He Took a Company Public. It Blew Up. He Came Back Solo. | Francis Davidson, Odessia

Francis Davidson started Sonder at 19, grew it into a multi-billion-dollar hospitality company with ~1,000 employees, took it public, was forced out by the board at 26 after Covid devastated the business, and immediately began building his next company, Odessia — an AI-powered travel planning platform he’s building as a solo founder.

• Sonder’s origin: Davidson was a student in Montreal who noticed students left apartments empty during the summer while the city’s tourism season peaked. He sublet those apartments to travelers, pocketing the difference. With no capital, network, or experience, he wrote a PDF playbook and recruited ~30 entrepreneurial student interns across cities to replicate the model, generating ~$1 million in sales the first summer.
• Scaling strategy: Started with what was achievable — student apartments — then progressively moved to leasing directly from landlords, then small developers, then national developers. This “beachhead” approach let a 19-year-old eventually build high-rise branded hospitality properties.
• Covid and its aftermath: Sonder had just signed massive long-term real estate leases and was in hypergrowth when Covid hit. Revenue dropped ~90% — worse than anything in a century of hospitality data. Davidson raised ~$400M in emergency capital (2020–2021), took the company public prematurely, and pursued a major Marriott partnership. The stock fell 99%. After the Marriott deal closed, the board forced him out. Four months later, the business unraveled entirely.
• Why solo this time: Davidson knew immediately he’d start again. He explored ideas including AI-enterprise transformation and data centers, but Sequoia’s investment team (who led his pre-seed) told him the enterprise AI idea was a “7 at best” and suggested he think about consumer travel. That feedback, combined with his own research showing trips take ~20 hours to plan internationally, led to Odessia. He chose to go solo because the opportunity was time-sensitive, finding the right co-founder would take 6–18 months he didn’t have, and coding agents now let one person output what previously required large teams.

Odessia is an AI travel agent that plans entire trips — flights, hotels, trains, activities, restaurants — in a single conversational interface with generative UI.

• Unlike traditional travel sites that constrain users to search boxes (destination, dates, results), Odessia lets users describe what they want in natural language. The agent then uses dozens of specialized tools to recommend destinations, pitch itineraries, compare options visually on maps with postcard-style stopovers, check prices across multiple weekends simultaneously, and surface optimization opportunities (points, expiring credits, price anomalies).
• The product was built in ~5 months (January to early June) by a team of 8, using coding agents that Davidson says compressed development time by ~30x. He argues coding agents are rewriting the startup playbook: instead of building a minimal MVP with finite resources, teams can now ship ambitious, delightful products from day one.
• The business model is straightforward: earn commissions on hotel bookings. If users come and the product works, the revenue follows — so the focus is purely on execution and product quality.
• Long-term vision goes beyond on-demand planning: Odessia wants to proactively plan trips for users based on their bucket list, loyalty programs, and preferences — running background searches to surface optimal trips at the right time, price, and weather. Davidson believes reducing the activation energy of travel planning will cause people to travel significantly more.

Davidson’s philosophy on solo founding has two sides.

• Bear case: If he had world-class co-founders grinding 16 hours a day alongside him, they might go faster. There’s real value in complementary skill sets and shared burden.
• Bull case: There’s usually one general on the battlefield — multiple leaders create debate and slow execution. Organizations are typically led by one person, and a kind of “dictatorship” can be valuable in a startup. With coding agents, a single founder can now have disproportionate output, making the opportunity cost of spending months “co-founder dating” extremely high. Davidson also argues it’s better to build first, gain traction, and then attract truly exceptional talent (CTO, executives) from a position of strength rather than convincing people to join a pre-launch idea.

On hiring and team culture as a solo founder:

• Davidson hired 8 people before launch, all with “founding” in their title. He wanted people with end-to-end ownership who are bought into the vision and can execute independently — not people waiting for co-founder-level strategic debates.
• He prioritizes senior people who combine engineering, product intuition, and design taste — what he calls the “100x engineer” (engineering + product) or even “1000x” (adding design). These people receive high-level instructions and deliver quality output autonomously.
• He screens candidates for advanced coding agent usage: running multiple agents in parallel, trusting the tools, and having established workflows. People who watch agents do every step one-at-a-time or don’t trust them are red flags.
• Culture was defined from day one with a written mission statement and four values in the onboarding doc. He believes solo founders need to be more intentional about culture since there’s no co-founder group to organically establish it.
• He sees a structural advantage: without co-founders, the founding team feels closer to the top and more ownership over the company’s identity and success.

Key lessons from Sonder applied to Odessia:

• Founder mode over executive delegation: Davidson’s biggest lesson is skepticism toward hiring external executives and delegating. At Sonder, he felt 90% of executive hires underperformed. He now believes in staying deep in the details — the “Zuckerberg/Chetsky style founder mode” — and hiring exceptional individual contributors who can grow into leadership, rather than importing senior managers who are better at accumulating power than building.
• Build from within: He hired too quickly and too externally at Sonder. He now wants to take rockstar ICs and give them leadership opportunities, referencing companies where VPs and C-level leaders are in their 20s.
• Personal output matters more than ever: With AI agents, every hour not spent building has enormous opportunity cost. Davidson dreads meeting-heavy days because his team can output ~20,000 lines of production code daily.
• Naming and brand: Both Sonder and Odessia took ~6 weeks of obsessive naming work each. Davidson tried hiring external help both times but ultimately came up with the names himself. Odessia maps to the Odyssey (epic Greek journey), is an invented word with a .com domain, ranks #1 on Google with domain authority of 3, and supports a brand identity around a fictional Greek goddess of travel.

On idea validation and the “idea maze”:

• Davidson initially pursued an enterprise AI transformation idea (an “AI Gartner”), speaking to CIOs and getting lukewarm-to-warm signals. Sequoia’s feedback was the critical pivot point — they told him the idea needed work and pointed him toward consumer travel.
• He distinguishes between ideas that are simple to express in one sentence (Sonder: branded apartments for travelers; Odessia: AI helps you plan a trip) versus ideas requiring nuance that are hard for people to quickly understand. The simple ones have been far more compelling.
• He references the book Monetizing Innovation as the sharpest framework for derisking ideas — going beyond “do you like this?” to willingness-to-pay questions and feature stack-ranking to quantify the dollar value customers ascribe to a product.
• On the “lukewarm signal” trap: Davidson notes that customers saying “yeah, that’s cool” to a theoretical idea is very different from actual behavior. He advises being quantitative about signal quality — a $5K/year contract vs. a $1M enterprise contract are radically different businesses.