The AI War Doesn't End with GPUs — The Secret Behind Cisco's $9B Order Surge
Summary
Cisco Systems (CSCO) reported record quarterly revenue of $15.84 billion for Q3 FY2026, representing 12% year-over-year growth, while simultaneously raising its AI infrastructure order target by 80% from $5 billion to $9 billion. All five major hyperscalers — Google, Microsoft, Amazon, Meta, and Apple — increased their Cisco orders by more than 100% year-over-year, confirming that AI data center investment has decisively shifted beyond GPU procurement into the networking infrastructure layer. On the same day as the record earnings announcement, Cisco disclosed the layoff of approximately 4,000 employees, exemplifying the emerging pattern in which AI-era corporate growth and mass workforce reductions operate as simultaneous, complementary strategies rather than contradictions. The company's shipment of its proprietary Silicon One G300 chip signals a deliberate push toward full-stack vertical integration of AI networking hardware, mirroring Apple's M-series silicon transition in both strategic intent and competitive implications. However, a critical margin paradox looms: AI infrastructure hardware carries 10-15 percentage points lower gross margins than Cisco's traditional high-margin software and services business, meaning the very success of its AI pivot may structurally compress profitability unless a rapid transition to high-margin subscription software offsets the hardware dilution.
Key Points
AI Infrastructure Orders Jump 80% in One Quarter — From $5B to $9B
Cisco entered FY2026 with a stated AI infrastructure order target of $5 billion — already an ambitious figure by historical standards for a company this size. But a single quarter later, the company raised that target to $9 billion, an 80% upward revision that reflects something far more significant than a single strong earnings quarter. The revision is concrete evidence that hyperscaler AI infrastructure investment has crossed a critical inflection point in the networking layer: all five major hyperscalers increased Cisco orders by more than 100% year-over-year simultaneously. The investment narrative around AI infrastructure from 2024-2025 was almost entirely GPU-centric, treating networking as a secondary consideration that would sort itself out. What Cisco's order data shows in stark terms is that in 2026, the bottleneck has structurally shifted — the constraint is no longer 'get more GPUs' but 'build the network fabric that lets those GPUs actually communicate at the scale AI training requires.' This is not a modest trend adjustment; it represents a reordering of AI capex priorities, and Cisco's order book is the first major primary data point confirming it at hyperscaler scale. Investors who have been watching GPU supply chains and semiconductor capacity as the primary AI infrastructure indicators need to update their frameworks.
Silicon One G300 — Cisco's Strategic Bet on Full-Stack Control
The shipment of Cisco's proprietary Silicon One G300 chip is not simply a product milestone — it is a strategic declaration about what kind of company Cisco intends to be in the AI infrastructure era. CEO Chuck Robbins made the competitive intent explicit when he publicly warned that 'AI infrastructure players without their own silicon will struggle to remain relevant,' a pointed statement directed at networking competitors who rely entirely on merchant silicon from Broadcom or Marvell. The Apple M-series analogy is genuinely instructive here: just as Apple regained control over its hardware performance roadmap by replacing Intel with custom silicon, Cisco is seeking an equivalent control advantage in AI networking hardware by owning the chip that runs its most critical systems. Silicon One integrates routing and switching functions on a single architecture, optimized for the low-latency, high-bandwidth requirements of AI workloads, reducing supply chain dependency while enabling tighter hardware-software co-design. This vertical integration strategy, if successfully scaled, transitions Cisco from a networking hardware assembler dependent on third-party components into a full-stack AI infrastructure platform company — a repositioning that justifies a meaningful valuation premium over commodity hardware vendors and creates structural differentiation that takes years for competitors to replicate.
Record Revenue and 4,000 Layoffs on the Same Day
Cisco's Q3 FY2026 earnings tell two stories that seem contradictory but are, in reality, expressions of the same underlying strategic logic. Revenue of $15.84 billion — a new all-time quarterly record, up 12% year-over-year — landed on the same day as the announcement of approximately 4,000 job cuts concentrated in traditional enterprise networking and switching divisions. These are the people who built and maintained the internet backbone for decades; they are being told that AI has shifted the value chain in ways that make their specific expertise less central to the company's future. This is the same pattern that Microsoft executed in 2024, that Google followed in cutting search-adjacent teams while accelerating Gemini investment, and that Meta deployed in eliminating middle management while scaling AI infrastructure. The 'AI pivot restructuring' playbook is: reduce headcount in legacy units, redirect capital to AI-aligned categories, and present the move to institutional investors as disciplined capital allocation. From an investor perspective, the logic is coherent. From a labor market perspective, it illustrates how the macro argument that 'AI creates jobs' provides cold comfort to the thousands of experienced engineers whose specific skills are being devalued in real time, before the promised new categories of AI-adjacent work have materialized at scale.
