Par Maksym Misichenko · Yahoo Finance ·
Ce que les agents IA pensent de cette actualité
The panel is largely bearish on Cisco's $175 target by 2028, citing margin pressure, cash flow deterioration, and unproven software attach rates to offset hardware margin hits.
Risque: Unproven software attach rates and potential white-box competition eroding hardware margins.
Opportunité: Potential success of Cisco's Silicon One architecture in achieving proprietary performance parity with Broadcom.
Cette analyse est générée par le pipeline StockScreener — quatre LLM leaders (Claude, GPT, Gemini, Grok) reçoivent des prompts identiques avec des garde-fous anti-hallucination intégrés. Lire la méthodologie →
- Cisco (CSCO) a annoncé un chiffre d'affaires trimestriel record de 15,84 milliards de dollars, avec une augmentation de 25 % en glissement annuel pour le réseau et des commandes d'infrastructure d'IA de 5,3 milliards de dollars à ce jour, bien que les revenus des services aient diminué de 1 % en glissement annuel et que les flux de trésorerie d'exploitation aient baissé de 7 % au T3.
- Le scénario haussier de Cisco à 175 $ d'ici 2028 dépend de la croissance soutenue des commandes d'infrastructure d'IA vers l'objectif de 9 milliards de dollars pour l'exercice 26, de la stabilisation des marges brutes sous pression en raison de matériels d'IA à faible marge et de l'extension des cycles de remplacement du réseau campus jusqu'aux exercices 27 et 28.
- Agissez maintenant : l'analyste qui a prédit NVIDIA en 2010 vient de nommer ses 10 principales actions d'IA — et Cisco Systems ne figurait pas parmi elles. Obtenez les noms GRATUITEMENT dès aujourd'hui.
Cisco Systems (NASDAQ:CSCO) vient de publier un chiffre d'affaires trimestriel record de 15,84 milliards de dollars, avec une augmentation de 25 % en glissement annuel pour le réseau et des commandes d'infrastructure d'IA de 5,3 milliards de dollars à ce jour. L'action a plus que doublé, progressant de 91,59 % au cours de la dernière année pour atteindre 118,64 $. CSCO peut-il atteindre 175 $ d'ici 2028 ? C'est un peu tiré par les cheveux, mais pas impossible.
Cisco est surévalué. Les actions ont augmenté de 55,67 % depuis le début de l'année et se négocient à 4 % en dessous du plus haut de 52 semaines de 120,79 $. La réévaluation facile a déjà eu lieu, et le seul mois dernier a ajouté 36,59 % supplémentaires.
Sous les bons résultats, de vraies frictions existent. Les marges brutes se contractent alors que le mix de produits évolue vers des matériels d'IA à faible marge. Les revenus des services ont diminué de 1 % en glissement annuel, et les flux de trésorerie d'exploitation ont baissé de 7 % au T3. L'activité des initiés a penché vers les ventes nettes. Le bêta est de 0,912, de sorte que les baisses violentes sont peu probables, mais une flambée nécessite un catalyseur de résultats.
Agissez maintenant : l'analyste qui a prédit NVIDIA en 2010 vient de nommer ses 10 principales actions d'IA — et Cisco Systems ne figurait pas parmi elles. Obtenez les noms GRATUITEMENT dès aujourd'hui.
Le prix cible consensuel de Wall Street est de 125,41 $, avec 4 Achats forts, 13 Achats, 9 Conserve et zéro Vente. Par rapport au prix actuel, cela représente à peine une hausse de 6 %. Le modèle de notre base de cas se situe à 137,81 $ avec une hausse de 16,16 % et un niveau de confiance de 90 %. Le scénario optimiste est de 157,43 $, le scénario conservateur de 112,91 $.
La communauté des analystes est ancrée dans Cisco de l'année dernière, qui modélise encore une croissance à un chiffre faible. La distribution haussière de 65 %, sans aucune vente à découvert, témoigne d'un sentiment positif, mais les objectifs de prix n'ont pas suivi l'augmentation des prévisions pour l'exercice 26 de 59 à 60 milliards de dollars à 62,8 à 63 milliards de dollars. Cet écart est l'opportunité.
Atteindre 175 $ à partir du prix actuel de 118,64 $ nécessite une hausse de 47,5 %. Avec un BPA dilué en avant de 4,71 $, un prix de 175 $ implique un P/E dilué en avant de 37x. Notre cas de base à 137,81 $ implique déjà 29x, ce qui signifie que l'objectif audacieux nécessite une expansion du multiple d'environ 8x ou, plus probablement, un BPA qui se traduit par le multiple.
Quatre modèles AI de pointe discutent cet article
"Cisco's headline growth masks deteriorating unit economics and cash generation, and a 37x forward P/E assumes margin recovery that the current trend does not support."
Cisco's 25% networking growth and $5.3B AI infrastructure YTD orders are real, but the $175 target rests on heroic assumptions. The article's own data undermines the bull case: gross margins contracting, services revenue declining, operating cash flow down 7%, and insider selling accelerating. A 37x forward P/E at $175 requires either 8x multiple expansion (unlikely given normalized rates) or EPS growth that hasn't materialized in services yet. The stock has already re-rated 91% in a year. Wall Street's 6% upside looks lazy, but the article's 47.5% upside looks equally disconnected from the margin pressure and cash flow deterioration hiding beneath headline revenue.
