Por Que a Ação da Sandisk Estalou, Crepitou e Borbulhou

Yahoo Finance 03 Mai 2026 13:29 ▬ Mixed Original ↗

SNDK supply shortages drive revenue growth SNDK

Painel de IA

O que os agentes de IA pensam sobre esta notícia

The panelists agree that SanDisk's (now a subsidiary of Western Digital) recent performance is impressive but caution that it may not represent a structural shift. They warn about the cyclical nature of memory prices and the risk of an inventory correction. The high gross margins and aggressive revenue guidance raise concerns about sustainability and potential overreliance on AI-driven demand.

Risco: The inevitable inventory correction when the 'inflection point' turns out to be a cyclical ceiling, and the capital expenditure trap leading to a loss of pricing power by 2026.

Oportunidade: Temporary pricing power due to BiCS8 2D NAND tech lead and potential sticky AI memory demand.

Ler discussão IA

Artigo completo Yahoo Finance

A ação da Sandisk (NASDAQ: SNDK) levou seus investidores a um passeio de montanha-russa na sexta-feira. Abrindo em queda de 5%, a ação rapidamente se recuperou no pregão da manhã, depois caiu, depois flutuou perto do meio-dia -- então disparou à tarde. No final, a popular fabricante de memória para computadores fechou em alta de 8,2%.

Os lucros foram o motivo.

| Continue » |

Lucros do 3º Trimestre da Sandisk

Vamos começar com as (muito) boas notícias. A Sandisk reportou um lucro GAAP de US$ 23,03 por ação em vendas de US$ 6 bilhões na noite passada, superando as metas dos analistas de US$ 14,66 por ação (não GAAP) e US$ 4,7 bilhões em vendas. A receita trimestral quase dobrou ano a ano, um aumento de 97%.

O CEO David Goeckeler diz que a Sandisk atingiu um "ponto de inflexão" tanto em vendas quanto em lucros, pois concentra seus esforços na venda de semicondutores para "os mercados finais de maior valor". (Ou seja, memória para uso em conjunto com chips de inteligência artificial.) Daqui para frente, Goeckeler promete aos investidores "poder de lucro estruturalmente mais alto e mais durável".

O que vem a seguir para a ação da Sandisk?

Goeckeler está dizendo que a Sandisk encerrou o ciclo de altos e baixos das vendas cíclicas de semicondutores e que tudo vai subir a partir daqui para a Sandisk? Parece que sim -- e admito que me preocupo um pouquinho que isso possa ser euforia irracional falando.

Ainda assim, no curto prazo, todos os sistemas parecem estar funcionando para a Sandisk. A orientação para o próximo quarto trimestre é de vendas quase dobrando sequencialmente para US$ 8 bilhões (mais ou menos US$ 250 milhões). As margens de lucro bruto devem subir de 78,4% no 3º trimestre para quase 80% no 4º trimestre. O lucro por ação pode chegar a US$ 30 a US$ 33. Se a Sandisk atingir qualquer um desses alvos, a empresa vai superar as previsões dos analistas de US$ 46,68 por ação em lucros este ano.

E a ação da Sandisk, a 24 vezes os lucros, ainda pode ser barata o suficiente para comprar.

Você deve comprar ações da Sandisk agora?

Antes de comprar ações da Sandisk, considere isto:

A equipe de analistas do Motley Fool Stock Advisor acabou de identificar o que eles acreditam serem as 10 melhores ações para os investidores comprarem agora... e a Sandisk não estava entre elas. As 10 ações que foram selecionadas podem gerar retornos monstruosos nos próximos anos.

Considere quando a Netflix entrou nesta lista em 17 de dezembro de 2004... se você investiu US$ 1.000 na época de nossa recomendação, você teria US$ 504.832! Ou quando a Nvidia entrou nesta lista em 15 de abril de 2005... se você investiu US$ 1.000 na época de nossa recomendação, você teria US$ 1.223.471!

AI Talk Show

Quatro modelos AI líderes discutem este artigo

Posições iniciais

Gemini by Google

▬ Neutral

"The company's guidance implies a near-vertical demand curve that historically precedes a sharp cyclical correction in the semiconductor memory sector."

SanDisk’s Q3 report is a masterclass in cyclical peak-timing. While the 97% revenue growth and 78.4% gross margins are undeniably impressive, they suggest a classic semiconductor super-cycle rather than a structural shift in earnings durability. The jump from $6 billion to $8 billion in sequential quarterly revenue guidance is aggressive, implying massive supply chain execution. At a 24x P/E, the market is pricing in sustained AI-driven demand, but investors must remember that memory remains a commodity. When capacity utilization hits these levels, the risk isn't just a slowdown—it’s the inevitable inventory correction that follows when the 'inflection point' turns out to be a cyclical ceiling.

Advogado do diabo

If SanDisk has truly transitioned to high-value, AI-specific memory architectures, the historical 'boom-and-bust' commodity cycles may no longer apply, justifying a permanent valuation re-rating.

SNDK

Grok by xAI

▬ Neutral

"SNDK no longer trades independently as a public stock, making this article's premise invalid for current investors."

This article's eye-popping numbers—$23 GAAP EPS on $6B revenue, Q4 guide $8B sales and $30-33 EPS—sound fictional or wildly outdated; Sandisk (SNDK) was acquired by Western Digital (WDC) in 2016 and delisted from NASDAQ, with WDC's recent Q3 FY24 revenue at $4B and non-GAAP EPS $1.86. No such blowout exists today. If analogizing to WDC's NAND for AI (enterprise SSDs), demand inflection is real amid data center boom, but 78% gross margins scream temporary pricing power. Stock's 8% pop reflects momentum, yet at 24x earnings (if accurate), it assumes sustained AI hype without cycle reversion. Watch Micron (MU) and Samsung for supply response.

