Full-Stack Bootstrap-Then-Divest Playbook
Temporarily own the whole ecosystem to launch a new standard, then divest the commodity layers once it takes hold.
- Difficulty
- Expert
- Time to result
- ~ongoing to results
- Steps
- 7
- Confidence
- 75%
When Qualcomm needed carriers to bet their networks on unproven CDMA technology, no single vendor's promise was enough — customers wanted proof that infrastructure, handsets, and chips would all actually exist together. Rather than wait for a mature ecosystem, Qualcomm patented the core IP years in advance, then temporarily built the entire value chain itself through majority-owned joint ventures with Nortel (base stations) and Sony (handsets) so customers could adopt the new standard without ecosystem risk. Once CDMA reached critical mass — including a full national rollout in South Korea — Qualcomm sold off the capital-intensive infrastructure and handset businesses to Ericsson and Kyocera, keeping only the highest-margin layers: fabless chip design and patent licensing. It's a playbook for entering markets where an entire ecosystem must exist before any single product can sell.
Origin
Grew out of Qualcomm's need to make carriers comfortable adopting CDMA in the early 1990s despite having no existing base station or handset ecosystem.
Core principles
- 01Patent the core invention years before a market for it exists
- 02When no ecosystem partner will take first-mover risk, temporarily own the whole value chain yourself via joint ventures
- 03Use lead markets or customers with concentrated buying power to prove the model at scale
- 04Divest capital-intensive, low-margin layers once the standard has independent momentum
- 05Keep and compound the IP/licensing and design layers where the real margin lives
- 06Continuously refile patents each successive technology generation so the moat doesn't expire with the first patent
How to run it
- 1
Patent the core invention early
File on the foundational technology years before the market that will need it actually exists.
Watch out Miss the adoption window for one generation and you may not survive to the next.
- 2
Prove feasibility with hard, public demos
Run live demonstrations for skeptical industry stakeholders to counter disbelief that the technology can work in practice.
Pro tip Accept the harder challenge test (e.g., a dense urban environment) when doubters raise it — passing it converts skeptics.
- 3
Temporarily own the entire value chain
Where ecosystem partners won't take the adoption risk alone, form majority-owned joint ventures to supply every missing layer (infrastructure, handsets) yourself.
Pro tip Structure JVs so you retain majority control and can unwind them later.
Watch out This requires raising very dilutive financing — be honest with the team and investors about the trade-off.
- 4
Win a concentrated lead market
Target a customer or country that can adopt the full stack in one decision (e.g., a government mandate) to build proof and scale fast.
- 5
Divest the commodity layers once momentum is independent
Once other credible suppliers exist and the standard is self-sustaining, sell off the capital-intensive manufacturing businesses.
Watch out Waiting too long to divest lets low-margin manufacturing drag down your stock and tie up capital needed for R&D.
- 6
Concentrate on IP and design
Keep only the highest-margin, most defensible layers — fabless chip design and patent licensing — and reinvest cash flow into R&D.
- 7
Refresh the patent portfolio each generation
File new, essential patents for every successive technology generation so your licensing leverage doesn't expire with the original patent.
In the wild
Qualcomm formed majority-owned JVs with Nortel to manufacture CDMA base stations and with Sony to make CDMA handsets, so carriers could adopt CDMA with a complete solution rather than assembling one themselves.
→ Carriers gained confidence to commit to CDMA, driving adoption in the US, South Korea, and beyond.
Once CDMA had reached critical mass, Qualcomm sold its infrastructure business to Ericsson and its handset business to Kyocera, exiting capital-intensive manufacturing entirely.
→ Qualcomm became purely a chip design and licensing company, unlocking much higher margins and eventually becoming the largest fabless semiconductor company in the world.
Common mistakes
Staying in manufacturing too long
Holding onto capital-intensive infrastructure and handset businesses after the standard was secure suppressed the stock and tied up cash that could have funded R&D.
Straying into unrelated capital-heavy bets
The $2 billion Mirasol display fab, outside Qualcomm's core competency, produced zero customers and wasted capital.
Over-squeezing ecosystem partners
Pushing licensing fees so hard that key customers like Apple sued and began building in-house alternatives, eroding long-term leverage.
Is it for you?
Not ideal for
Markets that don't require a multi-sided ecosystem to function, or founders without access to significant dilutive capital.
From the transcript
“I think we kind of gotta do everything ourselves. We need to offer the complete solution.”
“In March of 99, they sell their infrastructure business, the base stations, to Ericsson, which was formerly one of their competitors.”
From the episode
Qualcomm