Part IThe Story
The Man Named for the First Selfie
On the morning of March 2, 2017, someone took a selfie at the New York Stock Exchange. The wire photographers caught it. The Australian model Miranda Kerr, Evan Spiegel's partner, held up her phone at the opening bell and grinned into it with her friends, and the image went out on the Associated Press feed with a caption that did the job a caption is supposed to do. Moments earlier, the two co-founders of Snap Inc. had rung the bell. Within hours the stock had opened at $24 against a $17 offering price. By the close it sat at $24.48, and a company that had lost $514 million the year before was worth roughly $28 billion, about the same as CBS or Target. Spiegel was 26. The other co-founder was 28.
Nine years later, sitting with the podcaster David Senra in April 2026, Spiegel told a story he plainly enjoyed. When he and his co-founder set out "to reinvent the camera," he said, they studied the history of photography (Edwin Land, Polaroid, the whole lineage) and in the course of that study turned up a coincidence. "The first selfie ever was taken by a guy named Robert Cornelius," Spiegel said, "and my co-founder, Bobby, his name is Robert Cornelius Murphy."
That is the sort of thing a company tells about itself once it has decided to call itself a camera company. It is also, read a little differently, very nearly a joke at Bobby Murphy's expense. He is the namesake of the man who first turned the lens around, and he has spent fifteen years, by every available account, keeping himself out of the frame. Murphy has given a handful of interviews. Business Insider, writing the week of the IPO, could assemble his public life into a slideshow, and part of the slideshow was a photograph of his high school. At the moment of the offering he was worth an estimated $5.4 billion. He held, by the terms of the S-1, exactly as many shares as Spiegel (210.97 million) and exactly as much voting power, and under the company's charter, if one of the two men died, the other could control the whole enterprise alone. The Los Angeles Times, trying to explain this to investors who were being asked to buy stock with no votes attached, described him as "relative to Spiegel ... an unknown entity publicly."
So the question this profile keeps returning to is a simple one. How does a man end up holding half the keys to one of the most consequential consumer technology companies of his generation while almost nobody outside the building could pick him out of a lineup? And what, exactly, has he been doing in there?
The short answer is that he has been writing the code, and paying the server bills, and staying late. The long answer takes in a Catholic school in Berkeley, a fraternity house at Stanford, a summer house on Toyopa Drive in the Palisades, a $3 billion offer, a $157.5 million settlement, a decade of sunglasses nobody wanted, and a promise (things will disappear) that turned out to be very hard to keep.
By the Numbers
The Quiet Half of Snap
$485,000Lightspeed's first check into Snapchat, 2012
50%Share of his Revel Systems paycheck Murphy spent on Snapchat's server bills
$3BFacebook's cash offer, declined in November 2013, when Snapchat had six employees
$157.5MSettlement paid to ousted co-founder Reggie Brown
210.97MShares held by Murphy at the IPO, identical to Spiegel's stake
$1Murphy's base salary as CTO
95.8%Combined voting power held by Spiegel and Murphy
Who Stays to Clean Up
Spiegel threw parties as a teenager in Los Angeles, big enough that he hired a security team to police them. The head of that team, according to the Los Angeles Times, gave him a lesson Spiegel carried into adulthood: your true friends are the ones who stay behind afterward to help you clean up.
At Stanford, Spiegel kept throwing parties. And Bobby Murphy, a fraternity brother in Kappa Sigma who lived across the hall, kept staying behind. Late at night, after the house had emptied, he tidied up.
It is a small anecdote, the kind that turns up in the seventh paragraph of a newspaper story and is never mentioned again. It carries more weight than it looks like it should. Most of what anyone can say about Murphy's character with confidence comes down to some version of it: he is the one still in the room.
He came to Stanford from the other side of the Bay, and from a different sort of family than Spiegel's. Robert Cornelius Murphy grew up in Berkeley. His father, Cornelius Murphy Jr., was Irish-American. His mother, Rosie L. Go, had emigrated from the Philippines, and both parents held government jobs. Philippine outlets, which followed him with a proprietary pride the American press never matched, liked to note that she later worked for the city of Berkeley. He was educated by the Church: grade school at the School of the Madeleine, high school at Saint Mary's College High School, both Catholic, both private, both in the city where he was born. Asked later to explain himself and Spiegel, he gave Forbes a line that sounds like a confession and functions as a mission statement: "We weren't cool. So we tried to build things to be cool."
Spiegel was cool, or was working hard at it. He was the son of a prominent Los Angeles lawyer, a product design major who would later be described in print as "jet-setting, supermodel-dating, trend-making, once-in-a-generation." He was two years younger and somehow always in charge. He read about Edwin Land at an age when, as Senra put it, almost nobody that age knows who Edwin Land is. Murphy studied mathematical and computational science. That pairing (the designer who wanted to be Steve Jobs, the mathematician who wanted to be useful) would be the founding fact of everything that followed.
They tried things together. Murphy, by the Times's account, invited Spiegel to help on a social networking service. Spiegel, in turn, sought Murphy out to build a college admissions tool. The second idea had a name, Future Freshman, and a corporate shell, Future Freshman LLC, and in the summer of 2010 (Murphy just finished with his senior year, Spiegel with his sophomore) they split the equity and went to work on helping high school students get into college. It didn't work. Nobody much wanted it. In March 2011 they gave up.
Note what survived the failure: the LLC, the habit of splitting equity between the two of them, and the friendship. All three would matter enormously, and one of them would cost $157.5 million.
The Summer on Toyopa Drive
The man who had the idea is the one who has vanished most completely from the story.
Reggie Brown was another Kappa Sigma brother at Stanford and, by his own description, close friends with Spiegel. In April 2011, a month after Future Freshman died, Brown came to Spiegel with a notion. Billy Gallagher, another fraternity brother who later wrote How to Turn Down a Billion Dollars: The Snapchat Story, reconstructed the moment in vivid and unflattering detail: Brown wanted a way to send photos that would vanish, because people would be more willing to send certain kinds of photos if those photos didn't last. Spiegel had business experience, of a kind. Neither of them could build it. So they went to get Bobby.
