AI Orders Double the Target and Margins Hit a Record — But In-Line FY26 Guide Caps the Upside at the Highs
Key Takeaways
- Webscale AI infrastructure orders exceeded $800M in Q4 and $2B for the full year — more than double Cisco's original $1B FY25 target — while Q4 product orders grew 7% across all geographies. The architectural-shift thesis is no longer aspirational; it is in backlog.
- Q4 non-GAAP gross margin landed at 68.4% (+50 bps YoY) and non-GAAP operating margin at 34.3%, both at the high end of guidance, with EPS above the high end — a clean print on operating leverage, but the magnitude of the beat (~1¢ on EPS, ~$50M on revenue) was modest relative to a stock entering the print near 52-week highs.
- FY26 guidance — revenue $59.0–60.0B and non-GAAP EPS $4.00–4.06 — is roughly in line with Street and includes the estimated impact of current tariffs; the midpoint implies ~5% revenue growth, decelerating from the Q1 FY26 guide of ~6–7% growth, which management framed as a year-over-year comp issue rather than a demand shift.
- Security remains the swing factor: Splunk synergies drove a 14% YoY increase in new logos and the new/refreshed product family added 750 customers in the quarter, but headline security order growth of mid-single-digits is well below the long-term 15–17% Security+Observability framework, with U.S. Federal weakness still masking double-digit ex-Federal momentum.
- Rating: Initiating at Hold. The AI order book, record margins, and 94% FCF return are real positives, but with the stock near 52-week highs and the FY26 guide in line rather than ahead, the risk/reward is balanced — we need to see acceleration in security run-rate or a campus-refresh inflection before paying up for an Outperform call.
Results vs. Consensus
| Metric | Actual | Consensus | Guidance | Beat/Miss | Magnitude |
|---|---|---|---|---|---|
| Revenue | $14.67B (+8% YoY) | ~$14.62–14.64B | $14.50–14.70B | Beat | +0.3–0.4%; high end of guide |
| Non-GAAP Gross Margin | 68.4% (+50 bps) | ~68.1% | 67.5–68.5% | Beat | High end of guide |
| Non-GAAP Op Margin | 34.3% (+180 bps) | ~33.5% | 33.5–34.5% | Beat | High end of guide |
| Non-GAAP EPS | $0.99 (+14% YoY) | $0.98 | $0.96–$0.98 | Beat | +1¢; above guide range |
| GAAP EPS | $0.71 (+31% YoY) | — | $0.67–$0.69 | Beat | Above guide range |
| Operating Cash Flow | $4.2B (+14%) | — | — | Solid | Driven by revenue/earnings growth |
| Product Orders (YoY) | +7% | — | — | Strong | Growth across all geos |
Quality of the Beat
- Revenue: The $50M revenue beat is small but the composition is high quality. Product revenue grew 10% while services revenue was flat, with the product strength concentrated in webscale AI infrastructure and the new networking refresh cycle. Splunk had a full 13-week contribution in the prior-year Q4, so reported growth rates are fully comparable — this is organic 8% growth, not an acquisition tailwind.
- Margins: Non-GAAP gross margin expansion was structural rather than one-off. CFO Patterson attributed the product gross margin to "productivity improvements" and noted tariffs were "slightly favorable compared to our estimate" — a small mix help, not the swing factor. Operating leverage of 180 bps on the operating margin line is the more important signal — EPS growth (+14%) materially outpaced revenue growth (+8%).
- EPS: Clean operationally. Non-GAAP tax rate was 18.1% — broadly in line with the guided ~18% for the quarter. Below-the-line items contributed nothing extraordinary; Q4 buyback of ~19M shares at an average $64.65 for $1.3B provided modest support.
- Order quality: The 7% product order growth was driven by service provider/cloud +49% YoY, with four of the top six webscale customers each growing orders in the triple digits, and two webscale customers placing total orders exceeding $1B for the year across networking, security, collaboration, and observability. Concentration risk is real but the customer set is the highest-CapEx cohort in tech.
