Mesa Production 1958 Win Ibm Contract Commercial Volume
# Win the IBM Contract: Commercial Volume Production Begins
## 1. Why Signing the Order Is Easier Than Honoring It
This step is the moment Fairchild commits to supplying IBM with the 2N696 and 2N697 at a sustained volume and schedule, and committing to that number is a fundamentally different act than having demonstrated, once, that the line could produce it — a demonstration lot and a standing commercial order expose this process to a kind of risk none of the previous eleven steps had to manage: the risk of promising a rate and a cost before the line has proven it can hold either one indefinitely. Every step since Step 1 improved this line's capability in isolation; this step is where that capability has to translate into a number a customer can actually plan their own production around. The relevant economics here is not the single-wafer cost-per-device relation this project established in 1957, but a longitudinal one — cost falling with *cumulative* production experience, not merely with devices per wafer:
where $C_n$ is the cost of the $n$-th unit produced, $C_1$ the cost of the very first unit, $n$ cumulative volume, and $b$ a learning exponent set by $L$, the fractional cost reduction each time cumulative volume doubles. Unlike the 1957 series' per-wafer economics, which rewarded a bigger die count on one wafer, this relation rewards *time and repetition* — the line gets cheaper simply by making more of the same thing, as operators, tooling, and process control all improve with practice, independent of any change to the device itself.
## 2. Real Diagram: What a Standing Order Actually Commits the Line To
A single qualified lot, the kind Step 11 just validated, is a point in time. A standing commercial order is a promise about every week going forward — a sustained rate, a sustained yield, and a sustained cost, none of which any single lot's success actually guarantees on its own.
## 3. Why This Is the First Step in This Entire Project Driven by Someone Else's Calendar
Every step before this one — in the 1947, 1951, 1952, 1953, 1954, 1956, and 1957 series, and the first eleven steps of this series — was paced by this project's own process, this line's own furnace schedule, this line's own inspection cadence. This step introduces a pacing this process does not control: IBM's own production schedule, into which these transistors must arrive on time, in the agreed quantity, meeting the agreed specification, indefinitely. Fairchild's technical achievement across the previous eleven steps is what made this commitment possible to make in good faith; this step is where that achievement stops being a laboratory or factory-floor fact and becomes a contractual one, with consequences for Fairchild that no earlier step in this project's history ever had to answer to.
Step 12 does not improve the device or the process; it is the moment this project's story about *how a transistor is made* becomes, for the first time, a story about *whether a transistor can be made reliably enough, and fast enough, to keep a promise*.