The Palm Founders Didn't Come Home. They Were Carried Back on a Stretcher.

On a Monday in July 1998 — the first workday after the long holiday weekend — Donna Dubinsky sent an email to the entire Palm staff. She was leaving. So was Jeff Hawkins, the man who had invented the PalmPilot.⁷ The story most people remember stops roughly there: the founders walked out on a corporate parent that didn't understand them, built a better device across the street, and five years later came home in triumph to put the band back together. It is a clean, satisfying arc. It is also wrong at almost every joint.

The official version is that Palm and its founders reunited in a strategic homecoming. The real one, written in SEC filings and a Stanford case study, is that Handspring was a failing company and Palm bought it to get a product its own leadership had refused to build. The reunion was a rescue with a press release stapled to the front.

Three people walked out, not two

Start with the legend's smallest lie, because it tells you how the rest got smoothed over. The departure is almost always credited to a duo — Hawkins the visionary, Dubinsky the operator. But three founders left: Hawkins, Dubinsky, and Ed Colligan, who followed a few weeks after the first resignation email went out.⁷⁵ Colligan gets quietly dropped from the retelling because a two-person founding story is tidier — a genius and his lieutenant, not a committee. That instinct to compress, to round the inconvenient detail off, is the same instinct that later turned a distress sale into a homecoming. The context they left was itself layered: Palm had been incorporated in 1992 as Palm Computing, swallowed by U.S. Robotics in 1995, then swept up again when 3Com bought U.S. Robotics in 1997.⁴ By 1998 the founders were three acquisitions deep inside a networking conglomerate, unhappy with where 3Com was steering their handheld.⁵ So they left and built Handspring — and, in a detail the rivalry framing always misses, they licensed the very Palm OS they'd just walked away from, signing the agreement with 3Com on the day they announced the new company.⁷

The day Palm was worth more than the company that owned it

Before the rescue, there was the mania. On March 1, 2000, 3Com floated a slice of Palm on the public market, and the market lost its mind. Shares priced at $38 spiked to an intraday high of $165 before closing at $95, valuing Palm at $53 billion — more than 3Com itself, the parent that still owned the overwhelming majority of it, at $28 billion.⁶¹⁰ This was one of the strangest pricing dislocations of the dot-com peak: 3Com held roughly 94.3% of Palm's stock at the IPO, so on paper its stake alone was worth far more than the entire company carrying it.³ The IPO is often described as Palm's clean break from 3Com. It wasn't — it was a partial float. Full separation came only later that year, via a stock dividend to 3Com shareholders distributed in late July 2000.³ Independence took months, not a single bell-ringing morning. Remember the $53 billion. It is the high-water mark against which everything that follows should be measured.

By 2003, the rivals were both bleeding — one fatally

Three years after that euphoria, the handheld market had cratered and both companies were struggling. But Handspring's situation was worse: by early 2003 it was, in the Stanford case's flat language, in a 'dire financial situation.'⁵ Crucially, its board faced a fork — and it was not 'merge with Palm or stay independent and thriving.' It was 'merge with Palm, or take a private financing to survive a little longer.' Handspring pursued both a PIPE — a private placement of stock, the move companies make when public markets won't have them on decent terms — and the merger, simultaneously, with the board 'fairly evenly split' between them.⁵ That detail is the whole story. A board split down the middle between a rescue financing and a sale is not a board choosing a strategic future. It is a board triaging a patient. The merger won not because reuniting with Palm was a grand vision, but because it was the option that didn't require finding new money.

The merger closed on October 29, 2003. Handspring shareholders received 0.09 of a Palm share for each Handspring share — a fraction that quietly marks the distance traveled: the founders exchanged each share of the company they'd built for less than a tenth of a share in the one they'd left.¹²¹² And here is the mechanism that explains why Palm did the deal at all. It wasn't sentiment, and it wasn't engineering muscle — Palm had plenty of that. It was a product gap. Handspring had built the Treo, the smartphone that fused the organizer with the phone, and it was the future of the category. Palm's own acquisition announcement named "Handspring's highly regarded Treo product line and carrier relationships" as a primary strategic benefit of the deal — the acquirer was paying, in its own words, for the one thing its own roadmap had not delivered.⁹ The acquirer was paying for the one thing its own roadmap had failed to deliver.

The reunion that fragmented the company further

If this were truly a reunification, the merged entity would have emerged whole. It did the opposite. On the very same day the Handspring deal closed, Palm spun off its operating-system division as a separate public company, PalmSource.¹⁴ And the merged hardware company wasn't even called Palm. It was renamed palmOne.⁸ So the moment that supposedly put the original Palm back together actually split it into two public companies and stripped the founding brand off the surviving one. The name 'Palm, Inc.' did not return until July 2005 — and only because palmOne had to buy back PalmSource's share of its own trademark for $30 million.⁸ The team that left in 1998 spent two years after their 'homecoming' renting the right to be called Palm again.

Isn't a rescue that gets you the Treo still a smart deal?

The fair objection is that none of this makes the merger a bad decision. Distress on the seller's side is exactly when a buyer gets a bargain, and Palm got the Treo — the product line that defined its next several years — for a fraction of a share apiece.² If you can buy the capability you lack at a fire-sale price, calling it a 'rescue' rather than a 'strategy' is just semantics. That's a real point, and it concedes the thesis rather than refuting it. The deal can be both shrewd and a confession. The problem was never the price; it was the narrative the price got dressed in. By telling itself a reunion story, Palm got to skip the harder question the deal actually posed: why did a company that had invented the handheld need to buy its smartphone future from a startup three of its own founders had left to build? The Treo was the symptom of the gap, not its cure. A homecoming framing soothes; it does not diagnose. And the fragmentation that followed — palmOne here, PalmSource there, the brand itself in escrow for two years — is what a company does when it is reacting to events, not authoring them.

The tidy version of this story has the founders walking out a side door in 1998 and back through the front in 2003, the family made whole. The filings tell a colder one. They left because the parent wouldn't build what they wanted; they came back because the company they built couldn't pay its bills; and the entity that absorbed them promptly split in two and gave up the family name for two years. The lesson isn't that the merger failed — the Treo was real and it mattered. It's that the most expensive word in the whole affair was 'reunion,' because it let everyone involved feel like the wanderers had come home, when what had actually happened is that a company bought back, at the worst possible moment, the future it had once let walk out the door.