Segment Performance
| Product Segment | Q4 FY25 Revenue (est.) | YoY Growth | Order Color | Assessment |
|---|---|---|---|---|
| Networking | ~$7.6B | +12% | Double-digit order growth, 4th consecutive quarter | Strong — AI + early refresh |
| Security | ~$1.95B | +9% | Mid-single-digit orders; ex-Fed double-digit | Mixed — new product traction, headline trails LT target |
| Observability | ~$0.26B | +4% | Led by Splunk + ThousandEyes | Deceleration vs. Q3's +24%; Splunk anniversary effect |
| Collaboration | ~$1.05B | +2% | Driven by devices | Stable — not a thesis driver either way |
| Services | $3.8B | Flat | Deceleration from +3% in Q3 | Soft — but management expects pickup as product growth ratables in |
| Geography | Q4 Revenue Growth | Q4 Order Growth | Notable |
|---|---|---|---|
| Americas | +9% | +5% | U.S. Federal weakness offset by enterprise + webscale |
| EMEA | +4% | +10% | Enterprise mid-teens, SP mid-teens; UK/Germany/Saudi strong |
| APJC | +7% | +7% | Broadly balanced |
| Customer Market | Q4 Order Growth YoY | Notable |
|---|---|---|
| Service Provider & Cloud | +49% | Webscale triple-digit 4th straight Q; telco/cable +20% |
| Enterprise | +5% | Several 8-figure-plus deals; early AI pipeline ramping |
| Public Sector | -6% | Tough Fed comp vs. Q4 FY24; ex-Fed rest-of-world +10% |
Networking — The Beat's Center of Gravity
Networking is doing the heavy lifting. Networking revenue grew 12% in Q4 with growth across most of the portfolio, led by double-digit growth in internet infrastructure and enterprise routing as well as solid growth in switching. The order side is even stronger: Networking product orders grew double-digits for the fourth consecutive quarter, driven by webscale infrastructure, switching, enterprise routing, industrial IoT, and servers, with industrial IoT orders posting their fifth consecutive quarter of double-digit growth.
The medium-term catalyst is the campus refresh. The new family of Cisco C9000 Smart Switches powered by SiliconOne marks the start of a major, multi-year campus refresh opportunity, but management was explicit that this is not yet a revenue driver:
"The Cat9k is in year 8 of this transition. So it's — and lots of customers will take a lot of time to evaluate that and look at those products before they begin to deploy them. So we think that's going to kick in next year as well… if you look at products that were pre-Cat9k that are still installed in our customer base, there's tens of billions of dollars of installed base that's there that we can go after." — Chuck Robbins, Chair & CEO
Assessment: Networking is the bedrock and is firing on every cylinder that matters — webscale orders, enterprise refresh signal, industrial IoT, and a multi-year campus replacement opportunity that has not yet hit revenue. The tens-of-billions installed base of pre-Cat9k gear is the most important number management cited for FY27+ optionality.
Security — Mid-Single-Digit Headline Hides a Better Underlying Story
Security is where the bull/bear debate is sharpest. The headline is unimpressive — mid single-digit growth in orders in Q4 — relative to a long-term Security+Observability target of 15–17%. But the composition has improved materially.
"If you take out U.S. Federal, which had a tough year as we all know, rest of world security order growth in Q4 was up double digits… If you look at those new products and new and refreshed, again, including our refreshed firewalls, we saw order growth during the quarter in excess of 20%." — Chuck Robbins, Chair & CEO
The new/refreshed cohort (Secure Access, XDR, Hypershield, AI Defense, refreshed firewalls) is now collectively added 750 new customers in the quarter and represents roughly two-thirds of the organic security portfolio. Hypershield in particular is showing a strategic bundling pattern — the vast majority of new Hypershield enterprise customers are bundling with the N9300 Smart Switch, validating Cisco's "security in the fabric of the network" thesis. Splunk synergies delivered a 14% YoY increase in new Splunk logos in Q4, with management noting over 300 new Splunk logos in each of the last two quarters from customers Splunk had not previously sold into.
