In November 1990, Apple wrote a check for three million dollars to fund twelve engineers working out of a small office in Cambridge, England. The engineers came from Acorn Computers, a British firm most Americans had never heard of. Their assignment was to design a microprocessor for a handheld computer that Apple had been trying to build for three years without success. The chip had to be powerful enough to recognize human handwriting in real time and efficient enough to run on batteries all day.

Nothing like it existed. The joint venture was called Advanced RISC Machines. The handheld was called the Newton. The processor they built would power the most mocked product Apple ever shipped, a device the press savaged, comedians buried, and Apple itself eventually killed.

That same processor architecture now sits inside virtually every smartphone on the planet. The Newton began with a man who was bored. Steve Sakoman was a Hewlett-Packard engineer who joined Apple in 1984 to work on a Macintosh laptop. He was sharp, restless, and impatient with anything that felt incremental.

When Steve Jobs was forced out of Apple in 1985 after losing a boardroom fight with CEO John Sculley, the company changed. Apple remained profitable, but the culture of radical ambition that Jobs had imposed faded into something more corporate, more cautious, more interested in quarterly earnings than in reinventing the world. Sakoman felt the shift in his bones. He had not come to Apple to make the Macintosh a few percent faster each year.

By 1987 he was sketching something no one in the mainstream computer industry was seriously pursuing: a computer small enough to hold in one hand that could understand what you wrote with a pen and convert it into digital text. The idea caught the attention of Jean-Louis Gassée, Apple’s head of product development. Gassée had written publicly in 1985 that computers would recognize handwriting within five years, a wildly optimistic prediction but one that revealed where his thinking was headed. More urgently, Gassée recognized that Sakoman was ready to walk out the door, so he authorized a skunk works operation, a small secret team working outside the normal corporate structure.

He gave it a name pulled from Apple’s own origin story. The original Apple logo depicted Isaac Newton sitting beneath an apple tree. Gassée called the project Newton. Sakoman recruited Steve Capps, co-author of the original Macintosh Finder, and set up shop in a separate facility.

Their first prototype, code-named Figaro, used three AT&T Hobbit processors, a hard drive, an active matrix LCD screen, and infrared networking. It was years ahead of anything else in computing, but it was also enormous, heavy, power-hungry, and ruinously expensive. The Hobbit processor turned out to be completely inadequate for what the Newton needed to do. Then the project lost its creator.

In March 1990, Sakoman left Apple to co-found a new company with Gassée, who had also departed. The Newton was orphaned, a project without its chief architect or its executive protector. It was saved by a meeting in the spring of 1990. Bill Atkinson, one of the original Macintosh team members and creator of MacPaint, organized a gathering of Apple veterans to decide the Newton’s fate.

Andy Hertzfeld was there. Susan Kare, who had designed the Macintosh’s original icons, was there. And crucially, John Sculley attended. When Sculley heard what the Newton could become, he committed Apple’s full support and appointed Larry Tesler to lead the project.

Tesler was one of the most respected computer scientists in the world, a veteran of Xerox PARC. His first problem was the processor. The Hobbit was dead, and nothing on the commercial market was both powerful enough and frugal enough with battery power. Tesler found the answer in England at Acorn Computers, which had developed the ARM architecture, a processor design built for efficiency rather than raw speed.

For a device that had to run on batteries all day, ARM was not just a good option, it was the only option. On November 12, 1990, Apple purchased a 43 percent stake in a new joint venture formed between Acorn, Apple, and VLSI Technology. Acorn contributed the twelve engineers. Apple contributed three million dollars.

The Newton finally had its brain, but it still needed a deadline, and John Sculley was about to give it one. Sculley was not an engineer but a marketing executive who had built his career selling soft drinks at PepsiCo. He had a genuine fascination with technology and an appetite for grand pronouncements. In 1987 he had commissioned concept videos depicting something he called the Knowledge Navigator, a futuristic tablet with built-in artificial intelligence.

When he learned about the Newton, he saw his vision taking physical form. He set two mandates: the Newton had to launch by April 2, 1992, and it had to cost no more than fifteen hundred dollars. In January 1992, Sculley took the stage at the Consumer Electronics Show in Las Vegas and introduced a phrase that would enter the permanent vocabulary of the technology industry. He called it a personal digital assistant, a PDA.

The press frenzy that followed pushed expectations to a height no first-generation product could reach. The Newton team understood they were no longer building a computer. They were building a promise made by a CEO to Wall Street, and the timeline belonged to his ambition, not to their engineering. What followed was a kind of development hell Silicon Valley knows well.

The Newton software had swelled to 750,000 lines of code. By May 1993, the project had accumulated over 3,700 documented bugs. An engineer named Walter Smith built NewtonScript from scratch, an entirely new programming language designed to do more with less memory and less processing power than anything else available. It was brilliant work and one more thing the team had to invent from zero while the clock ran out.

