Broadcom Q1 Earnings AI Chips Now Account for Nearly 45% of Total Revenue.

 

Broadcom Shashes Records: AI Revenue Surges 106% as CEO Projects $100 Billion Milestone

Broadcom Inc. has reported stellar financial results for its first fiscal quarter of 2026, ended February 1. The technology giant posted consolidated revenue of $19.31 billion, a 29% increase year-over-year. Net income (GAAP) climbed to $7.35 billion, representing a robust 34% growth compared to the same period last year.

The AI Engine Driving Growth

CEO Hock Tan attributed the record-breaking performance to the relentless demand for AI-centric infrastructure. Revenue specifically from AI products skyrocketed 106% to $8.4 billion, fueled by massive orders for AI accelerators and high-speed networking chips. Broadcom expects AI-related revenue to surpass the $10 billion mark in the current quarter alone.

The $100 Billion Ambition

During the earnings call, Tan set an ambitious target, projecting that Broadcom’s AI revenue could exceed $100 billion by fiscal year 2027. Addressing skepticism regarding such a massive figure, Tan pointed to the firm’s locked-in pipeline with "hyperscale" data center clients. Broadcom’s current workload includes massive power-scale projects:

• Anthropic: 3 Gigawatts (GW)

• Google: 3 Gigawatts (GW)

• Meta: At least 2 Gigawatts (GW)

• OpenAI: 1 Gigawatt (GW)

With nearly 9 GW of total capacity across these AI titans, Tan remains confident that the $100 billion valuation is well within reach.

In the past, investors looked at the number of "chips" sold. But in 2026, "power capacity" will become a more accurate revenue indicator, because GW-scale data centers are massive "AI factories" that require enormous numbers of Broadcom networking and specialized ASIC chips to connect the systems together.

Broadcom doesn't just sell standard chips. However, as a leader in providing custom AI accelerators (ASICs) to giants, Google and Meta's expansion of multi-GW data centers means the demand for chips specifically designed in collaboration with them will grow permanently (recurring and sticky revenue).

As data centers grow larger, the bottleneck isn't the processors, but "data transfer." Broadcom's Tomahawk and Jericho chip families are key to 800G and 1.6T connectivity, which competitors struggle to match, allowing Broadcom to dominate almost the entire AI network infrastructure market.

Hock Tan's bold announcement of figures extending to 2027 is due to the fact that GW-scale data center projects require 2-3 years for planning and construction. This means these revenues are "already booked" (backlog), not just baseless sales estimates.

 

Google and Cloudflare Test Merkle Tree Certificates to Save the Post-Quantum Web. 

 

Source: CNBC 

 

💬 AI Content Assistant

Ask me anything about this article. No data is stored for your question.

 

Broadcom Shashes Records: AI Revenue Surges 106% as CEO Projects $100 Billion Milestone

Broadcom Inc. has reported stellar financial results for its first fiscal quarter of 2026, ended February 1. The technology giant posted consolidated revenue of $19.31 billion, a 29% increase year-over-year. Net income (GAAP) climbed to $7.35 billion, representing a robust 34% growth compared to the same period last year.

The AI Engine Driving Growth

CEO Hock Tan attributed the record-breaking performance to the relentless demand for AI-centric infrastructure. Revenue specifically from AI products skyrocketed 106% to $8.4 billion, fueled by massive orders for AI accelerators and high-speed networking chips. Broadcom expects AI-related revenue to surpass the $10 billion mark in the current quarter alone.

The $100 Billion Ambition

During the earnings call, Tan set an ambitious target, projecting that Broadcom’s AI revenue could exceed $100 billion by fiscal year 2027. Addressing skepticism regarding such a massive figure, Tan pointed to the firm’s locked-in pipeline with "hyperscale" data center clients. Broadcom’s current workload includes massive power-scale projects:

• Anthropic: 3 Gigawatts (GW)

• Google: 3 Gigawatts (GW)

• Meta: At least 2 Gigawatts (GW)

• OpenAI: 1 Gigawatt (GW)

With nearly 9 GW of total capacity across these AI titans, Tan remains confident that the $100 billion valuation is well within reach.

In the past, investors looked at the number of "chips" sold. But in 2026, "power capacity" will become a more accurate revenue indicator, because GW-scale data centers are massive "AI factories" that require enormous numbers of Broadcom networking and specialized ASIC chips to connect the systems together.

Broadcom doesn't just sell standard chips. However, as a leader in providing custom AI accelerators (ASICs) to giants, Google and Meta's expansion of multi-GW data centers means the demand for chips specifically designed in collaboration with them will grow permanently (recurring and sticky revenue).

As data centers grow larger, the bottleneck isn't the processors, but "data transfer." Broadcom's Tomahawk and Jericho chip families are key to 800G and 1.6T connectivity, which competitors struggle to match, allowing Broadcom to dominate almost the entire AI network infrastructure market.

Hock Tan's bold announcement of figures extending to 2027 is due to the fact that GW-scale data center projects require 2-3 years for planning and construction. This means these revenues are "already booked" (backlog), not just baseless sales estimates.

 

Google and Cloudflare Test Merkle Tree Certificates to Save the Post-Quantum Web. 

 

Source: CNBC