That summer the three young men lived at Spiegel's father's house on Toyopa Drive in Los Angeles. Murphy, by the account Forbes later published, worked eighteen-hour days to get a working prototype running. Spiegel handled design. Brown, according to a source who spoke to TechCrunch in 2013, went out a great deal. On July 8, 2011, the app launched on iOS as Picaboo. On July 13, Brown emailed Spiegel a draft press release that included the sentence, "Picaboo is a licensed product of the Toyopa Group, LLC."
The Toyopa Group, LLC was Future Freshman, LLC with a new name. Brown said in his deposition that they had discussed the rename together, "but I didn't understand it as the same company." Here was the trapdoor. To Brown, Toyopa was the new venture, and he was in it. To Spiegel and Murphy, Toyopa was their old venture with a new product inside it, and the equity had been settled back when Brown had nothing to do with anything.
Murphy's own deposition testimony on this point is the most revealing thing he has said on the record about the founding, and it is almost entirely about what he did not know. Asked whether he thought Brown understood he had no equity, Murphy said: "I don't know what he believed. All I know is that, again, he was invited to join us that summer, do some work." Spiegel would go further and describe Brown as essentially an unpaid intern. Murphy, by the court record's summary, said he had not fully understood what Brown's role was supposed to be, and then characterized it as an internship too.
The sequence that ended Brown's involvement has the shape of a farce. One night that summer the three had a drunken argument about Brown's contributions. The next morning they talked about it soberly, which by Brown's account meant "talking about switching me out for a different marketing person." On August 11, 2011, Brown filed a patent application titled "Timed, Non Permanent Picture Messages for Smart Phone Devices." It listed his home address as the sole contact and named the co-inventors in this order: Murphy, Brown, Spiegel.
On August 16 the three argued on the phone. Spiegel, according to the reporting, was offended that his name was listed last. He hung up. The passwords changed. Brown was out.
A month later, in September 2011, Picaboo relaunched as Snapchat. The team, according to Wikipedia's reconstruction from contemporary accounts, deliberately focused on usability and technical execution over branding, with one exception. They kept the ghost Brown had drawn, a mascot he'd named "Ghostface Chillah" after the Wu-Tang Clan rapper. The logo of the company that would eventually pay him $157.5 million is still, in essence, his doodle.
Brown sued in February 2013. Snapchat called the suit "devoid of merit" and a "shakedown." In court documents released that July, Spiegel acknowledged that Brown had "came up with the idea for disappearing picture messages." In a deposition video later published by Business Insider, Spiegel paused for thirty-seven seconds before describing his regrets about the friend he had pushed out. On September 9, 2014, the company announced a settlement, terms undisclosed, along with a statement crediting Brown as one of the app's original authors. The number surfaced on February 2, 2017, in the S-1: $157.5 million. Brown then did what the app had been designed to let pictures do. The Los Angeles Times ran a story under the headline "The guy who came up with the idea for Snapchat got $158 million and vanished from public life."
There is no evidence that Murphy drove the ouster. There is equally no evidence that he resisted it. He was the coder in a three-person company whose two non-coders had fallen out, and he went with the one who stayed to clean up with him, or rather the one he had stayed to clean up after. What the episode reveals about Murphy is something sharper than loyalty: when he picked a side, he stayed picked.
Half a Paycheck
The myth of the overnight app does not survive contact with the bank statements.
Murphy graduated in 2010. Spiegel still had his senior year ahead of him, which he spent, by Business Insider's account, promoting the app between classes. Snapchat had a product and a few users and no money. So Murphy went to work. He took a job as a software engineer at Revel Systems, a San Francisco company that made iPad-based point-of-sale registers for restaurants, about as far from the poetics of disappearing photographs as software gets. Toward the end of 2012, as usage began to climb, the cost of storing and relaying all those images began to climb with it. Murphy paid the server bills. He spent half of every Revel paycheck keeping Snapchat online until venture money arrived.
It is worth sitting with the arithmetic of that decision. A recent Stanford math graduate, living in one of the most expensive cities in the country, with a startup that had already failed once in a different form, chose to fund the infrastructure of his friend's college project out of his salary. Not a convertible note. Not a loan. Just the money. No one has published a word Murphy said about it at the time, which is itself characteristic.
The growth, when it came, came from below, and from the south. Spiegel's mother told his cousin about the app; the cousin attended a high school in Los Angeles where students' school-issued iPads blocked Facebook. Snapchat became the way they passed notes. Snapchat was processing about 25 images per second by May 12, 2012. By October 29 of that year, when the Android version launched, users were sending 20 million photos a day, about 231 per second, more than Instagram had been processing at a comparable stage.
The venture capitalists found it the way venture capitalists sometimes do, through their children. Barry Eggers, a managing director at Lightspeed Venture Partners, learned from his teenage daughter that the three most popular apps among her friends were Angry Birds, Instagram, and Snapchat. His partner Jeremy Liew went after the company and agreed to invest $485,000. Across its rounds Lightspeed would put in $8.1 million in total. At the IPO, that stake was worth $2.3 billion.
Not every investor answered. On November 2, 2012, Murphy sent an email to Chris Sacca, the early backer of Twitter and Uber, after the two had met. It is one of the very few pieces of Murphy's writing from that era in public circulation, and it surfaced only because Sacca posted it on Twitter the day after the IPO, as a kind of gracious self-flagellation:
"Hey Chris, I really enjoyed the chat last night – really respect your worldview. It'd be awesome if you could come by for lunch or hang out some afternoon. I'd love to get your opinion on LA vs the Bay and building out a company especially as it related to Snapchat now and in the next few months. Let me know if you have any free time in the next week."
Sacca never replied. Read the email again, though, for what it shows about the sender. It is courteous to the point of diffidence. It asks for opinions rather than money. It frames the company's central strategic question (Los Angeles or the Bay Area) as a thing to talk over at lunch. It is the email of a man who does not much enjoy selling, doing it anyway because someone has to. And note that the question he raised, LA versus the Bay, is the one Snapchat answered against every piece of Silicon Valley conventional wisdom, by staying south.