The handwriting recognition system, the centerpiece of the entire product, was the greatest source of dread. Apple had licensed recognition technology called Calligrapher from a programmer in Moscow. The software arrived with little documentation and even less predictability. It was supposed to learn each user’s unique handwriting over time.

In theory it was a breakthrough. In practice it was unfinished. The team worked fifteen to twenty-hour days. On December 12, 1992, a Newton engineer took his own life.

The team went back to work. Sculley’s April 1992 deadline passed without a product. At the summer Consumer Electronics Show in Chicago, Apple demonstrated the Newton publicly for the first time, but the devices on display were tethered to power sources and running carefully staged demonstrations. The first prototype famously failed to power on.

A second unit was rushed in and managed to work. Meanwhile, Sculley’s grip on Apple was slipping. On June 7, 1993, he was forced out as CEO and replaced by Michael Spindler. The man who had championed the Newton was gone, but the Newton was too far along to stop.

On August 1, 1993, at MacWorld Expo in Boston, the Apple Newton MessagePad went on sale. It cost $699, roughly sixteen hundred dollars in today’s money. It ran an ARM 610 processor at 20 megahertz, had 640 kilobytes of RAM, four megabytes of storage, and a monochrome touchscreen operated with a stylus. It could take notes, manage contacts, keep a calendar, perform calculations, send faxes, and recognize your handwriting.

That was the promise. The first 50,000 units sold by November 1993, respectable numbers for a product inventing a brand new category at a premium price. But the Newton was not judged as a respectable first attempt. It was judged against two years of Sculley’s promises, magazine covers, and breathless press previews.

The world had been told to expect a revolution. What arrived was a beta. The Calligrapher handwriting engine shipped before it was ready because the entire device shipped before it was ready. Users wrote “Meeting at noon” and received gibberish.

The mockery came fast. In August 1993, cartoonist Gary Trudeau devoted an entire week of his syndicated strip Doonesbury to savaging the Newton. In the most famous panel, a character writes “catching on” on the Newton screen, and the device interprets it as “egg freckles. ” The phrase became a permanent cultural punchline.

In 1995, The Simpsons piled on. Late night comedians made the Newton a running gag. Tech journalists who had breathlessly promoted the device now competed to write the cruelest reviews. Beta-quality software had been loaded into store demos, meaning the version potential customers actually tried was even worse than the one in the box.

Manufacturing defects compounded the humiliation. Faulty instruction cards shipped with the units, and rubber feet fell off devices and had to be repaired by hand. The cruel irony was that most of the criticism was both valid and wildly unfair at the same time. The Newton was attempting real-time handwriting recognition on a handheld computer in 1993.

No consumer product had ever tried anything like it. The technology was legitimately groundbreaking and legitimately unfinished. But somewhere in the development pipeline, a single engineer named Larry Yaeger had already begun rewriting the recognition engine from scratch. His code would eventually fix everything the comedians were mocking.

It would take years, and the world would have moved on before it mattered. The world in 1993 was simply not ready for what the Newton was selling. The internet barely existed for ordinary consumers. Wireless connectivity meant nothing to most people.

The idea of carrying a computer in your pocket was, for the vast majority of Americans, a solution to a problem they did not know they had. The Newton needed a world with wireless networks, always-on email, web browsers, and a culture of constant digital connectivity. That world was a decade away. Meanwhile, Apple was quietly funding its own competition.

In 1990, the same year Sculley committed to the Newton, he had authorized the creation of General Magic, a spin-off that recruited some of Apple’s most talented people. General Magic was building its own handheld platform, competing directly with the Newton for engineering talent, executive attention, and market position. Apple was splitting its resources between two visions of the same future, and neither would win. But one piece of the Newton story was quietly conquering the world without anyone noticing.

The ARM processor, the chip Apple had funded for three million dollars in a Cambridge office, was spreading. By the mid-nineties, ARM architecture had become the standard for low-power computing, and mobile phones across the globe were running on some variant of the ARM design. The chip born for a failed product was finding its way into millions of devices that had nothing to do with the Newton. While the press was busy writing the Newton’s obituary, the team inside Apple was quietly building a version of the device that would answer every criticism.

The MessagePad 100 dropped the price to $500. The MessagePad 110 refined the stylus and introduced a flip-over screen protector. The MessagePad 120 and 130 brought Newton OS 2. 0 and with it the most important change in the platform’s history.

Larry Yaeger had spent years rewriting the handwriting recognition system from the ground up. His engine, called Rosetta, took a fundamentally different approach. Where Calligrapher had tried to interpret natural cursive handwriting, one of the hardest unsolved problems in computer science, Rosetta focused on printed letters. It was backed by a dictionary of 78,000 words that helped the system predict what a user was likely writing and correct errors in real time.

The distinction sounds minor. It was not. Users who gave the updated Newton a fair chance found that the handwriting recognition worked reliably enough that taking notes on the MessagePad became faster and easier than pulling out a notepad and a pen. Yaeger’s code would later be ported to Mac OS X as an Inkwell feature, but almost nobody outside the Apple engineering community knows his name.

