Most people think of Broadcom Inc. (AVGO) as a chip company. And for good reason, it’s been one of the biggest beneficiaries of the AI hardware boom, riding a wave of custom silicon demand straight into the $2 trillion market capitalization club.
But on April 22, 2026, Broadcom did something that had nothing to do with chips. The company announced an expanded collaboration with Google Cloud. Not around semiconductors, or Tensor Processing Units, but around network observability.
The two companies launched Cloud Network Insights, a new first-party service built on Broadcom’s AppNeta technology, designed to give organizations end-to-end visibility across the kind of sprawling, multi-cloud, hybrid environments that modern AI workloads increasingly demand.
Broadcom’s own Chief Information Officer Alan Davidson put it plainly: “Our global workforce depends on seamless connectivity to a mix of on-premises and cloud-based applications. We rely on the AppNeta technology behind Cloud Network Insights to gain granular visibility into that traffic, allowing our IT teams to isolate network issues from application issues immediately.”
That quote from the company’s own CIO. Not a customer, not an analyst, says something about how seriously Broadcom takes this product. And it signals that this partnership with Google Cloud runs deeper than most investors realize.
How Broadcom’s AppNeta Tech powers Google Cloud Network
The product at the center of the announcement is Cloud Network Insights. A first-party observability service built exclusively for Google Cloud users, launched April 22, 2026, and powered by Broadcom’s AppNeta technology.
The timing is deliberate. Modern AI workloads, including the agentic AI systems that enterprises are rapidly deploying, don’t live in a single cloud. They span multiple clouds, on-premises environments, SaaS services, and high-performance networking infrastructure simultaneously.Â
Related: Goldman Sachs resets Broadcom stock forecast
Traditional monitoring tools weren’t built for that complexity. When something breaks, no one can tell whether it’s the network or the application, a problem that costs organizations significant time and money.
Cloud Network Insights is designed to solve exactly that. According to Broadcom, the service delivers:
- End-to-end network path visibility to monitor critical metrics and determine whether performance issues are network-related or application-related
- Agent and application experience monitoring across on-premises, cloud, and SaaS environments in a single interface
- Rapid issue isolation to reduce Mean Time to Resolution, with synthetic transaction monitoring that catches SaaS and web application problems before users are affected
Rob Enns, VP and General Manager of Cloud Networking at Google Cloud, framed the stakes clearly.Â
“As organizations increasingly deploy critical agents and applications across complex multi-cloud and on-premises environments, agent, application, and network observability has never been more crucial,” Enns said.
Google Cloud’s Cross-Cloud Network currently handles 27 exabytes of data monthly, according to Google Cloud. In fact, that is a volume that makes the observability gap not just an inconvenience but a genuine operational risk for enterprise customers.
Broadcom’s Google partnership runs deeper than one product launch
The Cloud Network Insights announcement is the visible tip of a much larger strategic relationship between Broadcom and Google. We both know that this has been quietly becoming one of the most consequential partnerships in the AI infrastructure buildout.
Broadcom and Google have extended their agreement to develop and supply custom Tensor Processing Units through 2031, Reuters confirms. Those TPUs, including the latest-generation TPU Ironwood racks, are the custom silicon powering Google’s AI infrastructure at scale.
According to the Broadcom March 4, 2026 report, the financial weight of that relationship is reflected in their numbers.Â
In the first quarter of fiscal 2026, Broadcom reported:
- Total revenue of $19.31 billion, up 29.5% year over year, per the company’s Q1 FY2026 earnings
- AI chip revenue of $8.4 billion in Q1 alone, representing 106% year-over-year growth, according to Broadcom
- A target of more than $100 billion in total AI-related sales by 2027, heavily supported by the Google TPU pipeline, per company guidance, CNBC noted.
That AI chip revenue growth, more than doubling year over year, is the financial engine behind Broadcom’s recent ascent into the $2 trillion market capitalization club, making it the sixth company in U.S. history to reach that milestone, according to MarketWatch.
AVGO’s stock also closed April 22 at $422.65, up 5.09% on the day, hitting an all-time high. According to Yahoo Finance, the stock’s performance relative to the broader market tells the longer story:
- Year to date: AVGO +22.37% vs. S&P 500 +4.27%
- One year: AVGO +151.25% vs. S&P 500 +34.99%
- Three years: AVGO +594.92% vs. S&P 500 +72.68%
- Five years: AVGO +926.46% vs. S&P 500 +72.62%
Jonathan Raa/NurPhoto via Getty Images
Anthropic deal adds another boost to Broadcom’s Google Cloud strategy
If the TPU agreement is the foundation and Cloud Network Insights is the latest layer, the Anthropic dimension is what makes the Broadcom-Google relationship genuinely unusual in scope.
Anthropic confirms that they have committed to 3.5 gigawatts of new compute capacity through a three-way agreement involving Google Cloud and Broadcom, beginning in 2027, according to partnership details.Â
The revenue implications are significant. According to estimates cited in Broadcom’s strategic disclosures, the company could recognize approximately $21 billion in AI revenue specifically from the Anthropic deal in 2026, potentially rising to $42 billion in 2027, according to Yahoo Finance.
More AI:
- Micron sits at the center of a red-hot chip rally
- IBM CEO sends blunt message on AI and quantum computing
- Anthropic CEO makes shocking admission about AI
Those figures sit against the backdrop of Google Cloud’s own scale. Google Cloud is currently running at an annual revenue rate exceeding $70 billion, growing at 48% annually, with a backlog of $240 billion driven largely by enterprise AI demand, according to Alphabet’s most recent earnings results.
Serge Lucio, VP and General Manager of Broadcom’s Agile Operations Division, described the expanded collaboration in measured terms.Â
“AppNeta offers proven network performance monitoring and operates with a small footprint, making it an efficient and powerful choice for users,” Lucio said. The understatement is almost charming given the scale of what surrounds it.
What Broadcom’s partnership means for you
The Cloud Network Insights launch is easy to read as a minor product announcement. It isn’t. It represents Broadcom extending its footprint inside Google Cloud from hardware, where it already holds a dominant position through custom TPU supply, into software and services.
That expansion matters because software and services carry different margin profiles, different revenue durability, and different competitive moats than chip supply agreements.
More Tech Stocks:
- Morgan Stanley sets jaw-dropping Micron price target after event
- Nvidia’s China chip problem isn’t what most investors think
- Quantum Computing makes $110 million move nobody saw coming
For you, Broadcom is quietly assembling a position inside the AI infrastructure stack that spans silicon, custom chip agreements running through 2031, a three-way compute capacity deal with Anthropic, and now observability software embedded directly into Google Cloud’s first-party product lineup.
The chip story at Broadcom has been well-told. The fuller story of a company threading itself into every layer of how the world’s largest AI platforms are built and operated is still being written.
Related: Bank of America resets Broadcom stock forecast