2026-08-13

Cisco (CSCO) Sinks 4.5% After Hours While Supermicro (SMCI) Jumps 17.6% - Two AI-Linked Earnings Beats, Two Opposite Verdicts

Written & reviewed by the TradeSmrt editorial team · 2026-08-13

In this article

What Happened
Why the Same Kind of Beat Read So Differently
What to Take Away From This
FAQ
Sources

What Happened

Two AI-adjacent stocks reported earnings within a day of each other this week - and got about as different a reaction as Wall Street can produce. Cisco Systems (NASDAQ: CSCO) posted fiscal fourth-quarter 2026 results after Wednesday's close, August 12. Adjusted earnings per share came in at $1.22, above the roughly $1.13-$1.17 analysts had modeled, and revenue of $17.25 billion cleared consensus near $16.8 billion. Revenue grew about 18% year over year and net income jumped 51%. On top of that, Cisco guided fiscal 2027 to $5.05-$5.11 in adjusted EPS and $72.2-$73.4 billion in revenue - both dramatically above Street estimates of roughly $4.28 and $62.9 billion. On paper, this was a textbook beat-and-raise.

The stock did the opposite of what a textbook beat-and-raise usually produces. Shares slid from a regular-session close of $123.88 to as low as $118.25 in after-hours trading, a drop of more than 4.5%. Every headline number topped expectations, and the stock still fell - a classic "sell the news" reaction.

One day earlier, after Tuesday's close on August 11, AI server maker Super Micro Computer (NASDAQ: SMCI) reported its own fiscal fourth-quarter 2026 results. Revenue came in at $11.12 billion, nearly double a year earlier and roughly in line with the ~$11 billion analysts were modeling. What actually stunned the market was profitability. Non-GAAP EPS hit $1.70, blowing past even the top end of the analyst range, which had run from $0.62 to $0.96. Non-GAAP gross margin landed at 17.6% - more than double the 8.2%-8.4% range the company itself had guided just a month earlier. On top of that, Supermicro booked more than $60 billion in new orders during the quarter, pushing its backlog to a record, and guided fiscal 2027 revenue to $65-$72 billion, roughly 30% above Street estimates. The reaction was immediate: shares jumped more than 10% after hours and carried that momentum into the next session, closing up 17.6% at $37.15.

Both companies beat expectations. Both sit squarely inside the AI infrastructure buildout theme. And within 24 hours of each other, one stock dropped over 4% and the other jumped nearly 18%.

Why the Same Kind of Beat Read So Differently

The key isn't the earnings print itself - it's what was already baked into each stock going in. Cisco had rallied roughly 60% year-to-date heading into the report, meaning a heavy dose of AI-networking optimism was already priced in. When a stock is already running that hot, "beating consensus" isn't enough on its own - it has to clear a second, unspoken bar: the expectations the market had quietly built on top of consensus. Judged against that higher bar, Cisco's numbers, strong as they were, could easily read as merely meeting expectations rather than exceeding them - or as a modest letdown relative to how much good news the stock had already absorbed.

Margin guidance added a real drag on top of that. Cisco guided fiscal 2027 non-GAAP gross margin to 65.5%-66.5%, which some analysts flagged as a step down from recent levels. Rising AI-related orders - the company lifted its cumulative order target to roughly $9 billion from a prior $5 billion - are unambiguously good news for revenue, but AI hardware and networking equipment typically carry thinner margins than Cisco's software and subscription businesses. A growing mix of lower-margin AI revenue reads, to some investors, as dilution of overall profitability. When growth arrives but its quality gets questioned, the stock can fall even as the headline numbers rise.

Supermicro was starting from the opposite place entirely. The company spent much of the past year and a half working through an accounting controversy, a Nasdaq delisting scare, and repeated guidance cuts - all of which had crushed investor confidence well before this report. That history is exactly why Supermicro's own gross margin guidance a month earlier was so conservative, sitting in the low-8% range. When the actual result came in at more than double that guided figure, the market didn't read it as a routine beat - it read it as evidence that the underlying business was normalizing faster than even management had expected. When expectations have been ground down that low, even a modest improvement produces an outsized reaction, let alone one that doubled the company's own forecast. The $60 billion-plus in new orders reinforced the point with something concrete: signed contracts, not forward-looking commentary, confirming that real demand for AI servers remains intact.