By February 8, 2013, Benchmark's Mitch Lasky had led a $13.5 million Series A at a post-money valuation between $60 million and $70 million, and users were sending 60 million snaps a day. Murphy had been on the board, and CTO, since May 2012. The server bills were no longer his problem.
Three Billion Dollars and Six Copies of Sun Tzu
Mark Zuckerberg first tried to buy Snapchat in 2012, according to Gallagher, for $60 million. When Spiegel declined, Zuckerberg showed him that Facebook had already built a clone. It was called Poke, and it launched on iOS on December 21, 2012, an app for sending expiring text, photos, and video. It was widely understood as a direct shot at a startup that had, at that point, raised less than half a million dollars.
Poke failed. Facebook pulled it from the App Store on May 9, 2014.
In between, the price went up by a factor of fifty. On November 13, 2013, the Wall Street Journal reported that Snapchat had spurned a $3 billion cash offer from Facebook. At the time, according to the Guardian's later reckoning, the company had six employees. Om Malik reported a day later that Google had offered $4 billion, also refused. Business Insider's summary of the Forbes reporting gives the scene a particular texture: Spiegel and Murphy met with Zuckerberg, who outlined how he planned to crush them; they declined; and then they bought copies of Sun Tzu's The Art of War for their six employees.
The official story is one of unwavering resolve. Gallagher's book complicates it. Spiegel, he reported, "seriously considered accepting" the $3 billion and at times favored the deal in his discussions with Murphy, before deciding that he couldn't spend his life as a Facebook employee answering to someone else. The book's sourcing is thin on Murphy's side of those conversations. We know Spiegel wavered. We know the two of them had the conversation. We know the deal didn't happen. What we don't know (and this is the recurring condition of reporting on Bobby Murphy) is what he said.
Here is what can be inferred. Spiegel's stated reason for declining was about autonomy: answering to someone else. Murphy would have been, as CTO and as the author of most of the code, the person whose working life would have changed least under Facebook ownership; engineers who sell their companies generally keep engineering. The Guardian's figure for what each founder would have received is $750 million. And yet the company did not sell. Whatever role Murphy played, he was one of two people who could have pushed it over the line, and it did not go over the line.
They had, it should be said, already taken some money off the table. When Institutional Venture Partners led an $80 million Series B in June 2013, $20 million of it was a secondary sale; Valleywag reported at the time that the co-founders had unloaded personal stock for $20 million. It's a detail that tends to go missing from the mythology of the refusal. Saying no to $3 billion is somewhat easier for a 25-year-old who has recently said yes to his share of $20 million. It was still, in retrospect, the right call. When the company listed, the Guardian observed drily that each founder now held "a paper fortune worth in excess of Zuckerberg's offer."
The refusal had a sequel. On August 2, 2016, Instagram launched Instagram Stories, and its CEO, Kevin Systrom, said out loud that the format was copied from Snapchat. Facebook started testing Snapchat-style ephemeral photo messaging in Ireland that October. The company that couldn't buy Snapchat and couldn't out-build it decided to absorb it feature by feature. This has been the weather in which Murphy has done his work ever since.
The Monk of Venice
David Kravitz was Snapchat's first hire. In 2014 he gave Forbes the description of Bobby Murphy that every subsequent profile has quoted, because it is very nearly the only one.
— David Kravitz, Snapchat's first hire, to Forbes, 2014I'd describe him almost like a monk. I don't think I've ever seen him upset.
The other descriptions, when they surface, rhyme. Garrett Gee, who sold his barcode startup Scan Inc. to Snapchat in 2014 (the price, $50 million, became public only because it turned up in emails stolen in the Sony Pictures hack), told the Los Angeles Times that Murphy was "Very chill. Very friendly. Real good guy." The handful of people the Times could find who were willing to discuss him called him "quiet, unpretentious and stoic." Business Insider's sources said smart, friendly, quiet. A Chinese encyclopedia entry, compiling the English-language record, settled on the phrase "the fair nice guy." The Asian American nonprofit Gold House, honoring him, called him "affable, even-keeled."
These are not the words people use about founders who become billionaires before they turn 27. They are the words people use about a well-liked senior engineer.
In 2013 Snapchat moved into its first offices in Venice, and Murphy bought a house nearby: a two-bedroom, 2,250-square-foot place at 43 Brookes Avenue, designed by the architect Larry Scarpa and the artist Randy West, for $2.1 million. It was, for the second-youngest billionaire in the world (Forbes put him at No. 2 in 2015, behind only Spiegel, at $1.5 billion, ranked 1,250th on the planet), a strikingly modest purchase. The houses would get bigger later. In 2018 he paid the full $19.5 million asking price for the former Pacific Palisades estate of Eddie Albert, the star of Green Acres. In 2019 he bought what the Los Angeles Business Journal counted as his sixth multimillion-dollar home, for $6 million. In 2021 he spent $14.5 million on a duplex on the Strand in Manhattan Beach. The real estate is the one part of his life that leaves a public trail, and it reads mostly as a man who likes the Southern California coast and does not much like to talk about it.
Spiegel, meanwhile, became one of the most photographed executives in the country. He dated, and married, Miranda Kerr. Gallagher's book reports that he brought Taylor Swift as his date to a Snapchat New Year's Eve party one year and that the two of them watched the bash from an elevated catwalk, and that the company's policy at those parties evolved from requesting "no photos" to confiscating guests' phones. In May 2014 the old Stanford fraternity emails leaked, and TechCrunch ran the headline "Confirmed: Snapchat's Evan Spiegel Is Kind Of An Ass." A former employee later alleged in a lawsuit that Spiegel had called the app "only for rich people." Spiegel absorbed the scrutiny. Murphy did not attract any.
When Time named its 100 most influential people in April 2014, Murphy was on the list, alongside Spiegel. Philippine outlets ran the news with headlines that put "Fil-Am" before his name; an American could read every major U.S. business profile of Snapchat from that era and never learn that one of its founders was half Filipino. He made his first television appearance on The Colbert Report, sitting next to Spiegel. The Times noted that the two of them continued to sit side by side "for big moments." That was the arrangement. Spiegel talked. Murphy sat next to him.