Assessment: Security is the most credible re-acceleration story in the deck. Robbins committed to either being "at" or "on a path to" the 15–17% Security+Observability range by year-end FY26 — a high-bar verbal commitment that becomes the watch item for every subsequent quarter. We are sympathetic but not yet paying for it.
Services — The One Genuinely Soft Line
Services revenue was flat YoY in Q4, decelerating from +3% in Q3, +5.5% earlier, and +6.5% before that. Management framed it as a delayed-effect issue:
"If you go back a year ago and even a little bit longer ago, you saw a lot of professional services that we did, really helping our customers and our partners… that drove nice growth in our services business for a while. What we usually do see, as you know, services usually trails or sort of follows on to what's happening in product… as we saw networking growth of 12% this quarter, I would expect that you'll start to see services pick up." — Mark Patterson, CFO
Assessment: The explanation is reasonable but unverifiable in real time. Services is 25% of revenue, so a 4-quarter glide from +6.5% to 0% is a material consumer of headline growth. If services does not lift through FY26, the FY26 revenue midpoint will face risk regardless of product strength.
Key KPIs & Capital Returns
| KPI | Q4 FY25 | YoY | Notes |
|---|---|---|---|
| Total RPO | $43.5B | +6% | 50% recognizable within 12 months; product RPO +8%, services +5% |
| Short-term RPO | $21.7B | +4% | Decent but not accelerating |
| Total ARR | $31.1B | +5% | Product ARR +8% — the more important growth metric |
| Subscription Revenue | $7.9B | +3% | 54% of total Cisco revenue |
| Software Revenue | $5.6B | +5% | Software subscription also +5% |
| Deferred Revenue | $28.8B | +1% | Product +2%, services flat |
| Capital Returned (Q4) | $2.9B | — | $1.6B dividend + $1.3B buyback at avg $64.65 |
| FY25 Capital Returned | $12.4B | 94% of FCF | 14th consecutive dividend raise |
| Buyback Authorization Remaining | $14.2B | — | No termination date |
Key Topics & Management Commentary
Overall Management Tone: Management was measured-confident rather than chest-thumping — appropriate for a quarter that beat on margins but only modestly on revenue and gave an in-line outlook. The AI order book ($2B vs. $1B target) is the only place where the language genuinely escalated; on every other vector (Q1 deceleration, security run-rate, services flatness, sovereign timing) the framing was patient and qualifier-heavy. Patterson's first call as CFO was on-message and notably more anchored on "durable, profitable growth" than aspirational TAM language, signaling continuity over reset.
AI Infrastructure — The Headline Number That Earned Its Spot
The AI franchise has gone from narrative to financial. AI infrastructure orders from webscale customers exceeded $800M in Q4, bringing the FY25 total to over $2B — more than double the original $1B target. Revenue recognized in FY25 from those orders was approximately $1B, with the balance flowing into FY26 as a starting backlog.
"The AI infrastructure orders we received from webscale customers in fiscal 2025 were more than double our original target, indicating a massive opportunity ahead as we lead the required architectural shift and build the critical infrastructure needed for the AI era." — Chuck Robbins, Chair & CEO
The product mix of AI infrastructure orders was approximately two-thirds systems and one-third optics, and management indicated it expects the mix to remain stable. The framework is now articulated in three pillars: (1) AI training infrastructure for webscale customers, (2) AI inference and enterprise clouds, and (3) AI network connectivity for agentic-AI traffic. Enterprise AI is described as "early" — a few hundred million of orders to date with a pipeline in the hundreds of millions.
Assessment: The 2x-target moment is the most important data point in this print. It validates Cisco as a credible co-architect of webscale AI build-outs — not the GPU story, but a participating system provider on systems + optics. The harder question is durability beyond FY26 once the comp gets harder and hyperscale CapEx growth normalizes — management's response (Robbins citing aggregate "~50% up" hyperscale CapEx growth) is defensible but provides limited forward visibility.
