If someone asked you to name a large AI chip manufacturer, you’d probably say Nvidia (NVDA), Advanced Micro Devices (AMD), or maybe Taiwan Semiconductor (TSM). Your first thought probably wouldn’t be Broadcom (AVGO).
But Broadcom is actually a giant in the AI hyperscaler space with a $1.6 trillion market cap. Up 54% over the past year, the company has found defining success by pursuing only seven to 10 main customers who need billion-dollar buildouts rather than chasing many small $10 million contracts.
That strategy is paying off in a major way. Broadcom reported an AI order backlog of $73 billion to be delivered over the next six quarters, with management expecting that figure to continue growing as new orders roll in.
JPMorgan recently raised its price target on the stock and maintained its “Overweight” rating, calling the recent dip a buying opportunity. So what exactly does Broadcom do, and why are Wall Street’s biggest investors loading up on shares?
Broadcom has achieved its success through category-leading semiconductor and infrastructure software solutions. The company operates two primary business segments: Semiconductor Solutions and Infrastructure Software.
On the semiconductor side, Broadcom is the leader in custom ASICs (application-specific integrated circuits) designed specifically for data center AI workloads.
ASICs are specialized chips built to exact specifications for hyperscalers such as Google (GOOGL), Meta (META), and Amazon (AMZN), which need massive computing power.
The infrastructure software segment, which includes the VMware acquisition, provides enterprise software for planning, developing, automating, managing, and securing applications across different platforms.
With roots in the technical heritage of AT&T/Bell Labs, Lucent, and Hewlett-Packard/Agilent, Broadcom combines more than 60 years of innovation with global scale and engineering depth. The company serves the world’s most successful tech companies by focusing on technologies that connect our world.
CEO Hock Tan explained the strategy at the Goldman Sachs technology conference: Broadcom is laser-focused on about seven customers building large language models and racing toward what he calls “super intelligence.” These companies are collectively spending around $30 billion annually on AI compute infrastructure.
Rather than pursuing thousands of enterprise customers who might spend $10 million each on merchant GPUs, Broadcom sees its biggest opportunity in providing custom chips and networking infrastructure for these hyperscalers building AI at unprecedented scale.
JPMorgan analyst Harlan Sur pointed to the continued surge in capital expenditure by hyperscalers and the consequent rise in demand for ASIC chips as key drivers for Broadcom’s growth.
Broadcom reported Q4 2025 revenue of $18.02 billion, up 28% year over year. The semiconductor segment jumped 35% to $11.07 billion, while the high-margin infrastructure software segment grew 19% to $6.94 billion.
For the full fiscal year 2025, Broadcom’s revenue hit $63.9 billion, a 24% increase driven overwhelmingly by AI, which saw revenue grow 65% year-over-year to $20 billion.
But here’s where it gets really interesting: Broadcom forecasts its AI business will hit $8.2 billion in Q1 2026 alone, marking roughly 100% year-over-year growth. And Tan revealed that his compensation package tied to 2030 targets includes achieving AI revenue over $120 billion — a sixfold increase from the $20 billion generated in fiscal 2025.
One customer prospect turned into over $10 billion in orders expected in the second half of fiscal 2026—a major driver behind this is that Broadcom designs and manufactures Google’s custom TPU chips, a specialized ASIC built specifically for AI tasks rather than general-purpose computing (GPU). Google’s latest TPUv7 chip is performing better than expected and attracting major customers like Anthropic and Meta, with Anthropic alone committing to potentially 1 million units.
Many investors have been trying to make the most of the AI trade, but when AVGO dipped over 27% in Q1 2025 — hitting a yearly low of $138 in April — that was a green light for some of Wall Street’s most famous money managers.
These investors spotted what Michael Burry later articulated in his new Substack: while 40%-50% of hyperscaler spending goes to Nvidia GPUs with “ever shorter product cycles,” Broadcom is positioned differently.
Burry specifically named Broadcom alongside Google, Amazon, and Microsoft as “very well-funded” and capable of challenging Nvidia’s dominance.
Major Q1 2025 buyers included:
-
Jeremy Grantham (GMO): Bought over 1.5 million shares despite calling AI a bubble. His fund holds a $291 million position in Broadcom versus just $21 million in Nvidia — nearly 14x more exposure to AVGO than NVDA.
-
Renaissance Technologies: Purchased over 2.4 million shares, massively increasing its stake. Regarded as the most successful hedge fund of all time, Renaissance’s bet signals strong conviction in Broadcom’s differentiated positioning.
-
Two Sigma Advisors: Bought around 2.5 million shares, bringing its total position to 3.9 million shares — the largest holding since Q3 2024 when Broadcom traded at just $159 versus Q1’s average of $211.
-
Bridgewater Associates (Ray Dalio): Added 280,000 shares, betting that Broadcom’s custom ASIC strategy offers structural advantages over commodity GPU manufacturers.
The stock has since recovered to around $330. While Burry’s warnings about hyperscaler capex accounting hit other AI stocks, Broadcom’s $73 billion backlog of multi-year contracts with just 7-10 customers provides revenue certainty that short-cycle GPU sales can’t match.
Broadcom stock has pulled back from December highs near $415 to around $330. If you’ve been on the sidelines, here are three bullish indicators worth considering:
1. The structural moat is different. While Nvidia sells commodity GPUs facing shorter product cycles and growing competition, Broadcom has locked in $73 billion in backlog with just 7-10 hyperscaler customers through multi-year custom chip contracts. These aren’t off-the-shelf components that competitors can easily replicate.
2. Wall Street’s smartest money doubled down. Renaissance Technologies, Jeremy Grantham, Two Sigma, and Ray Dalio all bought aggressively when the stock dipped in Q1. Grantham holds 14 times more Broadcom than Nvidia despite calling AI a bubble, betting that custom ASICs offer better risk-adjusted returns than merchant chip sales.
3. Broadcom (mostly) dodged the Burry selloff. Michael Burry’s warnings about hyperscaler capex accounting triggered the recent tech selloff, but he specifically named Broadcom as well-positioned to challenge Nvidia’s dominance. The company’s custom chip focus sidesteps the depreciation and replacement cycle concerns hitting pure-play GPU manufacturers.
The stock trades at 61x earnings, but with 100% AI revenue growth and signed contracts providing multi-year visibility, Broadcom offers a differentiated AI infrastructure play built on structural advantages rather than pure momentum and hype.
On the date of publication, Justin Estes did not have (either directly or indirectly) positions in any of the securities mentioned in this article. All information and data in this article is solely for informational purposes. This article was originally published on Barchart.com