The same kind of beat, in other words, meant something completely different depending on the starting line. Cisco's was "a modest surprise layered on top of an already-optimistic price." Supermicro's was "a dramatic surprise layered on top of rock-bottom expectations." Wall Street's formal ratings echoed the divergence without predicting it: Cisco carried a consensus Buy rating from 26 analysts and an average price target near $132.59 - implying roughly 8% further upside from the pre-earnings close - yet that bullish consensus wasn't enough to prevent the after-hours slide. Supermicro, by contrast, carried an average target of $37.25 from 19 analysts with a consensus rating that remained at Hold, not Buy - and still jumped enough in a single session to land right near that target. In both cases, the surprise relative to expectations moved the stock far more than the formal analyst rating did.

What to Take Away From This

"Good earnings" and "the stock goes up" are two different claims. As Cisco shows, revenue, profit, and guidance can all clear consensus and the stock can still fall. What matters is whether the numbers cleared the expectations already embedded in the price - which means checking how far a stock has already run before the print, not just the headline comparison to consensus.
Read the margin line in guidance, not just revenue and EPS. Cisco's headline guidance was strong, but a modest step-down in gross margin guidance was enough to raise "quality of growth" concerns. This matters especially for companies whose revenue mix is shifting toward lower-margin hardware or infrastructure.
Unusually conservative guidance from a company can be an expectations reset, not a real forecast. Supermicro's low-8% gross margin guidance a month before the print turned out to be far too conservative. A company still working to rebuild credibility - after an accounting issue, say - may guide deliberately low, which can set up outsized "beats" later. That's worth factoring in when a subsequent report looks almost too good.
New orders and backlog are a more reliable forward indicator than management commentary. Supermicro's $60 billion-plus in fresh orders is a far more concrete, verifiable signal than optimistic language on an earnings call. When evaluating a print, backlog trends deserve as much attention as the current quarter's headline numbers.
The direction and size of the surprise moves a stock more than the formal analyst rating does, in the short run. Neither stock's pre-earnings consensus rating predicted the direction of its reaction. For short-term positioning, gauging how far actual results diverge from the market's implicit expectations is more useful than the official rating itself.

FAQ

Does Cisco's stock drop signal that AI infrastructure spending is slowing down?

Not really. Cisco actually raised its cumulative AI order target from $5 billion to roughly $9 billion, and its fiscal 2027 revenue guidance came in well above Street estimates. The stock's decline looks less like a business problem and more like a "priced for perfection" reaction - the shares had already run up nearly 60% in 2026, and a modest step-down in gross margin guidance was enough to disappoint holders who had baked in flawless execution.

Is Supermicro's rally sustainable?

The gross margin improvement and $60 billion-plus in new orders are genuine signs of business normalization, not just a one-quarter blip. That said, Supermicro is still working to rebuild trust after its earlier accounting controversy, and the fact that the average analyst rating remains Hold rather than Buy across 19 covering firms suggests real caution persists around valuation and execution risk.

What do these two reports say about the AI investment boom overall?

Both companies' underlying results point the same direction: demand for AI infrastructure remains solid. Cisco's higher cumulative AI order target and Supermicro's record new orders both cut against the narrative that AI capital spending is cooling. But the diverging stock reactions are a reminder that even genuinely good news inside the AI theme can be digested very differently depending on how much optimism a given stock had already priced in beforehand.

Sources

This article is an original synthesis and analysis based on the reporting below, not a reproduction of the original articles. Please check the source articles directly for the most current figures.

⚠️ This article is for informational purposes only and is not investment advice. Market conditions change constantly - always verify the latest data before making investment decisions.