Sovereign AI — Real but Not Yet Revenue
The Middle East sovereign opportunities (HUMAIN, G42, Stargate UAE) were explicitly called out as in planning, not in backlog:
"We have not taken any orders from them yet. We've been in the planning phases with them. They're obviously working through getting the licenses for the GPUs. And we sort of expect that's going to — I would lean towards looking for that even the order flow to be sort of middle year into the second half [of FY26], and then revenue would follow." — Chuck Robbins, Chair & CEO
Notably, management indicated they are working with AMD on some of these sovereign deals alongside the better-publicized NVIDIA partnership — a meaningful diversification of the AI silicon partner story.
Assessment: Sovereign AI is a real call option that is not yet in the model — order flow mid-FY26 with revenue following implies the bulk of any contribution is FY27. This is a "free option" rather than a near-term catalyst, but it is the cleanest way for the AI thesis to extend beyond the current webscale concentration.
CFO Transition — Continuity, Not Reset
This was Mark Patterson's first quarter as CFO, succeeding Scott Herren. Patterson, a 25-year Cisco veteran, framed his priorities in terms that were notably grounded:
"You can expect that I will be focused on durable profitable growth, obviously, financial discipline and transparency and really just returning value to the shareholders." — Mark Patterson, CFO
Capital allocation priorities articulated on the call were ordered as: (1) support growth of the business, (2) support the dividend, (3) offset dilution via buybacks, (4) opportunistic capital return beyond that. The language is meaningfully more conservative than peers — note the absence of "M&A" as a discrete priority despite the freshly-announced industry CyberArk acquisition (by a competitor) that an analyst attempted to bait Robbins into commenting on.
Assessment: Patterson's debut signals continuity. The combination of an insider CFO with a "durable profitable growth" framing reduces the likelihood of a large-deal M&A surprise in FY26 and increases the probability of consistent capital return discipline. For a Hold-rated name, this is a positive marginal data point.
Public Sector / U.S. Federal — The Acknowledged Drag
U.S. Federal is the one segment where the headline number cleanly tells the story. Public sector orders were down 6% YoY, lapping an exceptionally strong Q4 FY24. Critically:
"Our teams are forecasting a return to growth for federal during this fiscal year. Now it's not going to be as high of growth as we saw a decline. So we'll still be below FY '25 levels. But the good news is they're forecasting growth." — Chuck Robbins, Chair & CEO
Assessment: Federal "returning to growth but below FY25" is a meaningful hedge in the FY26 guide — it suggests management is not banking on Federal as a tailwind, which de-risks the FY26 numbers if Fed weakness persists. The ex-Federal rest-of-world public sector +10% order growth indicates the issue is U.S.-specific budget cycle, not a Cisco-specific competitive loss.
Tariffs — Small Impact, Conservative Posture
Tariffs were a minor net help in Q4 — "slightly favorable compared to our estimate." The FY26 guide assumes current tariff regimes hold for the full year: China 30% (with semi/electronics exemptions), Mexico 25%, Canada 35% (non-USMCA components), and country-specific reciprocal rates on other geographies. Patterson explicitly declined to quantify the dollar impact, framing the disclosure as the assumption set rather than the exposure.
Assessment: The refusal to size tariff impact in dollar terms is a watch item. It leaves the door open for both upside (if tariffs de-escalate) and downside (if regimes worsen) without a clear baseline. Cisco's global supply chain flexibility is a structural advantage relative to less-diversified networking peers, but the lack of disclosure makes the gross-margin sensitivity hard to model.