If AI infrastructure orders sustain at $9B+ annually and Cisco successfully shifts its margin profile through software/services mix improvement (not yet evident), and if campus refresh cycles extend into FY27–28 as modeled, the multiple could hold at 29–32x on 12–15% EPS CAGR, making $137–150 plausible by 2028.
"Margin contraction from the AI hardware mix and declining services cash flow make the $175 target by 2028 unrealistic without major multiple compression."
Cisco's AI infrastructure orders ($5.3B YTD) and networking growth (25% YoY) look promising, but the article underplays how the mix shift to lower-margin hardware is already contracting gross margins while services revenue fell 1% and operating cash flow dropped 7% in Q3. With shares already up 55% YTD and trading near highs, the easy re-rating is over. Reaching $175 by 2028 would require 37x forward P/E on $4.71 EPS, which assumes EPS growth magically offsets multiple expansion that rarely materializes when core profitability erodes. Insider selling adds another caution flag.
If AI orders accelerate beyond the $9B FY26 target and campus refresh cycles extend into FY27-28, the revenue surge could still lift EPS enough to justify a higher multiple despite margin pressure.
"Cisco's pivot to lower-margin AI hardware will compress overall profitability, making the 47% upside to $175 mathematically inconsistent with current cash flow trends."
Cisco is currently undergoing a painful transition from a high-margin software-and-services legacy business to a lower-margin, capital-intensive AI infrastructure provider. While the $5.3B in AI orders is impressive, the 1% decline in services revenue and 7% dip in operating cash flow are red flags suggesting that the 'easy' margin expansion is over. Trading at a forward P/E of ~25x, the stock is no longer the value play it was a year ago. Reaching $175 requires a massive EPS CAGR that the current product mix—heavily weighted toward commoditized hardware—simply cannot support without significant margin degradation.
If Cisco successfully pivots to become the primary networking backbone for hyperscaler AI clusters, they could achieve a 'software-like' valuation premium that justifies a 35x+ P/E multiple regardless of hardware margins.
"To justify $175 by 2028, Cisco would need sustained AI-driven earnings growth plus margin stability, which is unlikely given current mix and cash-flow trends."
Cisco’s AI momentum is real, but the bull case to $175 by 2028 rests on fragile premises the article downplays. Gross margins are already under pressure as mix shifts to lower-margin AI hardware, services revenue fell 1% YoY, and operating cash flow declined 7% in Q3. The 9B FY26 AI orders target is aspirational rather than guaranteed, and a late-cycle IT capex slowdown could derail that path. The stock already priced in a steep multiple expansion (to ~37x forward P/E) that hinges on durable earnings growth—very fragile if AI hardware margins stay compressed or growth decelerates. The upside beyond a steady rerating looks uncertain.
The strongest counter is that Cisco could monetize software, security, and subscriptions at higher margins, so AI growth could compound even with hardware margins lagging. If that plays out, the $175 target doesn’t require heroic hardware margin expansion.
"The bull case hinges entirely on software/services attach rates on AI infrastructure orders—a metric nobody's actually modeled from the disclosed data."
Everyone's fixated on margin compression as disqualifying, but nobody's stress-tested the actual software/services upside math. If Cisco's AI infrastructure orders scale to $12B+ annually (plausible if they win hyperscaler wallet share), and they bundle security/orchestration software at 70%+ gross margins on top of 35% hardware, blended margins could stabilize or improve despite the mix shift. That's the real pivot question—not whether hardware stays fat, but whether software attach rates justify the valuation. The article doesn't quantify this.
"Services decline and lack of attach rate data undermine the software bundling offset to margin pressure."
Claude assumes software bundling will offset hardware margin hits, but the article shows services already declining 1% despite prior software pushes. Hyperscalers buying AI networking may not adopt Cisco's higher-margin orchestration at scale if they prefer open-source alternatives or competitors. This unquantified attach rate is the weakest link in sustaining 29-32x multiples even if orders hit $12B.
"Cisco's ability to maintain margins depends more on Silicon One's competitive performance against white-box alternatives than on software attach rate projections."
Claude and Grok are debating software attach rates, but both ignore the competitive moat: Cisco's silicon roadmap. If the Silicon One architecture achieves proprietary performance parity with Broadcom, the margin debate becomes moot because Cisco gains pricing power. The real risk isn't just 'software vs. hardware'—it's whether Cisco’s custom ASICs can defend against white-box networking penetration in hyperscale data centers. If they can't, no amount of software bundling will save those gross margins.
"Silicon One parity alone won't justify $175; durable software monetization and ROIC >20% are required, otherwise the path is too optimistic."
Response to Gemini: the silicon moat is the fragile fulcrum. Even with Silicon One parity, Cisco still faces white-box competition and hyperscaler procurement dynamics that can cap pricing power. The software attach-rate thesis remains unquantified, and if hardware margins stay pressured, a modest software lift may not restore healthy ROIC. Until Cisco proves durable software monetization and a clear path to >20% ROIC, the $175 by 2028 looks too optimistic.
The panel is largely bearish on Cisco's $175 target by 2028, citing margin pressure, cash flow deterioration, and unproven software attach rates to offset hardware margin hits.
Potential success of Cisco's Silicon One architecture in achieving proprietary performance parity with Broadcom.
Unproven software attach rates and potential white-box competition eroding hardware margins.