The Margin Paradox — How Winning Can Quietly Erode Profitability
Cisco built its financial identity on high-margin software licensing and services, consistently achieving 60%+ gross margins in its core business over the past decade. AI infrastructure hardware — optical transceivers, high-speed switching ASICs, data center fabric interconnects — operates at 45-50% gross margins, running 10-15 percentage points below the legacy business. As AI-related hardware grows to dominate the order mix, the product mix shift structurally depresses Cisco's blended gross margin, and the math compounds as the order pipeline scales up. This is the central financial tension in the Cisco bull thesis: the $9 billion order target revision generates obvious excitement, but the margin arithmetic behind it is quietly working against the company's profitability per revenue dollar. Cisco's answer to this paradox is to stack high-margin subscription software — ThousandEyes, Nexus Dashboard, AI-driven network automation — on top of the hardware layer, creating a blended margin that offsets the hardware drag. Whether that software transition happens fast enough to hold gross margins above 60% through FY2027-2028 is, in my judgment, the single most important financial variable determining whether the current stock valuation is a justified re-rating or a temporary premium that the earnings stream cannot sustain.
Networking as the Invisible Bottleneck of the AI Data Center
The public narrative around AI infrastructure constraints has centered almost entirely on GPU availability, semiconductor supply chains, and energy capacity — the visible, headline-generating constraints that are easy to communicate. But as AI model training scales to involve hundreds of thousands of GPUs communicating simultaneously inside a single training cluster, the networking layer connecting those GPUs becomes an equally critical constraint on overall system performance. Nvidia's next-generation GPU clusters require inter-node bandwidth of 800 Gbps to 1.6 Tbps — a specification that only a small number of networking vendors can actually meet at the required scale and reliability. Networking hardware accounts for 15-25% of total AI data center infrastructure costs, second only to compute, and inside any serious AI training cluster, the fabric interconnecting the GPUs determines whether those systems deliver their theoretical performance or spend cycles waiting on data transfer. Cisco, with 40 years of enterprise networking dominance and now a proprietary silicon platform, is the best-positioned incumbent to capture this layer as it transitions from an overlooked utility into an explicit competitive battleground. The market's reaction to the Q3 earnings — $60 billion of market cap added in a single session — confirms that the re-rating has begun, and in my view, the full revaluation of AI networking as a distinct investment category has not yet run its course.
Positive & Negative Analysis
Positive Aspects
- A 40-Year Enterprise Network Footprint Creates a Structural Moat
Cisco has spent four decades building the dominant position in global enterprise networking — a position that no AI-native startup or cloud-native competitor can replicate on a meaningful timeline. This is not simply about market share in a commodity sense; it is about deeply embedded customer relationships, multi-year service contracts, certified technical integrations across enterprise IT environments, and a global sales force trained to operate at the procurement level of the world's largest organizations. The fact that all five major hyperscalers increased Cisco orders by more than 100% simultaneously suggests that the company's technical capabilities and operational reliability have earned a level of trust at hyperscaler scale that newer entrants have not yet demonstrated under equivalent load and performance conditions. Over 100,000 enterprise customers globally represent natural upsell opportunities as those organizations build out their own private AI infrastructure, creating an installed base advantage that generates organic expansion revenue without proportional incremental selling cost. Cisco's competitive moat in AI networking derives not just from current technology but from the accumulated institutional trust and ecosystem integration that decades of incumbency creates — a structural advantage that is both genuinely defensible in the near term and genuinely difficult to quantify on a standard valuation model.
