Cisco Raises Hyperscaler AI Revenue Target to $9 Billion

Cisco raised its AI infrastructure revenue target from hyperscalers to $9 billion for fiscal 2027, up $1.5 billion from the $7.5 billion projection it issued in August and more than double the roughly $4 billion it delivered in fiscal 2026.

The networking company announced the higher target Thursday alongside first-quarter results for the period ended Oct. 24. The revision lands three months after the original number, and it arrives with a second, arguably larger signal: AI networking demand is spreading beyond the cloud giants.

What drove the higher target?

Product orders rose 21% in the quarter, Chief Executive Chuck Robbins said in prepared remarks for the earnings call. Orders excluding hyperscalers rose 14%. Enterprise orders grew 15%, on top of 4% growth in the same quarter a year earlier.

The pattern matters for the broader AI infrastructure market. For two years, spending on networking gear for large AI clusters came overwhelmingly from a handful of hyperscalers — Amazon, Google, Microsoft and Meta, plus a set of newer neocloud operators. Cisco's numbers now show enterprises placing meaningful orders of their own.

Robbins framed the cycle in expansive terms. "We are in the very early stages of a multi-year networking supercycle," he said in the company's results release. On the call, he said AI infrastructure orders from neocloud, sovereign and enterprise customers accelerated during the quarter, and he described the AI-driven networking cycle as still early, "with enterprise adoption just beginning."

Specific product lines back the claim:

• Orders for Nexus data center switches tagged for AI deployments rose 47%.
• Campus networking orders grew 14%.
• Cisco won three new AI design wins in the quarter — two for systems built on its Silicon One P200 chip and one for 1.6-terabit optics.

The optics win touches a segment of AI networking that has drawn heavy investment this month. Marvell agreed to acquire Lumentum in a $115 billion all-stock deal earlier in November, betting on the same high-speed optical interconnects that AI clusters require.

Is the $9 billion target comparable to prior numbers?

Partly, and the distinction matters. The target covers revenue, not orders. Cisco booked $9.3 billion in hyperscaler AI orders during fiscal 2026 but converted only about $4 billion of that into revenue in the same year. The $9 billion figure for fiscal 2027 measures recognized revenue from that business, not the order book behind it.

The gap between orders and revenue reflects how large AI infrastructure deals move through Cisco's books. Hardware is recognized as it ships, and hyperscaler deployments stretch over multiple quarters. A $9.3 billion order backlog converted into $4 billion of fiscal 2026 revenue means a substantial carryover supports the new fiscal 2027 target.

Cisco's shares fell in extended trading after the August report despite beating expectations, CNBC reported at the time. The company will look for a different reaction to this quarter's set of numbers.

How did the quarter perform overall?

Revenue and adjusted earnings grew at their fastest rates in more than 15 years, Cisco said in its results release.

• Revenue rose 24% to $18.5 billion. Analysts polled by LSEG expected $18.1 billion.
• Adjusted earnings rose 40% to $1.40 a share, against an LSEG consensus of $1.33, CNBC reported.
• Networking product revenue jumped 43%.
• Security revenue grew 7%.
• Services revenue grew 1%.

"Proud of our Cisco team for an incredible start to FY27. We delivered record quarterly revenue of $18.5B, up 24% y/y, and non-GAAP earnings per share of $1.40, up 40% y/y — our strongest growth rates in over 15 years," Robbins wrote on X on Nov. 12.

The networking segment's 43% growth dwarfs the rest of the portfolio. That concentration cuts both ways: it validates Cisco's position in AI buildouts, and it raises the stakes if the AI networking cycle that Robbins describes slows.

What happened to margins?

The cost of that growth shows up directly in profitability. Adjusted gross margin fell 250 basis points from a year earlier to 65.6%. Adjusted product gross margin dropped 350 basis points to 63.7%.

Chief Financial Officer Mark Patterson attributed the product margin decline mainly to a heavier mix of hardware and "elevated memory costs," partly offset by productivity improvements and price increases. A global memory-chip shortage has been raising component costs for hardware makers across the industry, SiliconANGLE noted.

The margin pressure sets up the central tension in Cisco's AI story: hyperscaler and enterprise AI business carries lower gross margins than the software and services lines that Cisco has spent years acquiring toward. Patterson said hyperscale AI should make up 12% of Cisco's total revenue this fiscal year, double its 6% share in fiscal 2026. As that share doubles again, blended margins face continued pressure unless pricing and productivity keep pace.

How did guidance change?

Cisco raised its full fiscal 2027 outlook across both lines:

• Revenue guidance moved to $74.8 billion to $75.8 billion, up from $72.2 billion to $73.4 billion guided in August.
• Adjusted earnings guidance rose to $5.47 to $5.53 a share, from $5.05 to $5.11.

On operations, headcount has fallen by a low single-digit percentage, and Cisco expects revenue per employee to rise about 19%. The company returned $3.7 billion to shareholders in the quarter through dividends and buybacks.

What comes next?

The next read on enterprise demand arrives Feb. 10, when Cisco reports its second fiscal quarter. The company has guided that quarter to $18 billion to $18.2 billion in revenue, above the $17.88 billion analysts expected in LSEG data cited by CNBC.

Security could draw attention sooner. Robbins said Splunk, which Cisco acquired in 2024, is exploring with an unnamed hyperscaler how AI agents can help security teams investigate threats. He said Cisco has more to share later this month.

For the AI infrastructure market, Cisco's raised target and its 14% order growth outside hyperscalers offer the clearest evidence yet from a major supplier that enterprise AI networking spend has started to move. Whether the "multi-year networking supercycle" holds depends on the margin math Patterson now has to manage.