Guidance & Outlook
| Metric | Q1 FY26 Guide | FY26 Guide | Implied vs. FY25 Actual |
|---|---|---|---|
| Revenue | $14.65–14.85B | $59.0–60.0B | +5% YoY at midpoint |
| Non-GAAP Gross Margin | 67.5–68.5% | Not formally guided | Implied flat-to-down 50 bps from 68.7% FY25 |
| Non-GAAP Op Margin | 33–34% | Not formally guided | Modest deceleration vs. Q4 34.3% high water mark |
| Non-GAAP EPS | $0.97–0.99 | $4.00–4.06 | +5.8% YoY at midpoint |
| GAAP EPS | $0.63–0.68 | $2.79–2.91 | +9% YoY at midpoint |
| Non-GAAP Tax Rate | ~19% | ~19% | Up from 18% in FY25 |
The shape of the FY26 guide is the analytically interesting piece. Q1 FY26 revenue midpoint of $14.75B implies ~6.5–7% YoY growth, decelerating to a full-year midpoint of ~5%. Management was asked to explain the apparent deceleration and attributed it primarily to comps:
"The dynamic that you're talking about is strictly connected to just year-over-year comps later in the year… You have to think back as well in prior year, other than Q4, that's the first quarter that really was apples-to-apples in terms of having Splunk in the prior year. So some of the growth rates before Q4 were obviously higher than they would be otherwise." — Chuck Robbins, Chair & CEO & Mark Patterson, CFO
Implied Q-over-Q ramp: Q1 midpoint of $14.75B vs. Q4 FY25 $14.67B = ~0.5% sequential. From Q2–Q4, management needs ~$14.9B/quarter average to hit the FY26 midpoint — consistent with modest sequential growth and no required acceleration.
Street at: FY26 revenue Street consensus was roughly $59.5B prior to the print — the guidance midpoint of $59.5B is therefore essentially in line. Non-GAAP EPS guidance midpoint of $4.03 is within ~$0.01–0.03 of Street.
Guidance style: Cisco's historical pattern is to guide conservatively and beat by 1–2¢ on EPS / 30–50 bps on revenue — the FY25 trajectory was consistent with that. The FY26 guide screens neither aggressive nor unusually cautious; it does, however, embed the assumption that tariffs hold in their current configuration for a full year, which is a non-trivial macro bet.
Analyst Q&A Highlights
FY26 Deceleration Math
- Aaron Rakers, Wells Fargo: Pressed on the apparent deceleration from ~6.5–7% Q1 growth to lower run-rate in Q2–Q4. Management attributed it cleanly to the Splunk comp (Q4 FY25 was the first fully comparable quarter) rather than any change in demand.
Assessment: Reasonable explanation, mechanically defensible. But it leaves no comp-based excuse for FY27 — Cisco needs to demonstrate organic AI/refresh acceleration by then.
Security Trajectory vs. Long-Term Framework
- Meta Marshall, Morgan Stanley: Asked about security growth outlook now that Splunk has anniversaried. Management offered three confidence-building data points: ex-Fed rest-of-world double-digit order growth, new/refreshed cohort +20% order growth (now ~⅔ of organic security), and 480+ new SSE customers in Q4.
- Ben Reitzes, Melius: Most pointed pushback on the call — challenged whether the long-term 15–17% Security+Observability target is still appropriate given the run-rate. Robbins committed to either being at the range or on a clear path by year-end FY26.
Assessment: This is the highest-stakes verbal commitment Robbins made on the call. The new/refreshed cohort math is supportive (if 67% grows 20%+ and the legacy 33% drag fades, blended growth converges toward the target). But it leaves zero margin for further Federal weakness or new-product execution slippage.
AI Revenue Contribution in FY26
- Michael Ng, Goldman Sachs & Amit Daryanani, Evercore: Both pressed for an AI revenue number for FY26. Patterson confirmed ~$1B of AI infrastructure revenue recognized in FY25 with the balance carrying as backlog. Robbins refused to give a discrete FY26 AI number, suggesting analysts "extrapolate" given the backlog and new order flow.
- Notable non-answer: The refusal to size FY26 AI revenue specifically is conspicuous given how much of the bull thesis depends on it. Reasonable inference: $2B+ in FY26 (carrying the $1B+ backlog plus new orders) is the implicit floor, but explicit guidance would have been a credibility-boosting disclosure.