The investment community, asked to make sense of it, reached for precedent. Steve Jobs and Steve Wozniak. Jerry Yang and David Filo. Ryan Jacob, a fund manager, put it to the Times in terms that sound almost envious:
— Ryan Jacob, Jacob Asset Management, to the Los Angeles Times, 2017It's often a case where the other person appreciates the fact that they can be unknown, stay behind the scenes and walk around without being noticed. Some can't believe it, but believe me, there's a lot of people who enjoy that.
The comparisons flatter and mislead in equal measure. Wozniak left Apple's operations early. Filo never held a governance position equivalent to his partner's. Murphy has held identical equity and identical votes for the entire life of the company. He is not the quiet one who stepped aside. He is the quiet one who did not.
The Camera, Not the Content
In 2014, at a Google Cloud Platform Live event, Murphy gave what remains one of the clearest statements of what he thought he was building, and it was not "disappearing photos."
"Evan and I got started Snapchat in the summer of 2011," he said, "basically understood that visual content that was the most engaging, interesting form of content there was. We wanted to create a way that would enable that to be a means of communication, rather than a piece of content around which communication actually happens."
Read that last clause twice. On Facebook, the photo was an object: you posted it, and then people commented on it. The conversation happened around the picture. Murphy's idea was that the picture should be the conversation, a sentence spoken in images rather than a thing displayed. Ephemerality, in this framing, was not the product. It was the grammar. Spoken sentences don't stick around, either.
Spiegel had been saying something similar in his more aphoristic way. In the company's first blog post, on May 9, 2012, he wrote:
— Evan Spiegel, Snapchat's first blog post, May 9, 2012Snapchat isn't about capturing the traditional Kodak moment. It's about communicating with the full range of human emotion—not just what appears to be pretty or perfect.
In a January 27, 2014, keynote he put it more austerely: "Snapchat discards content to focus on the feeling that content brings to you, not the way content looks." That speech, as the scholar Kimberly Hall later pointed out in the journal Amodern, was preserved on the company blog, a durable archive maintained by a company whose brand was impermanence. Hold on to that irony. It will come back.
The evidence suggests that Murphy was the one who saw past the original premise first. In the company's pre-IPO roadshow video, which ran Spiegel on screen about three times as long as Murphy, Spiegel went out of his way to credit his co-founder with the realization that Snapchat could become more than a photo-messaging app. The moment Murphy himself chose to describe in public came from a desert music festival. In 2014 he and Spiegel were at Electric Daisy Carnival in Las Vegas when the company launched the "Our Story" feature, which let users at a single event contribute snaps to one shared, public story. "Instead," Murphy said later, describing what they'd expected versus what they got, "we saw the most amazing content covering an event that we've ever seen." The company realized it could draw geofences around events and places and let the crowd fill them with footage. "It was just utterly unbelievable how much really cool and creative and fresh content was being submitted to this live story."
What emerges from that anecdote is a particular kind of technical leader. Not the architect who designs a system in the abstract and then defends it, but the observer who goes to the place where the system is being used, stands in the crowd, and notices what people are actually doing. Stories had launched on October 3, 2013, as a 24-hour chronological feed. By June 2014, Stories had overtaken person-to-person snaps as the most-used function on the service, with more than a billion views a day. The app that began as a way to send one photo to one person had become, through a series of small observed adjustments, a broadcast medium.
Then came the faces. In September 2015 Snapchat acquired Looksery, a startup with facial-tracking technology, and on September 16 launched Lenses: initially just seven augmented-reality filters, built in-house. One of them made users vomit rainbows. It is easy to laugh, and many people did, and the joke obscured what was happening, which was that a messaging company had just shipped real-time computer vision to a consumer audience at a scale nobody else had attempted. In October 2025, at the company's Lens Fest developer conference, Murphy and Spiegel would mark a decade of AR lenses going back to that rainbow.
By 2021, speaking to the Hindustan Times, which called him "in many ways the company's secret weapon when it comes to developing new features," Murphy could describe augmented reality in language that sounds like the Google Cloud quote grown up: "I think it represents a capacity to render or visualise software experiences in the way humans normally see the world and kind of imagine things in the world." The through-line from 2011 is clean. First the picture becomes the sentence. Then the camera becomes the screen. Spiegel's line for this, repeated for a decade, is that Snap is a camera company. Murphy is the person who has to make the camera do it.
Equal Votes
The most unusual thing about Snap's initial public offering was not its size (the largest U.S. listing since Alibaba in 2014) or its valuation or the fact that the company had never turned a profit. It was that the shares sold to the public carried no votes at all.
Plenty of technology companies had dual-class structures by 2017; Google and Facebook had normalized the idea that founders could keep control after selling most of the economics. Snap went one step further. Outside investors received Class A stock with zero voting rights. The founders' holdings, a mix of Class A, Class B, and Class C shares (97.16 million, 5.86 million, and 107.94 million respectively, for each of them, per CNBC), gave the two men more than 70 percent of the vote between them at listing. By 2025, according to the company's own filings as summarized on Wikipedia, their combined voting power stood at 95.8 percent.
British governance purists were appalled. Prospect magazine noted that the offering could not have happened in London, where institutional investors simply refuse non-voting shares. Martin Sorrell, then running WPP, defended it: "Perhaps surprisingly, corporate structures that seem to offend customary good corporate governance may deliver better long-term results." Investors bought anyway. The stock rose 44 percent on the first day.
The governance debate mostly concerned Spiegel, since Spiegel was the one making decisions in public. But the structure was designed around two people, not one, and its most striking feature is its symmetry. Spiegel and Murphy held the same number of shares. They held equal votes. The charter provided that if one died, control passed to the other. Spiegel received an additional 37.4 million shares as a bonus on completion of the IPO, worth roughly $750 million at the time, and that tilted the economic stake in his direction. It did not change the founding arrangement, which was that the company belonged to both of them or to neither.
It's worth noticing what kind of trust that represents. A CEO who wants unchecked control writes a charter that gives it to him. Spiegel wrote one that gave the decisive votes to two people, one of whom almost never spoke publicly and could, in theory, outvote nobody on his own but could block anything requiring both. For a founder whose early history included changing the passwords on a partner over the order of names on a patent application, this was a remarkable thing to do. The only plausible explanation is that he had already tested this particular partnership, at Future Freshman and on Toyopa Drive and on Revel's payroll, and it had held every time.
