Intel’s early microprocessor development systems brought in more revenue than the chips
Federico Faggin’s retrospective describes the tools customers needed before a programmable component could become a usable product.

Intel had a new kind of programmable chip to sell, and customers needed a way to make it do something. According to Federico Faggin’s retrospective, development systems were bringing Intel more revenue than microprocessor chips a year after the microprocessor’s introduction. The supporting equipment briefly had the larger receipt.
The account preserved here is Faggin’s own history of the early microprocessor years. It follows the work around the 4004 and 8008, including customer skepticism, package constraints and the effort to turn a general-purpose idea into something somebody could build into a product. The chip was only one part of that job.
A programmable component shifts work toward software. That flexibility is attractive until a customer needs to develop, debug and deliver the program. A working processor without a practical development route can leave the next team staring at a very sophisticated part. Congratulations on the silicon; now the blinking cursor needs a plan.
Faggin describes the commercial result as “more revenues from development systems than from microprocessor chips.” The comparison is revenue, not unit sales or profit. It also describes a particular early period, rather than an enduring rule that Intel’s tools business always exceeded its processor business. The distinction keeps an excellent fact from becoming an imaginary financial history.
The surrounding account shows why the path to adoption needed attention. Customers criticized architecture and performance, with more computer-oriented organizations sometimes especially dismissive. Faggin also describes wanting a later design to use a 40-pin package instead of the 8008’s 18 pins, to improve interfacing alongside other changes.
Those details put the development-system figure in context. A customer doesn’t buy a component in isolation from the work of connecting it, programming it and establishing that it behaves. Tools that remove those obstacles can be a meaningful product while the component’s market is still taking shape.
That isn’t proof that every platform company should sell expensive tools. It does show why counting only the headline component can miss where early customers spend their money. The dull-looking development apparatus may be what makes the exciting invention practical enough to adopt.
The account also describes the effort to secure approval for the successor that became the 8080. Architecture, fabrication, packaging and management decisions all appear in the same story. There wasn’t one finish line where inventing a microprocessor automatically completed the business around it (if only product launches worked that way).
For a modern builder, the useful parallel is the gap between showing a capability and making it usable by someone else. A developer kit, diagnostic tool or reliable example can determine whether that next person ever reaches their first result. Faggin’s revenue comparison makes that support work unusually visible.
The tools made their own sale.