Custom Silicon Kings: When to Choose Broadcom and When to Choose Marvell Technology
Broadcom and Marvell both design custom chips for the same hyperscaler clients, both raised their outlooks, and both trade on the AI ASIC boom. So why do their risk profiles, valuations, and investor theses point in completely opposite directions?
Broadcom (NASDAQ:AVGO | AVGO Price Prediction) closed fiscal Q2 with $22.2 billion in revenue, while Marvell Technology (NASDAQ:MRVL) reported fiscal Q2 revenue of $2.739 billion on August 27. Both design custom silicon for hyperscalers, both raised outlooks, and both ride the same XPU wave. Each just re-anchored investor expectations for what a custom ASIC franchise can deliver at very different scales.
Hyperscaler ASICs Power One. XPU Attach Powers the Other.
Hock Tan framed Broadcom’s quarter around $10.8 billion of AI semiconductor revenue, up 143% year-on-year, with bookings over $30 billion against that shipment base. Networking made up almost 40% of Q2 AI revenue, spanning Tomahawk 6 switches, 1.6T DSPs, and Jericho fabrics. VMware added $7.2 billion at 93% software gross margin.
Marvell’s story is narrower. Data Center delivered $2.17 billion, or 79% of total sales, growing 46% year-over-year. Matt Murphy highlighted “a significant acceleration in custom demand in the second half of this year” and disclosed an expanded hyperscaler warrant covering inference accelerators, storage controllers, NICs, and near-memory compute.
| Business Driver | Broadcom | Marvell |
| Core AI Bet | TPUs, MTIA, OpenAI silicon, Ethernet fabric | Google warrant, XPU-attach, scale-up optics |
| Margin Cushion | VMware software at 79% operating margin | Non-GAAP op margin 36.6%, targeting 38-40% |
| Named Customers | Google, Meta, OpenAI, Entropic | One key hyperscaler warrant, broad Tier 1 pipeline |
Scale Platform Versus High-Beta Growth Curve
Broadcom sells a platform. Tan pitched an AI XPV platform with Apollo and Blackstone to deploy more than 20 gigawatts of compute capacity through 2028, with fiscal 2027 AI silicon revenue guided in excess of $100 billion. Free cash flow ran at 46% of revenue, and forward P/E sits at 20x.
Marvell sells a growth curve. Management raised fiscal 2028 data center outlook to more than 60% year-over-year, with custom expected to more than double. Shares have run 237.41% over the past year, and beta of 2.246 reflects the volatility. Forward P/E of 60x demands the ramp materialize.
Fiscal 2027 Ramp Is the Real Referee
Watch whether Broadcom’s 10 gigawatts of planned 2027 shipments land on schedule for Google, Meta, and OpenAI, backing the $56 billion fiscal 2026 AI revenue target. For Marvell, the tell is whether scale-up optics and the warranted hyperscaler’s XPU-attach programs convert into revenue Murphy hinted could be “a lot larger than anybody’s been modeling” in fiscal 2029.
How the Risk Profiles Diverge
For lower-beta scale with pricing power, Broadcom screens as the steadier profile. The VMware margin cushion and structural networking dominance, paired with analyst target of $525.97, offer a franchise to hold through cyclical dips. For pure AI beta exposure, Marvell offers the higher-torque profile. Its smaller $194 billion market cap means each incremental Google TPU-attach socket moves the needle harder, and analyst target of $278.89 reflects that leverage (we studied what the early signals of a 100x tech winner look like and put them in a free playbook here). Both names face shared downside if hyperscaler capex guidance softens into 2027, since customer concentration is the vulnerability neither company can engineer away.
Contact [email protected] for any questions or corrections.
Alex Sirois is a financial writer with experience spanning both retail and institutional investing. He has written for InvestorPlace and held roles at BNY Mellon and Bernstein, giving him a perspective that bridges Main Street portfolios and Wall Street analysis.
Alex holds an MBA from George Washington University and has built his career across multiple industries, including e-commerce, education, and translation — a breadth of experience that informs how he breaks down complex financial topics for everyday investors. His writing is conversational, actionable, and grounded in long-term, buy-and-hold investing principles.
At 247 Wall St., Alex focuses on delivering analysis that is both accessible and useful, with a clear emphasis on helping readers make more informed decisions with their money.
