Company Overview
Broadcom has emerged as the AI infrastructure winner that many investors overlooked while chasing Nvidia’s meteoric rise. In October, the company announced a landmark partnership with OpenAI to deploy 10 gigawatts of custom AI accelerators and rack-level systems over multiple years—a deal that analysts estimate could generate $70-90 billion annually once fully deployed starting in late 2026. OpenAI will design the accelerators while Broadcom provides the manufacturing expertise and Ethernet-based networking infrastructure, creating vertically integrated systems optimized specifically for OpenAI’s workloads.
But that’s not all. Analysts have confirmed that Broadcom’s previously disclosed $10 billion “mystery customer” order is likely Anthropic, making Broadcom the exclusive AI accelerator provider for both of the world’s leading AI model developers. Combined with existing partnerships with three major hyperscalers (Meta, Alphabet, and ByteDance), Broadcom now has five major custom AI chip customers—positioning it as the picks-and-shovels provider for the entire AI infrastructure buildout. With fiscal Q4 2025 earnings scheduled for mid-December, investors are positioning ahead of what could be a significant catalyst.
Key Technical and Fundamental Drivers
OpenAI Mega-Deal → $300B+ Revenue Potential The October 13th announcement of 10GW AI accelerator deployment with OpenAI represents one of the largest AI hardware commitments ever, with Financial Times estimating the total deal value at $350-500 billion over multiple years.
Anthropic Partnership → $10B Order Confirmed Analysts confirmed Broadcom’s $10 billion “mystery customer” from Q3 earnings is Anthropic, involving TPU ASICs co-designed with Google plus networking and rack-scale deployment beginning in Q3 2026.
Multiple Analyst Upgrades → Targets to $460 Following the OpenAI deal, Mizuho raised its price target to $435, KeyBanc and Barclays to $450-460, and Goldman Sachs to $380, with consensus “Strong Buy” rating across Wall Street.
Expanding AI TAM → $146B by FY2027 Analysts project Broadcom’s AI Total Addressable Market expanding to ~$146 billion by FY2027 as custom AI accelerators gain share versus traditional GPUs, with AVGO offering typically cheaper alternatives to Nvidia’s pricier offerings.
December Earnings Catalyst → Q4 Results Ahead Broadcom reports fiscal Q4 2025 earnings in mid-December, with analysts expecting accelerated top/bottom-line growth at a CAGR of 40%+ driven by the expanding AI pipeline and VMware integration benefits.
Market Takeaway
Broadcom represents the ultimate AI infrastructure play that many investors missed while fixating on Nvidia’s GPU dominance. The company’s strategic positioning is brilliant: rather than competing head-to-head with Nvidia in the general-purpose GPU market, Broadcom is capturing the custom AI accelerator segment—chips designed specifically for individual customers’ workloads. This approach offers several advantages: deeper customer lock-in through co-development, higher margins on custom silicon, and reduced cyclicality through long-term multi-year contracts.
The OpenAI partnership is transformative. By designing chips optimized specifically for OpenAI’s inference workloads and embedding them in complete rack-level systems with Broadcom’s Ethernet networking, the solution delivers better performance-per-dollar than off-the-shelf alternatives. OpenAI CEO Sam Altman described the approach as controlling “from etching the transistors all the way up to the token that comes out when you ask ChatGPT a question,” enabling “huge efficiency gains” leading to faster, cheaper models. Deployment begins in late 2026 and runs through 2029, providing Broadcom with highly predictable revenue visibility.
The Anthropic deal adds another $10 billion layer, with shipments beginning in Q3 2026. Combined with partnerships with Meta (using custom chips for both training and inference), Alphabet/Google (custom TPUs), and ByteDance, Broadcom has systematically built a portfolio of the world’s most important AI companies as customers. This diversification reduces single-customer risk while positioning Broadcom to benefit regardless of which AI model provider ultimately dominates.
The stock has pulled back recently amid broader semiconductor sector weakness, creating an attractive entry point ahead of December earnings. Trading at 38x forward earnings—a premium to the market but in line with AI peers like AMD (37x)—the valuation appears reasonable given the explosive growth trajectory. Analysts project AI revenues reaching $41 billion in FY2026 and potentially doubling to $82.7 billion by FY2028, representing a CAGR well above 40%. With industry-leading gross margins, a diversified business model spanning AI networking, custom silicon, and the recently acquired VMware software business, Broadcom offers investors exposure to multiple secular growth drivers wrapped into a single stock. As one Motley Fool analysis concluded, if you’re looking for “the one AI semiconductor stock to buy hand over fist before December—hint: it’s not Nvidia”—Broadcom is the answer.