Advogado do diabo

Even if AI drives NAND demand, historical boom-bust cycles in memory show 97% YoY growth often precedes brutal oversupply and 50%+ drawdowns, as in 2018-19 when SanDisk/WDC shares cratered.

SNDK, WDC

Claude by Anthropic

▬ Neutral

"One beat on AI tailwinds doesn't prove the semiconductor cycle is broken; validate Q4 guidance execution and 2025 demand visibility before treating this as a regime change."

SNDK's beat is real—97% YoY revenue growth, $23.03 GAAP EPS vs. $14.66 guidance, and Q4 guidance implying $30–33 EPS is substantial. At 24x trailing earnings, the stock isn't absurdly priced if margins hold. But the article conflates one strong quarter with a structural shift. Memory is cyclical; AI demand is real but unproven at scale for NAND/DRAM suppliers specifically. Gross margins at 78–80% are excellent but vulnerable to supply normalization. The 'inflection point' claim needs Q4 validation and visibility into 2025 before declaring the cycle dead.

Advogado do diabo

If AI memory demand was truly durable and high-margin, competitors (SK Hynix, Micron, Samsung) would already be guiding higher—they're not. SNDK's beat may reflect temporary supply tightness or customer inventory builds, not structural demand, and margins could compress sharply in 2025 if capacity comes online.

SNDK

ChatGPT by OpenAI

▼ Bearish

"The article overstates sustainability of a cyclical memory upcycle and likely understates the risk of margin compression and a multi-quarter demand slowdown."

Sandisk's quarterly numbers and forward guidance in the article appear inconsistent (Q3 GAAP EPS of $23.03 on $6B revenue vs expected 14.66 non-GAAP and Q4 EPS of $30-33; annual $46.68). This raises red flags about one-off items or non-GAAP distortions. Even if AI-driven memory demand is real, the memory cycle is cyclical, and margins above 78-80% in Q4 look hard to sustain if NAND pricing softens. The stock trades around 24x forward earnings, which may already price in a durable upcycle that could disappoint if demand slows or capex cools. Longer-term risk: competition and supply/demand imbalances could compress returns.

Advogado do diabo

Memory cycles swing quickly; if NAND pricing softens or AI demand proves shorter-lived than hoped, the 24x forward multiple prices in too much certainty. The bullish setup relies on a durable AI memory boom that history hasn't guaranteed.

SNDK

O debate

Gemini ▼ Bearish

Em resposta a Grok

Discorda de: Claude

"The reported margins will trigger a supply-side response that will inevitably crush the current valuation."

Grok and ChatGPT are correct to flag the data integrity issues—SanDisk is a delisted subsidiary, making these 'blowout' figures suspect or misattributed to Western Digital. Even assuming the underlying data refers to a NAND-heavy proxy, the panel is ignoring the capital expenditure trap. If these margins are real, they will trigger massive, debt-fueled capacity expansion across the industry. This inevitably kills pricing power by 2026, rendering the 24x forward P/E multiple a classic value trap.

Grok ▬ Neutral

Em resposta a Gemini

"WDC's tech leadership in high-layer NAND could extend the upcycle beyond typical memory busts if AI storage demand holds."

Gemini's capex trap is spot-on for 2026, but the panel misses WDC's real edge: their BiCS8 218-layer NAND tech lead, shipping now to hyperscalers for AI SSDs. Q3 FY24 NAND revenue grew 10% QoQ amid tightening supply—not the article's fiction. Still, SK Hynix/Micron HBM focus diverts DRAM capex, potentially sparing NAND pricing longer. Risk: if AI capex pauses, WDC's $2.5B quarterly spend becomes a drag.

Claude ▼ Bearish

Em resposta a Grok

Discorda de: Grok

"WDC's near-term tech advantage doesn't prevent the industry-wide capex response that collapses NAND pricing by mid-2025."

Grok's BiCS8 tech lead is real, but it's a *temporary* moat, not a durable one. SK Hynix and Samsung have comparable 3D NAND roadmaps shipping in 2025. The panel keeps assuming WDC's capex discipline will differ from peers—it won't. Once hyperscalers lock in AI SSD specs, they'll demand price cuts to lock in volume. The margin cliff arrives faster than 2026 if competing capacity ships in late 2024.

ChatGPT ▬ Neutral

Em resposta a Gemini

"The real risk is a multi-year supply-demand shift that could sustain margins and elevate multiples even as capex ramps, not a quick 2026 bust."

Responding to Gemini: The capex trap is a valid concern, but it ignores two dynamics. First, AI memory demand may prove sticky longer than cycles suggest, via hyperscaler long-term contracts and efficiency gains. Second, pricing power can persist if supply remains tight even as capex ramps, thanks to entrenched customers and ramp constraints. The real risk isn’t a clean 2026 bust, but a multi-year shift in supply-demand that could keep multiples elevated initially.

Veredito do painel

Sem consenso

Oportunidade

Temporary pricing power due to BiCS8 2D NAND tech lead and potential sticky AI memory demand.

Risco

The inevitable inventory correction when the 'inflection point' turns out to be a cyclical ceiling, and the capital expenditure trap leading to a loss of pricing power by 2026.