- Silicon One Enables Full-Stack Differentiation That Rivals Cannot Quickly Copy
The development and initial shipment of Silicon One G300 gives Cisco a competitive lever that the majority of its networking rivals lack entirely: the ability to co-optimize hardware and software on the same silicon foundation designed specifically for the networking use case. By reducing dependence on external chip suppliers, Cisco gains control over its performance roadmap in ways that were architecturally impossible when the company was assembling systems around commodity components from third parties. Custom silicon enables product differentiation on the dimensions that matter most in AI networking performance — latency, throughput per watt, programmability, and upgrade cadence — rather than competing purely on the quality of system integration around components that every competitor can also access. The Apple M-series transition provides a useful reference: performance and efficiency improvements from custom silicon created product advantages that competitors struggled to match for multiple product generations. Cisco has the engineering depth, the existing software stack, and the hyperscaler relationships to execute the same transition in enterprise AI networking. If Silicon One scales successfully, the company offers AI networking performance and energy efficiency that third-party-silicon-dependent vendors simply cannot match on the same deployment timeline.
- Recurring Revenue Base Provides Stability Against Hardware Cycle Volatility
Cisco has spent five years deliberately transforming its revenue model from one-time hardware transactions toward subscription and recurring revenue streams, and that transformation is now a meaningful structural advantage. Annual recurring revenue already exceeds 50% of total revenue, creating a cash flow foundation that remains stable even when quarterly hardware order flows fluctuate. This matters significantly in the context of AI infrastructure investment cycles, which — as the 2022-2023 cloud capex pause demonstrated — can decelerate sharply and with limited warning when hyperscaler procurement priorities shift. Cisco's subscription base means that even in a quarter where AI hardware orders disappoint, the core recurring revenue stream continues funding operations, research, and strategic investments that hardware-only competitors would be forced to cut. Products including ThousandEyes, Meraki AI, and Cisco Security Cloud represent the vanguard of a subscription software portfolio that, if aggressively expanded alongside AI hardware wins, can sustain gross margins in the 63-65% range even as hardware mix grows. This recurring revenue foundation changes Cisco's underlying risk profile compared to a pure hardware cycle company in a way that deserves explicit weight in any long-term valuation framework.
- All-Time High Stock Price Signals a Genuine Strategic Re-Rating, Not Just Momentum
The fact that Cisco stock hit $118.21 — surpassing the dot-com era peak that the stock spent more than 20 years unable to reclaim — represents something more substantive than momentum trading or FOMO-driven speculation. It reflects a fundamental shift in market categorization: Cisco is no longer being priced as a legacy infrastructure vendor managing its installed base toward a graceful decline, but as an AI infrastructure platform company with a structurally expanding addressable market ahead of it. The $60 billion single-session market cap addition indicates the degree to which the market had previously underestimated both the magnitude of the AI networking transition and Cisco's competitive position within it. This re-rating carries practical implications beyond the stock price itself: a stronger balance sheet creates more latitude for acquisitions, accelerated research and development investment, and talent retention programs that reinforce the AI transformation roadmap. A genuinely re-rated valuation is a compounding advantage — it reduces cost of capital, improves employee equity compensation competitiveness, and signals to prospective partners and customers that Cisco is a long-term committed player in AI infrastructure, not a transitional vendor to be eventually displaced.
- Early Position in a Multi-Year AI Infrastructure Cycle Creates Compounding Pipeline Advantages
The global AI infrastructure investment cycle is still in its early expansion phase, with the most capital-intensive buildout years arguably still ahead. Hyperscalers have announced aggregate AI capex plans exceeding $200 billion annually through 2026-2028, and the networking infrastructure portion of that spending is growing as a share of total AI data center investment as GPU cluster sizes scale and bandwidth requirements per node increase. Cisco has established itself as a primary supplier in this cycle early — locking in orders, validating Silicon One platform performance at hyperscaler scale, and building the technical reference architecture relationships that determine future procurement patterns for years after the initial win. AI data center investment is not a single capex event; it generates sustained follow-on orders for capacity expansion, next-generation equipment upgrades, and software management tools as GPU performance roadmaps and cluster architectures evolve. Each major early-cycle win creates the technical integration familiarity and certified reference designs that make the next win easier to close. The structural runway for AI infrastructure investment extends well into the next decade, and Cisco's trajectory suggests it is accumulating the platform-level position within that cycle that drives sustained multi-year revenue growth rather than one-time episodic upside.
