The Airline That Decided to Move Like a Startup

Alaska Airlines is 94 years old and has never been through a bankruptcy. That record is not luck. It is the product of an institution that knows, in its bones, what it means to be accountable for an outcome.

Which makes what Pasha Saleh, Head of Corporate Development at Alaska, described on stage at Aviation Festival Americas today worth sitting with. He had to stand in front of the board of a 94-year-old airline and make the case for building startups, knowing that nine out of ten of them fail.

The answer was UP.Labs, a Los Angeles and Silicon Valley venture studio that builds companies by embedding directly with organizations that own the hardest problems. Among their partners, Porsche and J.B. Hunt. Stefan Teulon, President of Aviation Ventures at UP.Labs, described the model simply: spend a year inside the problem alongside Alaska’s domain experts, then bring in a founding CEO to own the outcome and spin the company out. Alaska stays as customer one, investor one, and 25% owner of the IP it helped create.

That 25% matters more than it sounds. Saleh described Alaska as the first airline in the world to sell a ticket on the internet. Its own technical pilots invented GPS ground navigation. In both cases, the value left, absorbed by partners and vendors. The intelligence walked out the door, and Alaska watched it go.

The venture model exists, in part, so that does not happen again.

Three years in, the aviation partnership has produced three companies. The proof that the model works came early, and sideways. Odysee was built to optimize schedules. Working from the problem rather than against a spec, the team discovered something they had not set out to find: IP that addressed guest disruptions in real time. A side benefit that became a measurable multi-million dollar outcome Alaska had not asked for and did not anticipate.

There is a quieter version of the same story playing out inside Alaska’s operations every day. Revenue management is one of the highest-turnover roles in an airline. The analyst who spent three years watching a competitor learns things the system never captures. When they leave, that understanding leaves with them. It is the same pattern Saleh described on stage: intelligence built inside Alaska, walking out the door. telos was built from that problem. Not to replace the analyst, but to make sure what they learned does not walk out when they do.

Each company that comes out of this model enters the market with something most startups spend years trying to earn: a real customer, a real problem, and an institution that built it because there was no other way.