How Apple Newton's $3M Gamble Created the ARM Empire Behind iPhone

The Navigator That Hit a Physical Wall

In 1987 Apple began the Newton project to realize then-CEO John Sculley's "Knowledge Navigator" concept. At the 1992 CES Sculley coined the term "PDA" (Personal Digital Assistant) and promised a handheld that could understand natural cursive handwriting, send faxes, manage calendars, and ship within a year. But the device immediately collided with 1990s physics: lithium-ion batteries had not yet entered consumer electronics, so portable devices relied on nickel-cadmium or AA alkaline cells. Real-time parsing of complex pen strokes required dense probabilistic computation, yet desktop processors of the era ran hot and power-hungry. Early prototype "Figaro" crammed three AT&T chips to barely run the interface, but drained batteries in minutes. The core deadlock: no existing chip could run a complex OS while surviving on a few dry cells.

To save Newton, project lead Larry Tesler decided to build a custom chip. In November 1990 Apple invested $3 million , partnered with 12 engineers from UK's Acorn Computers and foundry VLSI, and founded ARM. The sole requirement: a 32-bit low-power processor that could run on AA batteries. The result was the ARM610.

ARM610 RISC processor used in the first Apple Newton MessagePad. Source: Wikimedia Commons, photo by Chris Whytehead, CC BY-SA 3.0

Rushed Launch and Doomed Promise

The chip was ready, but software was collapsing. Sculley fixed the launch date for summer 1993. Engineers had built an entirely new OS and language; the codebase ballooned, thousands of bugs piled up, and at least one developer died by suicide under pressure. On August 1993 the first MessagePad shipped: 20 MHz ARM610, 640 KB RAM, four AAA batteries, priced at $699 . Initial sales of 50,000 units in three months seemed decent for a new category. Then disaster struck.

Newton's headline feature was "natural handwriting recognition," but the recognition engine had never been trained on enough real-world handwriting styles. Users wrote memos and got absurd gibberish — famously mocked by The Simpsons ("Beat up Martin" → "Eat up Martha") and the comic strip Doonesbury ("Catching on?" → "Egg freckles?"). Media ridicule was universal; pundits declared PDAs expensive, useless toys.

1993 Doonesbury strip satirizing Newton's handwriting failures. Illustration: AI-generated

Mirror Duel: The Opponent Who Taught Humans to Write for Machines

While Newton floundered, Jeff Hawkins launched the Palm Pilot in 1996. Hawkins saw Newton's fatal flaw: with 1990s mobile silicon, making a machine understand every person's idiosyncratic cursive was impossible. His solution: teach people to write letters the machine could recognize . Palm introduced Graffiti, a single-stroke alphabet where each letter is drawn without lifting the stylus. Users spent ~15 minutes learning it, but then achieved near-100% recognition. Graffiti demanded almost no compute — Palm ran on a weak 16 MHz Motorola chip, weighed half as much as Newton, and cost only $299 .

The market voted with wallets: Palm Pilot sold 1 million units in two years ; Newton struggled for five years and sold only ~ 250,000 units total . A day-one usable compromise crushed a still-half-baked pioneer.

Jobs' Guillotine

Newton's team didn't quit. They rewrote recognition with neural networks, dramatically improving accuracy. The 1997 MessagePad 2100 added a 162 MHz StrongARM chip, backlit screen, email and web browsing — essentially fulfilling the original vision. But the window had closed. In 1997 Steve Jobs returned to Apple. Facing bankruptcy, he drew a famous 2×2 grid (desktop/portable × pro/consumer) and decreed only four Mac models; everything else was cut. Newton was not in the grid.

On February 27, 1998 Apple officially ended the Newton line. To protesting staff and fans Jobs delivered the iconic line:

"God gave us ten styluses. Let's not invent another."

Ultimate Revenge via Resurrection

That same year ARM listed in London and on NASDAQ. Between 1998-1999, as Apple teetered near cash collapse with the original iMac, it sold its ARM stake in tranches, netting over $800 million — a 260× return on the $3M lifeline that had kept Newton alive. The deeper legacy took root in silicon.

2001: iPod shipped with two ARM processors inside.

2007: Jobs unveiled the first iPhone, its heart an ARM11 chip. The device that killed Palm was built on the architecture born from Newton's battery desperation.

From iPhone/iPad A-series to Mac M-series, the ARM architecture — conceived to let Newton run a few more minutes on AA cells — has now shipped over 350 billion units , becoming the energy core of the entire digital world.

Newton died because it ran fifteen years ahead of its time, shattered by reality. In Apple's 2019 40th-anniversary retrospective, Newton's name flashes on screen and is immediately crossed out — likely by an Apple Pencil.

Newton's late flagship MessagePad 2000 with StrongARM. Source: Wikimedia Commons, photo by Federigo Federighi, CC BY-SA 4.0