Concerns
- Low Hardware Margins Create a Structural Profit Compression Risk
The core financial tension in Cisco's AI pivot is not about revenue growth — the $9 billion order revision makes clear that top-line growth is happening and will likely continue. The tension is entirely about what that revenue actually generates in terms of economic profit. Cisco's traditional software licensing and services business carried 60%+ gross margins, reflecting the economic characteristics of intellectual property and recurring service relationships where marginal delivery costs are low. AI infrastructure hardware — optical transceivers, high-speed switching ASICs, chassis and fabric systems — operates at 45-50% gross margins, a structural feature that does not improve through volume alone. As hardware orders grow faster than software revenue in the near term, the blended gross margin faces sustained downward pressure that is arithmetically unavoidable given the current revenue mix. The critical question is whether Cisco can build the subscription software layer — network intelligence, automation, management platforms — fast enough to offset the hardware dilution before 2027 analyst earnings expectations bake in margin improvement that hasn't yet materialized. If the software transition lags, earnings per share deteriorates even as the revenue headline continues growing, creating a trap for investors who anchor on top-line momentum as the primary valuation driver and miss the underlying profitability compression developing below the headline number.
- Hyperscaler Capex Dependency Creates Concentrated Cyclical Exposure
Cisco's $9 billion AI order target is, at its core, a claim on a defined portion of hyperscaler AI infrastructure spending plans for FY2026. That concentration creates a specific and non-trivial risk profile: if any or all of the major hyperscalers reduce AI capex — whether due to disappointing AI return on investment metrics, macroeconomic pressure, regulatory intervention, or the kind of capex correction that occurred in cloud infrastructure in 2022-2023 — Cisco's order trajectory deteriorates with limited internal buffer against the impact. The 2022-2023 cloud infrastructure capex pause inflicted real damage on networking hardware vendors precisely because hyperscaler spending decelerated faster than supplier production and delivery pipelines could adjust, leaving significant inventory overhangs that took multiple quarters to clear. A comparable dynamic in AI networking in 2027-2028 would hit Cisco harder than it would affect more diversified software businesses, because hardware fulfillment cycles involve longer lead times and less elastic demand forecasting. The company's revenue dependence on AI capex commitments from five customers — however creditworthy those customers are — represents a single-variable vulnerability that investors should model explicitly with bear case scenarios rather than assuming the current trajectory extends indefinitely.
- Arista Networks and Juniper Are Competitive Threats With Real Momentum
The AI networking market is not Cisco's uncontested domain, and the competitive risks are more serious than a surface-level market share analysis suggests. Arista Networks has achieved faster revenue growth in data center networking than Cisco for multiple consecutive years and has established strategic supplier relationships at Meta's AI clusters and Microsoft Azure AI infrastructure — exactly the reference accounts that establish procurement patterns for future hyperscaler expansion cycles. Arista's cloud-native EOS network operating system architecture provides an agility advantage in AI cluster environments where operators prefer software-defined network management over traditional hardware-centric Cisco configurations. HPE's acquisition of Juniper Networks brings another well-funded competitor with a deepening AI networking portfolio directly into the fight. Beyond traditional competitors, the more existential long-term threat is hyperscaler in-house networking development: Google already has significant proprietary networking silicon capability, and Amazon's Nitro architecture represents broad networking internalization at AWS scale. If either accelerates the transition away from third-party suppliers, Cisco's two largest customers become its most dangerous competitors simultaneously — a scenario affecting both revenue volume and broader market confidence in the stability of Cisco's hyperscaler relationships.
- Layoff-Driven Talent Risk May Undermine AI Execution Capability Over Time
Cutting 4,000 employees on the day of record earnings is financially defensible on a spreadsheet but organizationally risky in ways that don't appear immediately on an income statement. In Silicon Valley's current AI engineering talent environment, where competition for qualified AI network architects, silicon engineers, and distributed systems specialists is more intense than at any prior point in the industry's history, the reputational signal from a large-scale layoff can measurably degrade a company's ability to attract precisely the talent it needs most to execute the AI transition it is pivoting toward. The engineers being let go from traditional networking and switching divisions carry deep expertise in systems that underpin every existing Cisco customer relationship — losing that institutional knowledge creates service quality and support risks that may not be visible in financial results for 12-18 months. The churn risk extends beyond the laid-off workers to the anxiety spreading among remaining employees, who now recognize that record revenue provides no protection from restructuring when the AI transition demands it. Cisco's most valuable AI infrastructure engineers need to believe the company offers a stable long-term career path — and conducting a large layoff on the best financial day in the company's recent history makes that message demonstrably harder to deliver to both current and prospective talent.
