Qualcomm: how CDMA patents built a wireless monopoly

Executive overview

Wireless spectrum is finite, and whoever controls the most efficient method of sharing it controls the economics of every phone on earth. Qualcomm's founders — academic engineers who studied under Claude Shannon himself — identified that code division multiple access (CDMA) was theoretically superior to every competing wireless standard before a single commercial network existed. They patented the method years in advance, bootstrapped revenue through satellite and trucking contracts, then forced the entire global industry onto their technology.

The core insight: control the standard, collect a royalty on every device forever — and never touch manufacturing once you no longer need it to bootstrap adoption.

The unlikely origins of spread spectrum

• Hedy Lamarr, Hollywood actress and covert observer of Nazi arms meetings, co-invented frequency-hopping spread spectrum in 1942 to defeat radio jamming of torpedoes.
• The patent was classified by the US military for nearly four decades.
• Claude Shannon independently formalised information theory in 1948, defining the theoretical maximum signal capacity of any medium — the mathematical foundation all of Qualcomm's work would rest on.
• Shannon's student Irwin Jacobs — originally enrolled in Cornell's hotel management programme — completed his PhD under Shannon in three years and wrote the first digital communications textbook, still in use today.

From consulting firm to foundational patent

• Jacobs and Andrew Viterbi co-founded Linkabit in the early 1970s to manage academic consulting work, mostly US military satellite projects.
• Linkabit's first commercial contract was building the satellite communication system for Walmart.
• They also designed pay-TV scrambling for HBO and the early cable industry.
• Linkabit sold to M/A-COM in 1980 for $25 million; Jacobs and Viterbi stayed five years, then left.
• In July 1985, seven engineers met at Irwin's house and founded Qualcomm — short for Quality Communications.
• In 1986 Qualcomm filed US Patent 4,901,307: CDMA applied to terrestrial cellular networks. The patent was designed to expire in 2006 — just as 3G was maturing.

Bootstrapping with OmniTRACS

• Unable to attack cellular directly, Qualcomm won a contract with OmniNET to build a mobile satellite network for commercial trucking.
• Qualcomm and OmniNET merged in 1988, raising $3.5 million in exchange for 50% equity — severe dilution, but the only viable path to capital.
• The product launched as OmniTRACS in late 1988; year-one revenue was $32 million (~$100 million inflation-adjusted).
• Walmart was again an early customer, using the system on its proprietary truck fleet.
• OmniTRACS cash flow gave Qualcomm the financial base to pursue CDMA in cellular.

The CDMA standard wars

• In September 1988 the US Cellular Telecommunications Industry Association published the 2G performance spec; the front-runner standard (TDMA, championed by Ericsson in Europe) could not meet it.
• Unlike Europe, the US did not mandate a single standard — carriers could use any technology that met the performance spec. Qualcomm confirmed this with regulators, then launched a carrier road show.
• The "holy wars of wireless" began: Qualcomm argued CDMA offered 3–5x more capacity than TDMA per unit of spectrum; sceptics doubted the required real-time signal processing was feasible.
• Pactel funded a $1 million prototype in early 1989; a live San Diego demo followed in November 1989.
• New York City demo in February 1990 secured nine-X mobile and Ameritech.
• South Korea's government mandated CDMA nationwide — at peak, South Korea represented ~40% of Qualcomm's revenue.
• Qualcomm went public in December 1991, raising $68 million.
• The US standards body formally adopted CDMA as a second standard in 1993 — by which point Qualcomm had already signed half the industry.

The complete-solution gambit

• Carriers needed four things to deploy CDMA: the core IP, base station infrastructure, consumer handsets, and the semiconductor chips to power both.
• Qualcomm held the IP but lacked manufacturing. Rather than find partners and hope, they built everything themselves.
• Joint venture with Nortel to manufacture CDMA base stations.
• Joint venture with Sony (51% Qualcomm / 49% Sony) to manufacture handsets.
• Qualcomm designed its own CDMA chipsets — the modem silicon for both infrastructure and phones — using fabless manufacturing via TSMC, which had only just become viable.
• This captured value at both ends of the smiling curve: IP and semiconductors; commoditised manufacturing went to foundries.

Shedding manufacturing, crystallising the business model

• By the late 1990s the joint ventures were consuming capital and depressing the stock.
• March 1999: Qualcomm sold its base station manufacturing business to Ericsson as part of a cross-licensing settlement.
• December 1999: Kyocera bought Qualcomm's mobile phone business.
• Qualcomm was left with two pure divisions: QCT (chip design) and QTL (technology licensing).
• In the year 2000 Qualcomm was the single best-performing stock in the US market, appreciating 2,621%.
• Revenue: $383 million in 1995, $814 million in 1996; today ~$44 billion annually.

The patent licensing model

• Every handset maker using CDMA-based standards (2G, 3G, 4G LTE, 5G) must license Qualcomm's patents.
• Royalty structure: a percentage of the phone's wholesale price — not a flat chip fee — meaning higher-priced phones generate higher royalties.
• Apple was estimated to pay $10–$18 per iPhone, on top of ~$30 per baseband chip.
• Apple attempted to switch to Intel modems for one iPhone generation; Intel fell behind on 5G, forcing Apple to settle in 2019 with a ~$4.5 billion six-year licensing deal and a commitment to continue using Qualcomm modems.
• Apple has since acquired Intel's modem division and is developing its own chips; Qualcomm's CEO has guided for near-zero Apple chip revenue after the current supply deal expires.

The snapdragon era and the nuvia bet

• During the 2000s Qualcomm expanded from modem chips into full Snapdragon mobile systems-on-chip, becoming the default high-end processor for Android phones.
• MediaTek emerged as a low-cost competitor; it now ships more units than Qualcomm, taking the mid- and low-end Android market.
• In 2021 Qualcomm acquired Nuvia for $1.4 billion — a team founded by the chief architect of Apple's A-series chips.
• Nuvia's mandate: custom CPU designs using the ARM instruction set, matching Apple Silicon's performance differentiation rather than using stock ARM cores.
• This positions Qualcomm to compete in laptops, automotive, and IoT with genuine performance rather than commodity silicon.

Strategic position and powers

• Cornered resource: ~17,000 patents covering essential wireless technologies across every generation of cellular standards.
• Scale economies: chip R&D costs amortised across billions of units make the business nearly impossible to replicate from scratch.
• Network effects (historical): controlling the infrastructure standard locked in handset makers, and vice versa.
• Process power (historical): the original Qualcomm brain trust — Shannon-trained engineers with unique domain knowledge — was a once-in-a-generation concentration of talent.
• Revenue today: ~$44 billion; 85% chips (QCT), 15% licensing (QTL) — but QTL runs at ~69% margin versus ~34% for QCT.

Risks and the road ahead

• Apple, Samsung, and others are building in-house modems to escape royalty payments.
• Qualcomm has historically failed in every market outside handsets: IoT, servers, smartwatches, displays.
• The bear case: customers finally irritated enough to invest in alternatives; licensing leverage erodes as competitors mature.
• The bull case: automotive (currently $2 billion in revenue), IoT ($7 billion), and the 5G RF front end ($4 billion) represent a credible $100+ billion addressable market; the intelligent connected edge thesis holds if wireless becomes the primary interface between cloud and physical world.
• A 2018 hostile takeover bid by Broadcom ($117 billion, largely debt-financed) was blocked by the Trump administration on national security grounds — widely seen as Qualcomm leveraging its long relationship with the US government.