Then came the eMate 300 in March 1997, a rugged translucent green clamshell with a built-in keyboard designed for the K-12 education market. Its design was the real story. The curves, the translucent plastic, the playful, confident industrial design looked nothing like anything Apple was selling at the time. The eMate was the work of a young British designer named Jonathan Ive, and its aesthetic was a preview of the design language Ive would bring to the iMac exactly one year later.

The design philosophy that would rescue Apple’s entire brand identity made its first public appearance on a Newton product aimed at middle schoolers. But even as the Newton improved, the broader market had moved on. In 1996, Palm Computing released the Palm Pilot. It cost $299, was smaller, lighter, simpler, and it outsold the Newton almost immediately.

The man behind it, Jeff Hawkins, had carried a block of wood in his shirt pocket for months, pretending it was his imaginary calendar, testing whether people actually wanted a handheld device before he built one. The Palm Pilot’s handwriting system, called Graffiti, asked users to learn a simplified alphabet of gestures. You had to meet the machine halfway, and in return the machine gave you near-perfect accuracy. Palm sold over one million Palm Pilots within two years.

The Newton, across its entire lifespan of four and a half years, sold roughly a quarter of that number. The market wanted less, not more. The Palm Pilot proved that in the consumer electronics business, a modest product that works on day one beats a brilliant product that works on day one thousand. While the Newton was losing its market to Palm, Apple itself was dying.

Michael Spindler had replaced Sculley and accomplished little. Gil Amelio had replaced Spindler and accomplished less. In early 1997, Apple created Newton, Inc. , a wholly owned subsidiary with its own board of directors, its own corporate logo, and a plan to operate independently.

Roughly 130 employees transferred from Apple to Newton, Inc. New offices were leased in Santa Clara. Then, on September 4, 1997, Steve Jobs told the Newton, Inc. executives not to move.

He was back as interim CEO after Apple had acquired NeXT, the company Jobs had built during his years in exile. Jobs began cutting with surgical precision. His philosophy was simple and merciless. Focusing is about saying no.

On September 9, 1997, Jobs announced that Apple was pulling the Newton product line back in house. By early 1998, even the eMate was dead. On February 27, 1998, Apple officially discontinued the entire Newton platform. A product that had consumed six years, over one hundred million dollars, and the working lives of some of the most talented engineers in Silicon Valley was erased in a single quarter.

Jobs had his reasons. The sales numbers were dismal. The development costs had been staggering. But Jobs had a deeper objection, one that would define everything Apple built for the next twenty years.

He hated the stylus. God gave us ten styluses, he said. Let’s not invent another. He wanted a device you operated with your finger, not a pen.

There was also one more factor everyone at Apple understood and nobody discussed publicly. The Newton was John Sculley’s product, the signature creation of the man who had driven Steve Jobs out of his own company in 1985. John Sculley, reflecting years later, offered a quiet assessment. Newton was fifteen years too early, he said.

He was off by one year. The iPhone launched fourteen years after the original MessagePad. The Newton was dead, but the technology was not. The ARM processor architecture that Apple had funded for three million dollars continued its quiet conquest of the world.

When Apple introduced the iPhone in January 2007, it ran on an ARM chip. When the iPad launched in 2010, it ran on ARM. When Apple transitioned the entire Macintosh line away from Intel in 2020 with the M1 chip, that chip was built on ARM architecture. The line runs straight and unbroken from twelve engineers in a Cambridge office in 1990 to the silicon inside every Apple product sold today.

Advanced RISC Machines went public in 2023 at a valuation of 54. 5 billion dollars. The Newton MessagePad sold perhaps three hundred thousand units across five years. The ARM chips it paid for have shipped in the billions.

The Newton software concepts proved equally prophetic. A digital assistant that managed your contacts, calendar, and notes was the core proposition of every smartphone operating system that followed. The third-party application ecosystem established the principle that a handheld device should be a platform, not just a product, a principle that generated hundreds of billions of dollars when the App Store opened in 2008. Phil Schiller, Apple’s longtime senior vice president of marketing, put it directly.

The insight and work to create Newton and eMate laid much of the groundwork for today’s iPhone, iPad, and Apple’s entire product lineup. The Newton team had built the blueprint. Someone else got to use it. In 2016, at a keynote celebrating Apple’s fortieth anniversary, the company played a short video flashing its most important products across the screen.

The Apple II, the Macintosh, the iMac, the iPod, the iPhone, and there in the middle of the montage was the Newton. But unlike every other product in the video, the Newton’s onscreen text appeared scratched out, as if someone had tried to erase it from the record. Even Apple could not resist one final joke. But the fact that the Newton was there at all, listed among the machines that defined the company, told the real story.

The Newton did not fail because it was wrong. It failed because it was first.