Pull-Forward Risk (Tariffs & Federal)
- Simon Leopold, Raymond James: Asked whether tariff fears or federal budget anxiety may have pulled orders forward, inflating the recent quarter. Management offered an unusually granular response: linearity within the quarter was normal, software activation timing-from-shipment had not lengthened, ship-date requests had not slid, and channel partner conversations showed no pull-forward signal.
Assessment: Best-in-class disclosure on this question — four discrete data points, not just a verbal denial. This materially de-risks the most plausible bear case for the order-book strength.
AI CapEx Sustainability
- Tal Liani, Bank of America: Asked whether webscale AI CapEx growth is sustainable or whether the comps get impossible next year. Robbins pointed to aggregate hyperscaler CapEx +50% YoY as the demand signal Cisco is operating against, declining to extrapolate beyond that.
Assessment: This is the unhedgeable macro risk for the AI story. Management's response is honest — they are riding the customers' signal — but it leaves the bear case (hyperscale digestion in CY26+) live.
Capital Allocation & Identity / Security M&A
- James Fish, Piper Sandler: Asked about capital allocation priorities and Cisco's view of the identity/privileged-access space in light of the CyberArk transaction announced by a competitor. Patterson laid out the priority stack (growth → dividend → dilution → opportunistic). Robbins pivoted away from a direct M&A response to instead discussing Cisco's organic identity strategy (Duo, ICE, Oort) and the importance of identity in agentic-AI architectures.
- Notable dodge: Robbins effectively declined to engage with whether Cisco would be a competing bidder for identity-security assets at scale. This is consistent with the capital-allocation discipline framing but leaves a strategic question open.
Silicon One ASIC Mix & Scale-Up Ethernet
- Karl Ackerman, BNP Paribas: Asked if Silicon One can represent half of Cisco's switch ASICs within three years. Robbins indicated "not too far off" — a meaningful directional confirmation.
- Sebastien Naji, William Blair: Asked about scale-up Ethernet within AI clusters as competing-silicon roadmaps evolve. Robbins acknowledged unannounced roadmap items and indicated Cisco could "definitely play" if scale-up shifts to Ethernet.
Assessment: These are the most forward-leaning architecture comments of the call. They suggest Cisco sees the systems-vs-merchant-silicon debate as genuinely in flux and is positioning for both outcomes.
EMEA Strength & Sovereign Pipeline
- Adrienne Colby (for Atif Malik), Citi: Asked about EMEA acceleration and sovereign-AI timing. Robbins highlighted UK, Germany, and Saudi Arabia as the sources of EMEA strength, with sovereignty/on-prem themes (on-prem Splunk, on-prem WebEx) as a structural EMEA tailwind. Confirmed sovereign AI orders are expected mid-to-second-half of FY26.
Assessment: EMEA strength is structurally sticky — sovereignty preferences favor Cisco's on-prem-capable architecture over pure-cloud rivals.
What They're NOT Saying
- No explicit FY26 AI revenue number: Despite repeated analyst attempts, management refused to commit to a discrete FY26 AI infrastructure revenue figure, deflecting with the FY25 $1B baseline and an "extrapolate" prompt. For the single most important growth driver in the deck, this is conspicuously thin disclosure.
- No quantified tariff impact: Patterson explicitly chose to share the assumption set (tariff rates by country) rather than dollar impact. This obscures the FY26 gross-margin sensitivity to trade-policy changes — both upside and downside.
- No Splunk standalone revenue or ARR disclosure: Now that Splunk has anniversaried, Cisco is no longer required to disclose its discrete contribution. The 14% new-logo growth and 300+ new logos per quarter are positive but unfalsifiable without a revenue figure.
- No discrete FY26 services growth target: Patterson said services will "improve" as product growth normalizes, but offered no specific growth rate. With services representing 25% of revenue and decelerating to flat, this matters for the FY26 revenue bridge.
- Limited engagement with the CyberArk / identity M&A question: Robbins pivoted to Cisco's organic identity capabilities rather than engaging directly with whether Cisco would compete for similar assets. A pure organic strategy is defensible, but the lack of explicit framing leaves the strategic posture ambiguous.
