Intel 1985: Exit Memory, Enter Microprocessors

Situation

It is mid-1985. Intel was founded in 1968 by Gordon Moore and Robert Noyce specifically to make memory chips. Memory — DRAM (Dynamic Random-Access Memory) — is Intel's identity, its origin, and the business that built the company. The Intel 1103 (1970) was the world's first commercially viable DRAM chip.

But the DRAM market in 1985 is collapsing:

1. Japanese manufacturers are flooding the market. Fujitsu, NEC, Hitachi, and Toshiba — backed by government coordination, cheap capital, and systematic reverse-engineering of American designs — have entered the DRAM market with relentless price competition. DRAM has become a commodity.
2. Intel is losing money on every memory chip it sells. DRAM revenue in 1985 is declining. The memory division is losing ~$200 million annually — a catastrophic drag on a company with ~$1.27 billion in total revenue.
3. Intel has something else. In 1981, IBM chose Intel's 8088 processor for the IBM PC. That decision — made partly because Intel was the only company that could supply at IBM's timeline — launched the IBM PC era. The x86 architecture now runs the world's dominant personal computers. But Intel still views microprocessors as a secondary business to memory.

Andy Grove is watching the spreadsheets. He is clear-eyed about what they say: at current trends, Intel runs out of cash within two years.

The pivotal conversation happens in Grove's office. He turns to Gordon Moore and says: "If we got kicked out and the board brought in a new CEO, what do you think they would do?" Moore answered without hesitation: "They would get out of memories." Grove replied: "Why don't we walk out the door, come back in, and do it ourselves?"

The decision moment

It is Q3 1985. Grove must drive the exit from memory. Three decisions:

1. Speed of exit. Shut down memory manufacturing immediately (aggressive — saves cash fastest, but destroys supplier relationships and strands 2,000+ employees overnight)? Or phase out over 18 months (slower — gives employees time to transition, but bleeds cash longer)?
2. Scope of commitment to microprocessors. Redirect all manufacturing capacity to the 386 processor (in development for IBM PC successors) and bet the company on x86? Or maintain a diversified chip portfolio (microprocessors + EPROMs + other logic chips) to reduce risk?
3. Communication. Memory is Intel's founding product. The workforce believes Intel is a memory company. How do you tell 25,000 employees that the reason the company exists is now being discontinued?

Grove knows this is not just a financial decision — it is a psychological one. Intel's engineers, its culture, its hiring pipeline, and its manufacturing processes were all built around memory. Exiting memory is not a product discontinuation. It is an identity change.

You are Andy Grove.

Key financial datapoints (for reference)

Frameworks invoked

• Strategic Inflection Point (Grove). Grove's own framework: a 10X change in a competitive force (Japanese government-subsidized pricing is a 10X shift in cost structure) that fundamentally alters the basis of competition. When a strategic inflection point hits, the old strategy stops working and continuing it leads to decline. The inflection point does not announce itself — it must be recognized.
• Portfolio Theory. Intel's resource allocation in 1985 is split between a declining, money-losing business (memory) and a growing, profitable one (microprocessors). Basic portfolio theory says kill the loser, feed the winner. The barrier is identity, not analysis.
• Core Competency Theory (Prahalad & Hamel). Intel's core competency is semiconductor design and manufacturing — not specifically memory. The competency transfers to microprocessors. But the organization has to be convinced that "Intel makes semiconductors" is the identity, not "Intel makes memory."
• Competitive Dynamics. The Japanese manufacturers are not competing on innovation — they are competing on cost, volume, and government subsidy. In a commodity market, the winner is the low-cost producer. Intel's cost structure cannot match the Japanese. The correct response is to exit the commodity and move up the value chain.

Discussion questions

1. Grove and Moore's "walk out the door" thought experiment is a classic tool for cutting through organizational inertia. What makes it effective, and when does it fail?
2. Intel's identity was built on memory. Abandoning memory means abandoning the story the company has told itself for 17 years. How do you manage that change without losing the culture that made the company successful?
3. The Japanese price competition was partly government-subsidized — arguably unfair. Intel could have lobbied for trade protection (and did). Should the answer to government-subsidized competition be political, strategic, or both?
4. Grove made the decision quickly once the analysis was clear. What information would you have needed to make this call in 1984, before the losses became existential?
5. Intel's bet on x86 depended on IBM PC becoming the dominant platform — which was not certain in 1985. Apple had ~15% PC market share. How do you bet the company on a platform bet when the platform outcome is uncertain?

The real outcome (revealed at session end)

1985: Grove begins the exit from DRAM. Manufacturing plants are converted to microprocessor production. ~2,000 employees are laid off. It is painful and fast.

1986: Intel launches the 386 processor — a breakthrough that defines the next generation of IBM-compatible PCs. It is faster, more powerful, and fully backward-compatible with the 8086/8088 family.

1987: Grove becomes CEO. He institutes the "Intel Inside" branding campaign — one of the most successful B2B-to-consumer brand initiatives ever.

1992: Intel becomes the most valuable semiconductor company in the world.

1997: Intel's market cap reaches $113 billion. Revenue hits $25B+. The DRAM business they abandoned is now a low-margin commodity war between Samsung, Micron, and the Japanese manufacturers.

Grove's legacy: Only the Paranoid Survive (1996) — Grove's account of the strategic inflection point concept — becomes required reading at every major business school. The core insight: companies that survive inflection points are those that recognize them while still healthy enough to act.

The lesson: The willingness to exit your founding business — your identity — is the rarest and most valuable strategic capability. It requires asking "what would a new CEO do?" — and then having the courage to do it before you have to.

Sources

• Andrew S. Grove, Only the Paranoid Survive (1996).
• Gordon Moore, "Intel: A Case Study" (Intel archives).
• Intel Annual Reports 1984-1992.
• Tim Jackson, Inside Intel (1997).
• MIT Sloan/Kellogg case: "Intel: Strategic Decisions at a Crossroads" (various editions).