Then the market tested it. On May 10, 2017, Snap reported its first quarterly earnings as a public company: a $2.2 billion loss, revenue and user growth below expectations. The stock fell about 23 percent after hours, to $17.66, a few cents above the IPO price. CNBC calculated that Murphy had taken a $1.1 billion haircut in less than a day. Any $5 drop in the share price now moved each founder's fortune by more than $1 billion. Nobody reported Murphy saying anything about it.
In February 2017, before all this, he and Spiegel had pledged up to 13 million shares of Class A stock over fifteen to twenty years to a new Snap Foundation, supporting arts, education, and youth nonprofits. Murphy's base salary as CTO, the Los Angeles Business Journal reported, is $1. He is not paid for his board seat. In 2023 his total compensation from the company came to $801,334. His fortune rises and falls with the stock, which is to say with the work.
The Glasses That Keep Coming Back
The hardware started in secret. In early 2014 Snapchat quietly bought Vergence Labs, a small Los Angeles startup that had made camera-equipped eyewear called Epiphany, for $15 million in cash and stock. The world learned about it only that December, from those same Sony Pictures emails. Business Insider later described the acquisition as the genesis of everything that followed: Snapchat "secretly bought a struggling startup, then bet the future on it."
The bet surfaced on September 23, 2016, when Snapchat Inc. renamed itself Snap Inc. and unveiled Spectacles: $130 sunglasses with a camera that recorded ten-second clips in a circular format. Spiegel called them "a toy." They were sold out of yellow vending machines that appeared in unannounced locations, and for a few weeks the lines were long and the coverage was rapturous. The company had opened a research and development office in Shenzhen in December 2016, focused on the glasses. Spectacles went on sale online on February 20, 2017.
Then nobody wanted them. Snap sold about 220,000 pairs of the first version and later took losses of more than $40 million on unsold inventory. A second version arrived in April 2018; Murphy's Snap Labs team was behind it. On April 28, 2022, the company announced a $230 pocket-sized selfie drone called Pixy. That August the Wall Street Journal reported that development had been halted. The same month, Snap laid off 20 percent of its 6,400-person workforce, cuts that fell heaviest on the hardware division and on Zenly, the location-sharing app it had bought in 2017 for somewhere between $250 million and $350 million and whose concepts it had already folded into Snap Map. Original scripted programming, which Snap had poured money into since 2017, was ended entirely.
A reasonable company would have quit hardware in 2022. Snap kept going, and the person who kept it going in public, as much as anyone did, was Murphy. Every year at Lens Fest he and Spiegel sat for an unscripted Q&A with the developers who build augmented-reality lenses on Snap's platform, a practice the immersive-media podcaster Kent Bye called "an amazing tradition. You don't see this a lot from a lot of the other companies." Murphy, who doesn't do press, does this. He goes to Augmented World Expo, the industry's main gathering. In 2021 he gave a talk there called "Reinventing the Camera"; in 2022 he spoke at MIT Technology Review's EmTech on "Enhancing Our World Through AR"; in June 2024 his AWE session was titled "Advancing Snap AR Through Generative AI." This is his circuit. It is not the circuit of a man trying to be famous. It is the circuit of a man trying to persuade a few thousand developers that the platform will still be there in five years.
In January 2026 Snap spun its augmented-reality glasses into a subsidiary, Specs Inc. In June 2026, the company announced a new generation of AR glasses called Specs. The BBC headline read: "Snap unveils new Specs smart glasses after previous disasters." At AWE that month, after Spiegel's morning keynote, Murphy took the main stage to explain the software. The numbers he and his colleagues offered were engineer's numbers, specific and unglamorous. An optical system 35 percent more transparent, so wearers could make natural eye contact. Displays with a 30 percent larger active area in a design 40 percent lighter. Motion-to-photon latency cut nearly in half, to 7 milliseconds, which is how long it takes for digital content to keep up with the turn of a head well enough to seem anchored to the physical world. Ten Snap OS updates and more than forty new features and APIs in eighteen months. A migration agent to translate Unity projects into Lens Studio. A native development kit for C and C++ libraries, which Mapbox had used to port its entire navigation engine to Specs in under two hours.
The camera on your face, Spiegel tells interviewers, will eventually replace the screen in your pocket. Ten years and at least one write-down in, it hasn't yet. Murphy keeps showing up to the developer conferences anyway.
The Price of Memory
The promise was that things would disappear. It was never entirely true.
On December 31, 2013, hackers published the usernames and phone numbers of about 4.6 million Snapchat users, exploiting a vulnerability that an Australian security firm, Gibson Security, had disclosed to the company on August 27 and then published in frustration on Christmas Day. Snapchat apologized a week after the breach. In May 2014 the company settled Federal Trade Commission charges that it had misled users: messages it said disappeared could in fact be retrieved, and the app had collected address books and transmitted locations without notice. Snapchat was not fined. It agreed to have its privacy claims monitored by an independent party for twenty years. Its privacy page was rewritten to concede that the company "can't guarantee that messages will be deleted within a specific timeframe," and a company blog post noted helpfully that "with the right forensic tools, it's sometimes possible to retrieve data after it has been deleted."
For a CTO, this is the hardest category of failure: the gap between what the product says it does and what the system actually does. The founding promise was an engineering promise. It was also, it turned out, partly a marketing one.
Then the company began, by degrees, to keep things. In December 2013 users were allowed to replay one snap a day. Stories lasted twenty-four hours. In July 2016 Snapchat launched Memories, a searchable private archive of saved snaps, with a PIN-locked folder called My Eyes Only. A company built on forgetting had become a place where people stored years of their lives. In September 2025 Snap announced that users storing more than five gigabytes of Memories would have to pay. The internet called it a "memory tax." Long-time users said the archive was the main reason they still had the app.