- No FY26 segment-level guide: Consistent with Cisco's historical practice, but worth noting given how much of the bull/bear debate hinges on the Networking vs. Security mix specifically.
- No commentary on a Q1 FY26 EPS midpoint that is flat-to-down sequentially vs. Q4 FY25's $0.99: Q1 midpoint is $0.98 — effectively flat — despite revenue at the high end of the Q1 range exceeding Q4. The implied operating margin compression to 33–34% from Q4's 34.3% was not addressed.
Market Reaction
- Pre-print setup: Stock closed at approximately $71.38, down 1.42% on the day, trading near a 52-week high of $72.55. CSCO had run hard into the print on AI optimism following Q3's $600M of AI infrastructure orders (already exceeding the original $1B FY25 target a quarter early). The Q3 print itself had been received negatively in May despite a beat, suggesting a market that needs more than incremental beats to reward the name.
- After-hours move: Initial reaction was negative — shares slipped roughly 1–2% on the headline, with reports placing the after-hours price near $69.55 (down about 1% from the regular-session close). The thinness of the revenue beat (~$50M) combined with an FY26 revenue midpoint of $59.5B (essentially in line with Street ~$59.55B) left no clear upside catalyst against a stock priced near 52-week highs.
The after-hours move is the rational response to the print, not a punishment. EPS beat by 1¢, revenue beat by ~$50M, and the FY26 outlook was in line — none of those data points alone justify multiple expansion at a 52-week high. The market is telling us that "AI orders doubled the target" was already in the stock, and that the durable acceleration story now has to come from the next vector (campus refresh, sovereign AI, security re-acceleration, or services recovery). None of those vectors materializes in the next 90 days, so a sideways-to-down reaction is consistent with the information set.
Street Perspective
Debate: Is AI Revenue Durable Beyond FY26?
Bull view: Cisco has now demonstrated it can win at the systems + optics layer of webscale AI build-outs, with order growth from cloud and AI customers in triple digits for four consecutive quarters and two customers placing $1B+ in total orders for FY25. The sovereign AI optionality (HUMAIN, G42, Stargate UAE) plus an emerging enterprise AI pipeline extend the runway materially.
Bear view: Webscale CapEx is at historic levels, +50% YoY in aggregate. The risk is not whether AI is real but whether FY27 lapping the current order cohort is even possible without a meaningful broadening of the customer base. Concentration in a handful of hyperscalers exposes Cisco to the same digestion risk that has historically defined networking cycles.
Our take: The bull case is more credible than it was a year ago — the $2B vs. $1B-target outperformance was a structural achievement, not a quarter-end push. But durability beyond FY26 requires either sovereign-AI revenue to materialize (back-half FY26 at the earliest) or enterprise AI to scale (still "pilots" today). The next 12 months are well-supported by backlog; the 12 months after that depend on vectors not yet in numbers.
Debate: Is the Security Long-Term Target Still Realistic?
Bull view: The new/refreshed product cohort is growing 20%+ on orders and now represents two-thirds of organic security. Splunk synergies are delivering 300+ new logos per quarter. Ex-Federal, security orders are already growing double digits. Robbins committed to being at or on a path to 15–17% by year-end FY26 — a verbal commitment that creates accountability.
Bear view: The headline security revenue growth of +9% is a long way from 15–17%. Hypershield bundling with N9300 smart switches is intriguing strategically but unproven at scale. Federal weakness has persisted longer than expected and there is no guarantee of recovery in FY26 magnitudes.
Our take: The composition math is plausible — if the 67% new/refreshed cohort grows 20%+ and the 33% legacy fades, blended growth converges toward the target by exit-FY26. But execution from here has zero margin for error, and the long-term target was constructed in a different macro and federal-budget environment.
Debate: Should the Stock Re-Rate Higher on AI?
Bull view: Cisco is the only major networking incumbent with credible systems + silicon + security + software stack for AI-era infrastructure. Margins are at record levels, capital return is 94% of FCF, and the FY26 guide includes meaningful tariff conservatism. The setup mirrors prior platform-shift inflections where the incumbent re-rates as the new TAM becomes visible.