- Elevated Valuation Prices In Multiple Simultaneous Execution Successes
The 15-20% single-day stock surge to an all-time high of $118.21 has materially compressed the forward error margin in Cisco's AI thesis, and investors entering at current levels need to understand what the current multiple actually requires. At a P/E of roughly 25-28x, Cisco is priced as an AI growth platform rather than a mature networking incumbent — and growth platform valuations assume sustained execution across multiple independent variables simultaneously: AI order growth continuing through the cycle, margins recovering on a defined schedule, Silicon One gaining broad market adoption, and subscription software scaling faster than hardware dilution. Any single one of those variables missing consensus creates a re-rating risk that was not present at lower multiples. The historical precedent is worth naming directly: Cisco fell from $82 to under $8 after the dot-com peak — a decline of more than 90% — when the gap between narrative expectations and financial reality became impossible to sustain. The current situation is not a direct analogy to 2000, but the lesson that Cisco specifically can become substantially overvalued relative to its actual growth cycle durability is encoded in the company's own market history. Investors entering at current levels should ensure their scenario modeling includes both a bear case probability and a realistic timeline assumption for the multiple to re-compress if the AI hardware investment cycle decelerates before the subscription software transition is complete.
Outlook
Looking at Cisco's near-term trajectory, the next 1-3 months hinge almost entirely on Q4 FY2026 earnings, expected around August 2026. The core question is whether the AI order momentum that produced the Q3 breakout — 80% upward target revision in a single quarter — can be sustained without the underlying hyperscaler demand signal reversing or decelerating. My expectation is that Q4 revenue clearing $16 billion is realistic, for a straightforward reason: hyperscaler AI capex commitments for the second half of 2026 were largely locked in before Q3 even closed, and those commitments flow into Cisco's order pipeline with limited short-term variability. The risk profile, however, is asymmetric to the downside from here: the stock is now priced for continued perfection. The 15-20% single-day surge almost certainly included a meaningful FOMO premium from momentum-driven investors, and a 5-10% technical pullback in the next 2-3 months would not surprise me even in a fundamentally healthy scenario. Strong support likely forms in the $100-$110 range, and any correction into that zone represents a more rational entry point for investors who missed the initial post-earnings surge.
Over the 6-month to 2-year timeframe, the most important structural variable is how the competitive landscape in AI networking evolves alongside Cisco's margin trajectory. The global AI data center networking market, in my view, will reach $35-$50 billion annually by the end of 2027 — roughly 2-3x its current run rate — as AI model training scales and enterprise AI adoption moves from pilot to production deployment. If Cisco maintains its estimated 35-45% market share through that expansion, AI-related annual revenue could reach $12-$20 billion, which is a genuinely transformative figure for a company that generated $53 billion in total revenue in FY2025. The threat from Arista Networks is real and deserves serious analytical weight: Arista has established strong footholds in Meta's AI clusters and Microsoft Azure AI infrastructure, and its cloud-native EOS architecture gives it responsiveness advantages in AI cluster environments that Cisco's more traditional enterprise sales motion can't easily replicate. HPE-owned Juniper Networks is also aggressively building its AI networking portfolio, creating a three-way competitive fight where market share is genuinely contested rather than predetermined.
On margins over the medium term, this is the variable that, in my view, separates the bull case from the bear case more than any other single factor. My base case projects Cisco's gross margin declining modestly from today's roughly 65% level to 62-63% by 2027 as the AI hardware mix grows in the revenue stack, then recovering toward 64-65% by 2028-2029 as subscription software revenue scales up and partially offsets the hardware dilution. The instruments of that recovery already exist within Cisco's portfolio: ThousandEyes for network intelligence, Nexus Dashboard for AI infrastructure management, and a growing range of AI-driven network automation platforms. The critical variable is execution speed — specifically, whether Cisco can accelerate subscription software adoption at hyperscaler accounts alongside hardware wins, rather than treating them as sequential motions. If the subscription transition lags and gross margins fall below 60% before the software layer scales, the stock faces a 20-30% correction scenario from current levels. Conversely, if the hardware-plus-software bundle strategy executes on schedule and gross margins hold above 65%, a realistic path to $150-$170 per share emerges as the street completes its re-rating.