Other things persisted that nobody wanted to persist. The "speed filter," which displayed how fast a user was moving, was cited in lawsuits after a 2015 crash in Georgia at 107 miles per hour and a 2017 crash in Wisconsin at 123 miles per hour that killed three young men. In May 2021, in Lemmon v. Snap, the Ninth Circuit held that Section 230 did not shield Snap from claims about a feature it had designed itself. The filter was removed in June 2021. The FBI opened a probe in early 2023 into the sale of counterfeit fentanyl pills on the platform. In November 2024 the British children's charity NSPCC reported that, according to police data, Snapchat was the app most favored by online groomers. Australia barred it for users under sixteen in December 2025. The ghost that was supposed to make photos vanish has proven hard to make anything vanish, including liability.
The money has its own arc. Snap's market capitalization crossed $100 billion for the first time on February 22, 2021. Murphy's net worth that year was roughly $7.9 billion. In 2022 it fell by about 70 percent. By 2023 and 2024 the Los Angeles Business Journal listed him at a steady $2.4 billion, the 35th and then 36th richest person in Los Angeles; he had sold just over $46 million of stock since the fall of 2022. In 2025 Snap had revenue of $5.93 billion and an operating loss of $532 million, with 5,261 employees. The Business Journal's 2026 entry on him noted, with no particular malice, that the company was nearing a billion monthly users and that it had never once claimed a profitable year. A roughly $400 million deal to integrate Perplexity's AI search into Snapchat, signed in 2025, had been delayed and excluded from the company's outlook.
That ledger is the honest one. Snap invented Stories and watched every competitor copy them. It brought augmented reality to hundreds of millions of faces and has yet to sell a pair of glasses people line up to buy twice. It refused $3 billion, then $30 billion (Google reportedly offered that in 2016), and it has spent much of its public life worth less than at its debut. The two founders have kept control of all of it, through every quarter, with votes the public cannot touch.
And through all of it, one of them has kept his name nearly out of it. Search the record for a public argument between Spiegel and Murphy and you will not find one. Search for a Murphy interview about the redesign of 2018, when Kylie Jenner's tweet ("sooo does anyone else not open Snapchat anymore?") erased $1.3 billion in market value in a day and 1.2 million people signed a petition to undo it, and you won't find that either. What you find, over and over, is the same posture: the CTO next to the CEO, the developer Q&A, the engineer's numbers, the even keel.
The security chief's lesson was about who stays after the party is over. Spiegel heard it as a teenager and went to Stanford and threw more parties. The music would stop sometime after midnight, and the guests would scatter across the dark campus, and whatever had happened in that house would already be passing into the ephemeral condition of things half-remembered and never written down. And in the emptied room, putting it back the way it had been before anyone arrived, was the boy from Berkeley who would one day hold half the votes, with the lights still on.
Part IIThe Playbook
What follows is an attempt to extract operating principles from a career that its subject has almost never explained. That is a limitation, and also the point: much of what Bobby Murphy has to teach lives in what he did rather than what he said: the paycheck he split, the votes he holds, the conferences he attends and the interviews he declines. Each principle below is grounded in the documented record from Part I and ends with a single actionable tactic.
The principles
- 1.Earn trust in the unglamorous hours.
- 2.Fund your conviction before you ask anyone else to.
- 3.Partner across a real asymmetry.
- 4.Treat anonymity as a structural asset.
- 5.Build the medium, not the artifact.
- 6.Go stand in the crowd.
- 7.Lock control in symmetrically.
- 8.Write down the founding deal before it is worth anything.
- 9.Never let the promise outrun the system.
- 10.Keep a hardware bet alive through its write-downs.
- 11.Show up for the builders, not the press.
- 12.Price yourself at a dollar.
S
The Murphy Record
Key moments in the quiet half of Snap, drawn from the documented sources.
2010
Graduates from Stanford in mathematical and computational science; launches Future Freshman with Spiegel, splitting the equity.
2011
Future Freshman abandoned in March; codes Picaboo on Toyopa Drive; launches July 8; relaunched as Snapchat in September.
2012
Works at Revel Systems, spending half his paycheck on server bills; joins the board and becomes CTO in May; Lightspeed invests $485,000.
2013
Series A ($13.5M) and Series B ($80M); Facebook's $3B offer declined in November.
2014
Named to TIME 100; witnesses Live Stories at EDC Las Vegas; Vergence Labs acquired for $15M; Brown settlement announced in September.
2015
Forbes names him the world's second-youngest billionaire at $1.5B; Lenses launch in September.
2017
IPO on March 2 with 210.97M shares, equal to Spiegel's; non-voting public stock.
2022
Pixy halted; 20% of staff laid off; net worth falls roughly 70%.
2026
Presents the Specs developer platform at AWE: 7ms motion-to-photon latency.
Principle 1
Earn trust in the unglamorous hours
The most durable partnership in consumer technology over the past fifteen years was formed after the party, not during it. Spiegel had learned, from the head of his own security detail, that real friends are the ones who stay to clean up, and Murphy, across the hall at Kappa Sigma, kept being that person. The decision Spiegel later made, to write a corporate charter granting Murphy equal votes and succession to full control, would be inexplicable without that history. Charters codify trust. They don't create it.
The lesson for operators is less sentimental than it sounds. Every founding team runs a long, informal audit of who does the boring work when no one is watching: who stays for the post-mortem, who answers the 2 a.m. page, who reconciles the expense report nobody else wants to touch. Those observations accumulate into a credit score that later determines who gets the board seat, the equity, the benefit of the doubt in a dispute. Reggie Brown, by one source's account, was out partying while Murphy coded. When the moment came to decide whose version of the founding to believe, the audit had already been completed.
Tactic: For the next month, deliberately volunteer for the lowest-status task after every high-stakes meeting or launch, and notice who else is in the room with you.
Principle 2
Fund your conviction before you ask anyone else to
Murphy spent half of every Revel Systems paycheck on Snapchat's server bills until venture money arrived. That is the founder's version of a costly signal, and it did two things at once. Practically, it kept the product alive through the late 2012 growth spike that eventually caught Lightspeed's attention through Barry Eggers's daughter. Psychologically, it removed any doubt, in Murphy's own mind and in Spiegel's, about how serious he was. It's hard to imagine Murphy later wavering on a $3 billion offer for a company whose servers he had personally paid for.
