Small Cost, Big ConsequenceWide moat

The purest expression of Broadcom's pricing power is this: many of its chips are a trivial fraction of a system's total cost while being absolutely essential to whether that system works at all. A networking chip or a specialized component might account for a few percent of a server's or a phone's bill of materials, yet if it fails or underperforms, the whole expensive product is worthless. When a part is cheap but mission-critical, the customer's overwhelming priority is that it work flawlessly — not that it be a few dollars cheaper.

This asymmetry hands Broadcom enormous latitude on price. A customer building a hundred-thousand-dollar piece of equipment will not risk its reliability to shave a few dollars off a critical component, so Broadcom can price to the value the chip delivers rather than to the cost of making it. The buyer's incentive is to pay up for certainty, and Broadcom has built its business around exactly the parts where that logic holds most strongly.

The limit is that 'small cost, big consequence' protects the price of a given design but does not guarantee the design is chosen next time. Customers accept the pricing on parts already engineered into their products, but for the next generation they can seek alternatives, design the function in-house, or integrate it elsewhere — which is why Broadcom must keep winning the technical contest, not merely harvest the sockets it holds. The advantage is real and durable within a product cycle, and must be re-earned at the start of each new one — as the ~$20B AI franchise is being re-earned right now¹.