Looking further out on the 2-5 year horizon, Cisco's ultimate trajectory depends on the magnitude of AI's displacement of traditional enterprise IT infrastructure spending. My projection is that AI-related categories will grow from roughly 15% of total enterprise IT budgets today to 35-40% by 2028-2030, driven by the combination of AI model deployment at enterprise scale, real-time inference at the edge, and the associated networking requirements of connecting distributed AI infrastructure. In that structural shift, networking infrastructure transitions from a commodity connectivity layer to a core strategic component of AI computing itself — enabling real-time inter-GPU communication, edge-to-datacenter data pipelines, and dynamic multi-cloud AI workload routing. The company that locks in the platform position at the beginning of this transition will hold it through the maturation cycle: reference architectures get embedded, integrations get certified, switching costs accumulate. If Cisco successfully scales its silicon strategy and completes the subscription software conversion, I see a credible path to $500-$600 billion market capitalization by 2030 — roughly 50-80% above today's levels.
The simultaneous record-profits-and-mass-layoffs pattern that Cisco exemplified in Q3 FY2026 will, I believe, become a standard template for Fortune 500 management over the next several years rather than remaining an exceptional event. By 2027-2028, more than half of Fortune 500 companies are likely to execute some version of what Cisco did here: reduce traditional business headcount by 10-20% under the banner of AI transformation, redirect that freed capital toward AI infrastructure and automation investment, and present the move to institutional investors as capital efficiency improvement. The social consequences of this playbook operating at scale are substantial and largely unpriced: large-scale dislocation of experienced workers from traditional IT and operations roles, surging demand for workforce retraining programs that current institutions aren't equipped to deliver, and growing political pressure on policymakers to redesign safety net systems for a labor market transition velocity the existing infrastructure wasn't designed to handle. This dimension will become an increasingly material variable for enterprise valuations as ESG frameworks and regulatory attention catch up with what AI-era workforce strategies actually look like in practice — and the companies that get ahead of this narrative will face meaningfully lower regulatory and reputational risk than those that don't.
Let me put three explicit scenarios on the table. In the bull case, Cisco delivers FY2027 revenue above $70 billion, AI-related revenue exceeds $20 billion, gross margins hold at 65%, and the stock reaches $160-$180 per share. The preconditions are: hyperscaler AI capex continues growing at 20%+ annually through 2027, Silicon One achieves broad market adoption, and subscription software acceleration outpaces hardware margin dilution. In this scenario, Cisco re-enters the S&P 500's top 20 by market cap — roughly 25% probability. The base case delivers approximately $65 billion in FY2027 revenue, $15 billion in AI revenue, 62-63% gross margins, and a stock range of $120-$140. Competition with Arista intensifies but doesn't break Cisco's share position; margins dip temporarily before recovering on subscription growth; the current valuation proves roughly fair to mildly expensive. This is the most likely outcome at roughly 50% probability, and the one I would anchor a long-term position around if Q4 earnings confirm the trajectory.
In the bear case, AI capex decelerates sharply in H1 2027 as hyperscalers face pressure to demonstrate ROI, AI revenue falls short at below $10 billion, margins compress to 59-60%, and the stock retraces to $85-$95. A macroeconomic recession, AI investment disappointment, or regulatory shock would be the trigger, and the historical precedent — Cisco falling more than 90% from its dot-com peak — is a necessary reminder that even dominant market leaders can get substantially overvalued relative to cycle durability. I assign this roughly 25% probability. For investors considering a position: staged accumulation tied to Q4 earnings validation, rather than chasing the post-earnings breakout, is the approach most consistent with the actual risk profile. There is also the scenario I should name honestly — hyperscaler in-house networking development accelerating, turning Cisco's largest customers into its most dangerous competitors — which remains a tail risk worth monitoring as Google and Amazon continue building internal silicon capabilities. This analysis is informational only.
Sources / References
- Cisco Q3 FY2026 Earnings Report: Record $15.84B Revenue, AI Infrastructure Orders Raised to $9B — CNBC
- Cisco Cuts Nearly 4,000 Jobs While Reporting Record Quarterly Revenue — TechCrunch
- Cisco Outlines $9B FY2026 Hyperscaler AI Infrastructure Orders and Q4 Revenue Guidance — Seeking Alpha
- Cisco Announces Silicon One G300: Next-Generation AI Networking Silicon Platform — Cisco Newsroom
- Cisco Q3 FY2026: AI Networking Momentum Drives Raised Outlook — Futurum Group
- Cisco CEO Warns AI Infrastructure Players Without Silicon Will Struggle to Stay Relevant — Benzinga