Self-funding has gone out of fashion in an era of easy pre-seed rounds, and for good reasons; nobody should mortgage a house for an untested idea. But the specific discipline here is narrow: pay for the infrastructure of traction yourself, in the window between "people are using this" and "someone will fund this." That window is often short, it is where the strongest ideas die of neglect, and money spent there buys an unusually large amount of leverage in the conversations that follow.
Tactic: Identify the single cost that will kill your project if usage grows before funding arrives, and set aside a fixed share of your personal income to cover it for six months.
Principle 3
Partner across a real asymmetry
Spiegel studied product design; Murphy studied mathematical and computational science. Spiegel talked to investors and Stephen Colbert; Murphy wrote much of the core code and, by Business Insider's account, still authored much of it years after the IPO. Spiegel had more patents to his name; Murphy had the servers. This was not a partnership of two similar people who divided labor by preference. It was a partnership of two people who could not do each other's jobs.
⇄
The Founding Asymmetry
How the two halves of Snap divided the work, per the documented record.
| Dimension | Evan Spiegel | Bobby Murphy |
|---|---|---|
| Stanford major | Product design | Mathematical and computational science |
| Summer 2011 role | Design | Coding, 18-hour days |
| 2012 role | Promoting the app between classes | Day job at Revel; paying server bills |
| Screen time in IPO video | Roughly three times Murphy's | Praised "the brilliance of Evan" |
| Shares at IPO | 210.97M (plus 37.4M bonus) | 210.97M |
| Public circuit | Keynotes, media, podcasts | Developer conferences, Lens Fest Q&As |
Asymmetric partnerships fail when either side starts to covet the other's role or resent its visibility. There is no public record of Murphy wanting Spiegel's job, and no public record of a fight between them. The asymmetry held because both men seem to have regarded it as the design, not a compromise.
Tactic: List the three capabilities your company most needs from its founders and confirm that no two founders are primarily strong in the same one; if two are, decide now which of them owns it.
Principle 4
Treat anonymity as a structural asset
Murphy's low profile is usually described as a personality trait. It also functions as a corporate asset. Every public controversy at Snap, from the leaked fraternity emails to the "only for rich people" allegation to the Kylie Jenner tweet, attached to the person who spoke for the company. The CTO, meanwhile, retained the ability to recruit, to meet developers, and to lead research teams without his every move being read as a signal to the market. Ryan Jacob's observation that some co-founders genuinely enjoy walking around unnoticed is half the story. The other half is that their invisibility protects the institution's technical core from the weather that surrounds the CEO.
This only works if the quiet partner has real power. Murphy is not a cautionary tale about being sidelined, because his anonymity was paired with equal shares, equal votes, and a board seat since May 2012. Anonymity without leverage is just obscurity.
Tactic: If you are the less visible co-founder, make sure your governance rights are written down in proportion to your contribution, then let your partner have the microphone without competing for it.
Principle 5
Build the medium, not the artifact
Murphy's clearest statement of strategy, from 2014, was that he and Spiegel wanted visual content to become "a means of communication, rather than a piece of content around which communication actually happens." That distinction (medium versus artifact) explains almost every product decision that followed. Ephemerality made photos feel like speech. Stories made a day of photos feel like a conversation. Lenses made the face itself something you could talk with. Each move treated the camera as a keyboard rather than a gallery.
Most product teams build artifacts: posts, documents, listings, things that sit still and get commented on. The deeper opportunity is often in collapsing the artifact into the exchange itself, so that the thing users make is the interaction. That reframing is hard to copy, because competitors tend to copy features rather than the grammar that makes them work together, though Instagram's Stories showed that a determined rival with a bigger network can copy a lot.
Tactic: Write one sentence describing what your users make, then rewrite it describing what they say with it, and prioritize the features that serve the second sentence.
Principle 6
Go stand in the crowd
When Snapchat launched its Our Story feature in 2014, Murphy and Spiegel were not watching a dashboard in Venice. They were at Electric Daisy Carnival in Las Vegas, where Murphy saw "the most amazing content covering an event that we've ever seen." That firsthand observation led directly to the geofenced, crowd-sourced Live Stories model and, a year later, to an NFL partnership and Olympic coverage with NBC. The data would eventually have shown the spike. Being there showed why.
The same habit runs through the company's origin story. The growth that justified Lightspeed's first check came from a high school where Facebook was blocked on school iPads, and the investor heard about it from his own daughter. Snapchat's best signals have rarely come from analytics. They came from people physically near the users.
Tactic: Before your next major launch, put at least one senior technical leader in the physical place where the feature will be used most heavily, with instructions to watch rather than present.
Principle 7
Lock control in symmetrically
Snap's non-voting public shares drew criticism from governance advocates and a defense from Martin Sorrell. The more instructive detail is the symmetry inside the founders' block. Equal share counts. Equal votes. A provision that if one died, the other would control the company. That design made the partnership itself, rather than either partner, the sovereign. Neither man could be removed by investors, and neither could easily act against the other.
Symmetric control has costs. It concentrates risk in a relationship that has no external referee, and it asks public investors to trust two people instead of a board. But for founders who intend to make long, unpopular bets (refusing $3 billion, building glasses for a decade) it removes the most common failure mode, which is a board losing patience mid-bet.
Tactic: If you and a co-founder share a long-term vision, negotiate your voting arrangement with each other before you negotiate it with investors, and include an explicit succession clause.
Principle 8
Write down the founding deal before it is worth anything
The Reggie Brown dispute turned on an ambiguity that took about five minutes to create and about $157.5 million to resolve. Future Freshman LLC was renamed Toyopa Group LLC; Brown thought he was joining a new company, and Spiegel and Murphy thought he was helping with a product inside their old one. Murphy's own deposition captures the problem precisely: "I don't know what he believed." Nobody had asked. Nobody had written it down.
The cost compounded with the company's success. A disagreement that could have been settled with a few percentage points in 2011 became a lawsuit in 2013, a confidential settlement in 2014, and a disclosed nine-figure number in the 2017 S-1, along with an official acknowledgment that Brown helped originate the product whose ghost logo he drew. Ambiguity about founding equity grows at roughly the same rate as the company's valuation.
