A 26-YEAR WAIT AND WHY PATIENCE STILL MATTERS

(CSCO), (ADBE), (MSFT), (CRM), (ANET), (CIEN)

Cisco (CSCO) finally broke through its March 2000 dot-com bubble peak of $82 last quarter, a mere 26 years later. I’ve owned the stock since 2006, which means I’ve spent two decades watching it trade sideways while the rest of tech tripled. 

The AI infrastructure narrative was supposed to change all that. Instead, Cisco reported solid earnings, beat guidance, and promptly dropped 11% the next day because Q3 guidance showed flat sequential growth. 

Some things never change.

The mechanics of why the stock sold off are straightforward enough. Cisco’s sitting on $2.1 billion in hyperscaler AI infrastructure orders from Q2 alone, up from $1.3 billion in Q1 and equal to what they booked in all of fiscal 2025. 

Full-year guidance implies these orders will hit $5 billion, though the first-half run rate suggests they’ll exceed that. Service provider and cloud orders grew 65% in Q2. 

Enterprise orders up 8%, public sector up 12%, telco and cable up 20%. The demand is real and accelerating.

But order growth doesn’t show up in revenue immediately, especially when you’re dealing with enterprise infrastructure sales that involve multi-quarter deployment cycles. 

Cisco guided Q3 revenue to $15.5 billion, essentially flat from Q2’s $15.35 billion. The full-year guidance of $61.45 billion implies Q4 revenue of $15.72 billion, which represents 7.2% growth over Q4 2025. 

The growth is coming, just not fast enough for a market that apparently expects networking equipment to ship at the speed of software downloads.

The memory supply constraint narrative got more attention than it deserved. Memory prices spiked because hyperscalers are hoarding DRAM for data centers, and Qualcomm warned about phone manufacturers cutting production. 

But Cisco’s not selling smartphones. The same hyperscalers driving memory prices higher are ordering switches and routers for those data centers. They’re not canceling networking equipment because memory got expensive. 

Cisco’s size gives it preferential supplier treatment, and its lower UCS server exposure means less memory dependency per revenue dollar.

The Splunk situation is getting misread entirely. Security revenue down 4%, Observability flat, Services down 1%. 

On the surface, the $28 billion acquisition looks questionable. Except Splunk’s going through the exact transition every SaaS company navigated in the early 2010s: converting perpetual licenses to cloud subscriptions. 

You trade big one-time revenue for smaller recurring payments, and cumulative cash flows take three to five years to break even. 

Adobe (ADBE), Microsoft (MSFT), Salesforce (CRM) – they all went through this, got hammered during transition, and came out worth multiples of pre-transition valuations. 

Splunk’s providing network monitoring and security for enterprises upgrading infrastructure for AI workloads. The transition pain is temporary.

The valuation argument against Cisco is legitimate and worth taking seriously. At $75, the stock trades at 18.2 times fiscal 2026 earnings, with consensus expecting 8% annual EPS growth through 2028. That’s a PEG ratio of 2.28. 

Arista (ANET) trades at a PEG of 1.82 with 22% growth, Ciena (CIEN) at 2.04 with 27% growth. On profitability metrics, Cisco’s return on equity of 24% is respectable but trails Arista’s 32%. 

The only place Cisco wins is the dividend – 2.2% yield with annual increases since 2020, versus zero for the competition. 

They’re also returning capital through buybacks at 2.3% of market cap annually, though they’re currently spending more on dividends and buybacks combined than free cash flow generates, which suggests buybacks will need to come down.

For Cisco to be compelling at current prices, one of two things needs to happen. 

Either fiscal 2028 EPS estimates need to move up by about ten cents to reflect sustained 9% growth instead of 8%, which would drop the PEG to 2.0 and make it competitive with peers. 

Or the stock needs to fall to around $65, which accomplishes the same valuation improvement through price rather than estimates. 

The Q4 results in August, along with fiscal 2027 guidance, will determine which path we’re on.

The AI infrastructure upgrade cycle is real. Older Catalyst 4000 and 6000 series switches are hitting end-of-support, forcing upgrades to Catalyst 9000. 

WiFi 7 deployments are accelerating to handle increased data flows. These next-generation products are ramping faster than previous product cycles, and the order book supports continued growth. 

The question isn’t whether Cisco benefits from AI infrastructure spending – it clearly does. 

The question is whether that benefit is worth paying 18 times earnings for 8% growth when faster-growing alternatives trade at lower PEG ratios.

I’m holding my position because I’ve already got a 15% CAGR over two decades, and the dividend keeps compounding. But I’m not adding at $75. 

If the stock falls back to $65 or fiscal 2027 guidance in August shows acceleration toward 9-10% earnings growth, that changes. 

For now, it’s a hold, which, after 26 years of waiting to break even with the dot-com peak, feels about right for Cisco’s speed.