Bear view: The stock entered the print at 52-week highs after a strong run on AI optimism. The multiple already reflects elevated expectations and the FY26 guide in line with Street offers no fundamental basis for further expansion. Services flatness and security run-rate below LT target are real overhangs.
Our take: The re-rating debate is the right debate, but premature. The print confirmed the AI thesis without changing its slope. With the stock near highs and no obvious upside surprise pending in the next 90 days, the risk/reward sits balanced — we'd prefer to re-engage either on a pullback or on tangible Q1 FY26 evidence of security re-acceleration.
Model Update Needed
| Item | Current Assumption | Suggested Change | Reason |
|---|---|---|---|
| FY26 Revenue | +5–6% YoY | Hold at +5% (midpoint of guide) | Guide is in line; comp dynamics support trajectory |
| FY26 Non-GAAP GM | 68.5–68.7% | Trim to 68.0–68.3% | Q1 guide of 67.5–68.5% implies modest compression from FY25's 68.7% |
| FY26 Non-GAAP Op Margin | 34.0%+ | Hold at 33.5–34.0% | Q1 guide 33–34% suggests reset from Q4's 34.3% peak |
| FY26 Non-GAAP EPS | $4.05 | Hold at $4.03 (midpoint) | Aligned with company guide; some upside from buyback cadence |
| FY26 AI Infra Revenue | Implicit $1.5B | Raise to ~$2.0–2.5B | $1B+ backlog from FY25 plus new order momentum supports higher floor |
| FY26 Tax Rate | ~18% | Lift to 19% | Per company guide |
| Capital Return | $11B | Raise to ~$12B | FY25 was $12.4B (94% of FCF); FY26 likely similar |
Valuation impact: Our model implications point to modest EPS revisions (+1–2%) on the AI infrastructure mix shift, offset by the gross margin reset. Net change to fair value: limited — the print does not move our target materially in either direction. The case for a higher multiple is the AI durability story; we want one more quarter of evidence before paying for it.
Thesis Scorecard Post-Earnings
| Thesis Point | Status | Notes |
|---|---|---|
| Bull #1: AI infrastructure is a multi-year, multi-billion-dollar order opportunity | Confirmed | $2B vs. $1B target; backlog supports FY26 |
| Bull #2: Margins can expand as software/subscription mix grows | Confirmed | Record FY25 non-GAAP OpM; Q4 OpM 34.3%, +180 bps |
| Bull #3: Campus refresh cycle accelerates revenue in FY26+ | Neutral | Management explicit: "kick in next year" — FY27 is the inflection |
| Bull #4: Splunk synergies drive security re-acceleration | Neutral → Confirmed | 300+ new logos per quarter, 20%+ growth in new/refreshed; headline still soft |
| Bear #1: Webscale concentration creates digestion risk | Neutral | 2 customers $1B+ in FY25 = real concentration; no signal of digestion yet |
| Bear #2: Security can't reach 15–17% LT target | Challenged → Neutral | Mgmt verbally committed to path by exit-FY26; execution risk high but not impossible |
| Bear #3: Services flatness signals underlying weakness | Confirmed (mild) | 0% YoY in Q4 is real; mgmt explanation reasonable but unverifiable |
| Bear #4: Stock is priced for perfection at 52-week highs | Confirmed | Modest beat + in-line guide = no upside catalyst; AH reaction muted-negative |
Overall: Thesis is unchanged — moderately strengthened on the AI confirmation, moderately weakened on valuation re-rating optionality given the muted reaction. Net: a wash that supports a balanced Hold posture.
Action: Initiate at Hold. Re-engage on either (a) a pullback to the high-$60s creating a more attractive entry, or (b) tangible Q1 FY26 evidence of security accelerating into double-digit headline growth ex-Federal recovery. Maintain monitoring on sovereign AI order conversion (back-half FY26) and campus refresh signals (FY27 inflection).