Tactic: Within one week of anyone contributing meaningful work to a new project, put their role, equity, and vesting in a short signed document, even if the answer is "none."
Principle 9
Never let the promise outrun the system
Snapchat's core promise was that messages disappeared. The system didn't fully honor it: snaps could be recovered, address books were collected, locations were transmitted. The 2014 FTC settlement imposed twenty years of independent monitoring. Gibson Security had disclosed the vulnerability behind the 4.6 million-account leak on August 27, 2013; the breach came on December 31. For a company whose brand was a technical claim, the gap between marketing and engineering was an existential risk.
This is the CTO's burden in its purest form. The product's story is written by the person who speaks for the company; its truth is determined by the person who runs the systems. When those diverge, regulators, plaintiffs, and users eventually close the gap, usually on terms the company doesn't choose. The speed-filter litigation, in which the Ninth Circuit held that Section 230 didn't shield a feature Snap had designed, is the same lesson applied to product design rather than privacy copy.
Tactic: Have the engineering owner of every user-facing claim (privacy, deletion, safety) sign off on its exact wording before it ships, and re-verify it each quarter.
Principle 10
Keep a hardware bet alive through its write-downs
Snap's hardware history reads like a casualty list, and it is also the clearest case in the company of a long bet surviving short-term failure.
◎
A Decade of Glasses
Snap's hardware efforts and their outcomes.
| Year | Effort | Outcome | Status |
|---|---|---|---|
| 2014 | Vergence Labs acquired for $15M | Foundation of Spectacles | Scan |
| 2016 | Spectacles V1, $130 | ~220,000 sold; $40M+ inventory losses | Risk |
| 2018 | Spectacles V2 | Iterated by Snap Labs | Try |
| 2022 | Pixy drone, $230 | Development halted within four months; hardware layoffs | Risk |
| 2026 | Specs AR glasses; Specs Inc. subsidiary | 7ms latency, 40% lighter, full developer platform | Try |
What made persistence possible was not optimism but structure: founders with decisive votes, and a CTO whose public identity is bound to the developer platform rather than to quarterly results. The bet was narrowed each time it failed. The selfie drone was cut, the scripted shows were cut, Zenly was cut, but the glasses and the AR platform behind them survived. Persistence, in this version, is ruthless about the adjacent bets so that it can afford to be stubborn about the central one.
Tactic: When a long-term bet misses, cut at least one adjacent project to fund it rather than cutting the bet itself, and write down which one you protected and why.
Principle 11
Show up for the builders, not the press
Murphy's public calendar is a map of his priorities: AWE in 2021, 2024, and 2026; EmTech in 2022; the annual Lens Fest Q&A where developers ask the founders whatever they want. He has given almost no general-interest interviews. The audience he courts is the few thousand people who will decide whether to build lenses and spatial applications on Snap's platform rather than someone else's, and those people want specifics: TypeScript support, a C and C++ native development kit, a Unity migration agent, a navigation engine ported in under two hours.
Platform companies live or die on developer trust, and developer trust is built by technical leaders who answer hard questions in person, year after year. Kent Bye's observation that this practice is rare among large companies is the point. It is cheap to do and almost nobody does it.
Tactic: Commit your most senior technical leader to one unscripted, open-question session with your external developers or power users every year, and publish the answers.
Principle 12
Price yourself at a dollar
Murphy's base salary is $1. He is not paid for his board seat. His 2023 compensation was $801,334, a rounding error against holdings that have swung from roughly $7.9 billion to $2.4 billion. The arrangement is common among founder-CEOs and less discussed among founder-CTOs, and its effect is the same: the person most responsible for the technical roadmap has no meaningful income that is independent of the company's long-term value.
That alignment does not guarantee good outcomes; Snap has never posted a profitable year. But it does make one class of bad outcomes less likely. A CTO paid primarily in salary has an incentive to keep the engineering organization comfortable. A CTO paid primarily in founder equity has an incentive to keep the company alive, which may mean laying off 20 percent of the staff, as Snap did in 2022, and keeping the glasses.
Tactic: Restructure at least one senior technical role so that the majority of its expected compensation depends on multi-year equity value rather than annual cash.
Free playbook
Get The Business Model Playbook
58 business models, one visual page each: how the money flows, the metrics that matter, and who runs it. Free when you join the Faster Than Normal email.
Free. No spam. Unsubscribe anytime.
Part IIIQuotes and Maxims
In their words
— Bobby Murphy, Google Cloud Platform Live, 2014Evan and I got started Snapchat in the summer of 2011, basically understood that visual content that was the most engaging, interesting form of content there was. We wanted to create a way that would enable that to be a means of communication, rather than a piece of content around which communication actually happens.
— Bobby Murphy, to ForbesWe weren't cool. So we tried to build things to be cool.
— Bobby Murphy, on the launch of Live Stories at EDC Las Vegas, 2014It was just utterly unbelievable how much really cool and creative and fresh content was being submitted to this live story.
— Bobby Murphy, to the Hindustan Times, 2021I think it represents a capacity to render or visualise software experiences in the way humans normally see the world and kind of imagine things in the world.
— Bobby Murphy, email to Chris Sacca, November 2, 2012I'd love to get your opinion on LA vs the Bay and building out a company especially as it related to Snapchat now and in the next few months.
Maxims
- Stay after the party. Trust is built in the hours nobody photographs, and it is redeemed in the documents everybody signs.
- Pay the server bill yourself. Funding the gap between traction and capital is the cheapest conviction you will ever buy.
- Choose a partner who can't do your job. Asymmetric skills make a partnership necessary; symmetric governance makes it durable.
- Hold power quietly. Invisibility is a strength only when it sits on top of real votes.
- Make the picture the sentence. The deepest products turn what users make into how they speak.
- Watch from inside the crowd. The dashboard tells you that something happened; the festival tells you why.
- Write down who owns what on day one. Founding ambiguity compounds faster than any valuation.
- Let engineering sign the promise. A product's claims are only as true as the systems beneath them.
- Cut the side bets to save the main one. Persistence means narrowing the wager each time it fails, not abandoning it.
- Answer the builders in person. A platform's future is decided by the developers who believe you'll still be there in five years.
