On Thursday, the S&P 500 crossed 7,500 for the first time in its 68-year history. The Nasdaq Composite tagged a fresh record close at 26,635. And the catalyst was exactly what you’d expect — AI spending that has gone from speculative to profitable in under 18 months. Cisco alone added $30 billion in market cap in a single session because hyperscalers can’t stop buying its networking gear. This wasn’t a broad-based rally. It was AI pulling the entire market uphill.
Cisco Told You Everything You Need to Know in One Earnings Call
Cisco reported Q3 fiscal 2026 revenue of $15.84 billion, beating Wall Street estimates by nearly 2%. Earnings per share came in at $1.06, up 10% year over year. The stock surged over 13% to an intraday record of $119.36. But the number that moved the market wasn’t in the income statement — it was in the order book.
AI-related hyperscaler orders are now tracking at roughly $9 billion for fiscal year 2026. CEO Chuck Robbins said hyperscaler orders rose by a triple-digit percentage. Networking revenue jumped 25% to $8.82 billion, blowing past the $8.47 billion consensus. Cisco raised its full-year revenue guidance to $62.8–$63.0 billion and bumped EPS targets to $4.27–$4.29 — both well above what analysts were modeling.
Here’s the translation: the companies building the world’s largest AI data centers — Microsoft, Google, Amazon, Meta, Oracle — are spending so aggressively on networking infrastructure that Cisco can’t fill orders fast enough. This isn’t projected demand. This is purchase orders with delivery dates.
The Quiet Part: AI Spending Actually Generates Revenue Now
For the past two years, the AI trade had a credibility problem. Hundreds of billions were flowing into GPU clusters, model training, and data center construction, but the returns were theoretical. Investors had to take it on faith that the spending would eventually produce matching revenue. That argument is officially over.
Alphabet, Meta, Microsoft, and Amazon collectively committed $630 billion to AI capex in the current cycle. Their latest earnings showed it working. Microsoft’s cloud revenue accelerated. Google’s AI-powered ad products lifted margins. Meta’s recommendation engine drove higher engagement per session. Amazon’s AWS AI services grew faster than the rest of the cloud business combined.
And now, the infrastructure suppliers — Cisco, Arista, Broadcom, the entire networking and server stack — are posting record quarters because the spending shows no sign of decelerating. The S&P 500 crossing 7,500 isn’t speculation. It’s the market pricing in confirmed cash flows.
Why 7,500 Matters More Than Any Other Round Number
The S&P 500 took 18 months to go from 5,000 to 6,000, then six months to reach 7,000, and just four months to hit 7,500. The acceleration tells you something important about market structure: AI is compressing the entire cycle. Companies that usually take three to five years to grow into their capex are doing it in three to five quarters.
The concentration risk hasn’t disappeared — the Magnificent Seven still account for a disproportionate share of index gains. But the second derivative is changing. Cisco isn’t a Magnificent Seven stock. Neither is Arista. Neither is Vertiv, Eaton, or the dozens of infrastructure companies posting record results because they happen to sell things that AI data centers consume. The rally is broadening precisely because AI spending has become operational expenditure, not a venture bet.
The Contradiction Nobody Wants to Talk About
There’s a dark thread running through all of this. Cisco announced 4,000 job cuts alongside those record earnings. The company is restructuring to pivot resources toward AI products and away from legacy networking lines. It’s the same pattern we’ve seen at Microsoft, Meta, Google, and virtually every large tech company this cycle: AI creates revenue and destroys headcount simultaneously.
The S&P 500 at 7,500 represents an economy where corporate profits have never been higher and workforce efficiency has never been more aggressive. Those two facts aren’t coincidental. They’re the same fact, measured differently. The companies driving this rally are posting record margins specifically because they’re replacing human labor with AI at every layer of the stack — from customer service to code generation to ad targeting.
Wall Street doesn’t have a word for this yet. It’s not a bubble, because the earnings are real. It’s not organic growth, because it’s funded by once-in-a-generation capital deployment. It’s something new — a market that keeps setting records while the labor market underneath it quietly restructures around fewer people doing more with machines.
What Comes Next
The immediate catalyst calendar is stacked. Nvidia reports earnings later this month, and every analyst on the street is modeling another blowout quarter. If Jensen Huang confirms that data center demand is accelerating — again — the market will price in 8,000 before summer. WWDC is in June, with Apple expected to unveil its most significant AI push yet. Google I/O just showed an aggressive agent and hardware roadmap.
But the bigger question is durability. AI capex cycles historically overshoot before correcting. The dot-com boom built fiber capacity that wasn’t fully utilized for 15 years. The current AI infrastructure build is the largest capital deployment cycle in technology history, and the returns are arriving faster than any previous cycle — but “faster than before” doesn’t mean “fast enough to justify the next $630 billion.”
The verdict: The S&P 500 at 7,500 is real. The earnings behind it are real. The AI revenue is real. But the speed of this rally is pricing in a future where every dollar of AI spending produces a dollar of return. History says that’s never how infrastructure cycles end. The market is right about the direction. The question is whether it’s right about the magnitude — and that won’t be answered until the next company posts a miss, not a beat.