PodcastSteve Eisman PlaybookSam Badri

What the AI Boom Looks Like From Inside Cisco with Sam Badri | The Real Eisman Playbook Ep 77

Badri describes Cisco's AI-related growth alongside a longer-running silicon program, a refresh cycle for older enterprise equipment, and rising hyperscaler demand. He says power density is increasing faster than data-center size, while AI adoption and token use remain early and may expand across industries. The host also raises risks around customer concentration, dependence on frontier models, and a possible AI price war.

Preview · 18 more points in the app

Read the Full Summary

Key Takeaways

Cisco’s growth acceleration reflects both hyperscaler demand and replacement orders for enterprise gear: it reported 18% quarterly revenue growth and guided to 15% growth at the midpoint for FY27, while Badri said hyperscaler revenue was expected to nearly double but remain below 20% of the company’s total. Badri linked larger Silicon One-based system orders, which began six to nine months before the interview, to multiple wins in hyperscaler networks; Cisco’s own silicon is used in its data-center switches, and the broader silicon program has been running for about 10 years.

Cisco and Arista have different exposure to growth trends: Badri says Arista is more heavily weighted toward hyperscalers and faster-moving markets, while Cisco’s security, collaboration, services, and other businesses give it more exposure to slower trends. That mix sits alongside Cisco’s reported acceleration, which the host attributed in large part—but not entirely—to hyperscaler AI-related business.

AI activity and demand are concentrated around OpenAI and Anthropic, according to Badri, who compares their position to Google Search if it were private; the host separately cites Nvidia’s 10-Q, where five accounts represented 70% of accounts receivable, and raises concern about reliance on the two frontier-model companies. Badri says some workflows can use less costly open-weight models, while demanding work such as silicon engineering may warrant frontier capabilities; the host also flags a possible price war among Anthropic, OpenAI, and Chinese models as a risk.

GEO-POLITICS

Texas Grid Audit — Badri says Texas is auditing its grid-allocation queue after data-center construction drove a buildup in power requests. The backlog is many times larger than the capacity under construction and may include duplicate requests, so the audit aims to determine how much of the queue is genuine.

“So whenever there are when you're submitting a request to the grid for power that goes through a process almost like not not an auction but just a queue, a…”—

TECHNOLOGY

Silicon One History — Cisco announced its G300 Silicon One chip in February 2026. Its broader silicon program had been running for about 10 years.

“This specific chip is very new. This was announced in February of 2026. but the actual silicon program is not new at Cisco. It's been around for about 10 years.”—

Cisco Growth — Cisco reported 18% revenue growth in its most recent quarter and guided to 15% growth at the midpoint for FY27, following low-double-digit growth over the preceding 12 months. Badri said the hyperscaler business was expected to nearly double in FY27 while remaining below 20% of company revenue. The host attributed much, but not all, of the recent acceleration to Cisco's hyperscaler AI-related business.

“So our fiscal year 27 guidance is actually 15% revenue growth at the midpoint…”—

Data-Center Connectivity — Cisco sells hyperscalers switches that carry traffic through data centers, along with coherent pluggables that connect fiber or Ethernet to switches. Badri says Cisco's own silicon goes into its switches.

“So we sell them it comes down to really two main products…”—

Intro & Sam’s Background

Sam Badri traces his path from consulting and Wall Street equity research covering data centers, networking, and software to becoming Cisco’s head of investor relations.

What Cisco Does

Badri explains Cisco’s core role in enterprise connectivity through switches, routers, and Wi-Fi, and distinguishes its network-processing silicon from GPUs and complete computing systems.

Cisco Revenue Growth & What’s Changed

Cisco moved from a normalized mid-single-digit growth profile and an inventory digestion period to stronger growth, driven by hyperscaler demand and a long-delayed enterprise campus refresh.

How AI Factors In

Cisco’s AI inflection became clearer as demand grew for Silicon One-based systems; Badri says large hyperscaler orders and new use cases made Cisco’s AI relevance more tangible.

Cisco vs Arista

Cisco’s in-house silicon differentiates its hyperscaler switches from Arista’s typically merchant-silicon approach, while Arista grows faster because hyperscalers represent a much larger share of its business.

The Negativity Surrounding the AI Story

Badri connects frontier-model popularity to accelerating token demand, while the discussion weighs customer concentration, open-weight alternatives, and whether the AI boom depends too heavily on a few model companies.

Data Center Pushback

Badri says Cisco is not a front-line data-center builder, but describes grid-connection queues, local opposition, and sharply rising power density as constraints on expanding facilities.

What Sam Brought to Cisco

Badri says his sell-side experience helped Cisco recalibrate investor communications, improve disclosure choices, and bring shareholder questions directly into internal discussions.

Last Word on Cisco & the Sector

Badri argues Cisco has improved productivity and set records for revenue, margins, and earnings per employee; he sees AI adoption and token use as early, with potential to spread across industries.

Outro

After the interview, Eisman highlights Cisco’s accelerated growth and AI-related hyperscaler business, views token expansion as key to durability, and flags a possible price war among frontier and Chinese models as a risk.

Related Episodes

Important Disclaimers

Not Investment Advice: For education only — not financial, tax or legal advice.

AI-Generated: Made by AI from source transcripts and may contain errors.

No Guarantees: Past performance does not predict future results; investing carries risk.

Our Commentary · Request Removal