscriptions. The multiple should follow the mix.
• Whether the multiple compressed or the earnings expanded. Two stocks converging on one number can happen because the expensive one got cheaper or the cheap one got dearer. The cause changes the conclusion.

A purely illustrative way to frame the risk: at 392.99, Broadcom would need to recover the 24.83 it just gave up simply to return to its previous close of 417.82. That is arithmetic, not a forecast — but it shows how much a single session can reset the entry price on a crowded position.

What to watch from here

Watch whether Cisco's networking strength shows up as a second consecutive strong quarter or reads as a one-off pull-forward. Watch whether Broadcom's AI revenue growth rate decelerates in sequential terms even while the absolute numbers stay large — that is usually where the multiple gets decided. And watch the relative performance of the two against the broad indexes, which were essentially flat into the Aug. 14 close. When benchmarks are still and a single mega-cap moves nearly 6%, the market is telling you where the disagreement lives.

Key facts

• CSCO last close: 111.68, -1.58% (Aug. 14, 2026, 20:00 GMT)
• AVGO last close: 392.99, -5.94% (Aug. 14, 2026, 20:00 GMT)
• Broadcom day range: 388.50–412.50, prior close 417.82
• Market backdrop: SPY $776.34 (-0.20%), QQQ $731.07 (-0.14%), DIA $536.80 (-0.21%)

Frequently asked questions

How did Cisco and Broadcom shares perform after their latest results?

Both fell. Cisco closed at 111.68, down 1.58%, with a session range of 110.06 to 113.68. Broadcom closed at 392.99, down 5.94%, from a prior close of 417.82, with a range of 388.50 to 412.50. Those were the last traded prices as of Aug. 14, 2026 at 20:00 GMT.

Why does it matter that both stocks trade near the same multiple?

A valuation multiple compares price to earnings and normally reflects growth. According to 24/7 Wall St., Cisco and Broadcom now sit at nearly the same multiple despite wildly different growth trajectories. When growth rates differ but prices do not, one of the two is likely mispriced — either the faster grower is cheap or the slower one is expensive.

Was Broadcom's decline part of a broader market selloff?

No. On the same session, the S&P 500 tracker closed down 0.20%, the Nasdaq 100 tracker down 0.14% and the Dow tracker down 0.21%. With benchmarks essentially flat, Broadcom's 5.94% drop was a stock-specific repricing rather than a market-wide move.

What did each company report?

Per the source, Cisco delivered what was characterized as a networking super cycle quarter, while Broadcom posted AI semiconductor figures described as bordering on absurd. Specific revenue and earnings line items were not disclosed in the material available, so the comparison here rests on the reported characterizations and on verified closing prices.

How do the two businesses differ within AI infrastructure?

Broadcom sits closer to the silicon layer, supplying chips and networking components tied heavily to large cloud operators, which concentrates demand. Cisco sits at the systems and switching layer with a broad enterprise and service-provider installed base that buys on refresh cycles. That makes Broadcom's revenue potentially lumpier and Cisco's steadier.

What should investors watch next in this comparison?

Three things: whether Cisco's networking strength repeats in a second quarter or proves to be pulled-forward demand; whether Broadcom's AI growth rate decelerates sequentially even as absolute revenue stays large; and customer concentration and backlog quality at both firms, since committed orders are worth more than indicated ones.

Sources

• Cisco vs. Broadcom: One Stock Looks Like the Better AI Play — 24/7 Wall St

Photo: Brett Sayles · Pexels Licence — source