When looking at stocks with some of the strongest upside potential over the next few years, Broadcom (AVGO +2.51%) tops the list. The chipmaker has been caught in the recent market sell-off, down around 20% in 2025 as of this writing. However, the stock is still up more than 45% over the past year.
Let's look at why Broadcom's stock could double, or more, over the next three years.
NASDAQ: AVGO
Key Data Points
A huge AI opportunity
Broadcom operates in two segments: semiconductor solutions and infrastructure software.
Originally focused solely on hardware, the company entered the software business through various acquisitions over the past few years, including CA Technologies, Symantec, and most recently VMware. With CA Technologies, the company acquired a large portfolio of mainframe and enterprise software solutions, while Symantec brought it an enterprise security business.
It biggest acquisition, though, was VMware, which it acquired in November 2023 for $69 billion. VMWare is a market leader in virtualization and cloud computing solutions. Its core virtualization product allows multiple virtual machines to be run on a single server, helping save customers costs. Meanwhile, its VMware Cloud Foundation (VCF) platform helps customers build hybrid and multicloud environments where they can manage workloads across public clouds and their own on-premise data centers.
Broadcom has set out to transform VMware by streamlining its operations and product solutions, while transitioning customers from perpetual licenses to a subscription model. At the end of its fiscal first quarter, it had transitioned 60% of its customers to a subscription model. In the process, it has also been upselling customers from its compute virtualization product, vSphere, to its VCF platform.
As a result, Broadcom's infrastructure software revenue climbed 47% to $6.7 billion last quarter and it forecast the segment to grow by 23% in fiscal Q2. Its software businesses should be a solid contributor moving forward, especially as enterprise customers start to turn to hybrid cloud environments to run artificial intelligence (AI) workloads.
On the hardware side, Broadcom makes components for various markets, including wireless devices, broadband internet, and automotive. However, its two most important businesses are networking and ASICs (application-specific integrated circuits).
Within its networking business, Broadcom makes components such as ethernet switches, optical receivers, digital signal processors (DSPs), and network interface cards (NICs). These components are critical parts of AI infrastructure.
Ethernet switches let AI chips rapidly exchange massive amounts of data, while optical receivers and DSPs work together to send data across longer distances. NICs, meanwhile, help connect servers to a network. Together, these components are essential in managing the flow of data and distributing AI workloads across servers and ensuring that high-performance computing resources are being best utilized.
As AI chip clusters grow in size and complexity, these components become even more important in helping maintain performance and efficiency.
Given the growing size of AI chip clusters, Broadcom's networking business has strong growth prospects ahead. However, Broadcom's ASICs business represents its biggest opportunity. With this business, Broadcom helps customers create their own custom AI chips designed for specific tasks.
Custom AI chips have better performance and consume less power than the mass-market graphic processing units (GPUs) from Nvidia, which are the primary chips used to train AI models and run inference. The downside is that there are a lot of upfront costs involved and it takes time to design a custom chip. These chips are also less flexible than GPUs, which can be programmed for a variety of tasks.
However, with the skyrocketing cost of Nvidia's GPUs, more companies have been turning to Broadcom to help them design custom AI chips. While custom chips require significant upfront costs, their reduced power consumption can lead to a lower overall cost of ownership over time.
Broadcom's first custom AI chip customer was Alphabet, for which it helped design the Tensor Processing Unit (TPU). These chips were developed specifically to optimize AI workloads within Google Cloud's TensorFlow framework, and have been highly successful at optimizing AI workloads and reducing costs.
Broadcom's success with Alphabet has led to a slew of new custom AI chip customers. The company sees its three AI chip customers furthest along in their development having a $60 billion to $90 billion serviceable market opportunity in its fiscal year 2027 (ending October 2027) alone. Meanwhile, it has added several other custom AI chip customers along the way, including Apple. While Broadcom won't capture this entire market opportunity, it should lead to very strong growth in the coming years.
Image source: Getty Images
Why Broadcom's stock can more than double in the next three years
Broadcom generated $51.6 billion in revenue in fiscal 2024, so if it were able to capture a nice chunk of its custom AI chip opportunity (say, $40 billion in additional revenue) and continue to grow VMware, it perhaps could grow its revenue to surpass $100 billion in fiscal 2027. At a 75% gross margin and few other added expenses, that could add an additional $10 in earnings per share (EPS) to the $4.87 in adjusted EPS it produced in fiscal year 2024.
Let's assume some added expenses and round down its fiscal 2027 EPS to $14.50. With Apple and other customers likely coming on in fiscal 2028, let's assume another $15 billion in revenue growth and add another $2.25 in EPS in fiscal year 2028, bringing the total to $16.75.
With a 25 times forward price-to-earnings (P/E) multiple at the end of 2027 on $16.75 in fiscal 2028 EPS, the stock would be worth more than $400 in three years. The stock currently trades at a forward P/E of near 28 times, so that multiple seems appropriate.
As such, Broadcom's stock looks like it has a nice opportunity to more than double in the coming years helped by a surge in custom AI chip and networking revenue and the continued growth of VMware.
