Google wins a bankruptcy auction with a $10M bid to acquire deidentified business data, software code, and more from Spirit Airlines to improve its AI models

Bloomberg Law 2026-08-18 James Nani

Context & Ripple Effects

Google’s bid extends an AI dealmaking pattern that already includes reported talks to license technology and hire talent from AI coding-agent startup Mechanize. It also follows Google’s rollout of natural-language flight search in Google Flights, making airline-derived assets strategically adjacent to an existing travel-AI surface.

The transaction matters less as a conventional airline acquisition than as a purchase of data and software from a distressed company: Google is using a bankruptcy auction to obtain inputs intended for its models.

First-order effects

Google gains Spirit Airlines’ deidentified business data and software code for model-improvement work, while Spirit’s bankruptcy estate receives the winning $10 million bid.
The acquired assets move under Google’s control rather than remaining available for competing buyers to license or acquire through the auction.

Second-order effects

Google’s pursuit of both distressed-company assets and outside AI technology gives it multiple routes to assemble proprietary model inputs alongside its internal products.
The deal reinforces the commercial value of deidentified operational data and production software as assets that can be separated from a failing company and sold to AI developers.

Third-order effects

If bankruptcy auctions become a repeatable source of data and code for large model builders, distressed-company estates may increasingly treat those digital assets as distinct monetizable property rather than incidental operating records.
AI competition would then turn further on access to proprietary real-world datasets and software artifacts, not only compute and frontier-model research.

The trend: Large AI platforms are broadening their supply chain for model inputs, using licensing, talent deals, and asset purchases to secure proprietary data and software.

Discussion

@levie Aaron Levie on x
When you hear that data is the new oil, this is ultimately what that looks like. AI has such a thirst for data that we're entering an era where it's valuable almost in any form. In a world of AI, information actually belongs as an asset on the balance sheet. This data sale is
@chrisjbakke Chris Bakke on x
“Gemini build me an ultra-low-cost airline that charges $59 to fly across the country + $300 per checked bag. Make the seats carved out of actual rocks. Make sure the staff (including pilots) are semi-functioning alcoholics. Make no mistakes.”
@lizwessel Liz Wessel on x
This speaks well to how I feel about most (not all!) data labeling businesses
@yangkeving Kevin on x
there are obvious structural reasons why mercor was not willing to bid more than $7.5m for this data. they have the money
@deepdishenjoyer @deepdishenjoyer on x
this is a really interesting precedent lots of companies go bankrupt every year
@_arohan_ Rohan Anil on x
Model trained on data a company that seized to exist. Would this make the model that trains on it have existential dread?
@giyu_codes @giyu_codes on x
I need to raise my prices
@ypatil125 Yash Patil on x
Don't do this
@thezvi Zvi Mowshowitz on x
It sounds cheap but we'll see when and if the data actually safely arrives at its destination.
@abhijaymrana Abhijay Rana on x
Wow, seems like Google is buying Spirit Airlines' enterprise data for $10m (outbidding Mercor at $7.5m). Basically includes every internal document, email, workflow, and codebase for a once $6B company. Honestly, $10m for 34 years of operational data really seems like a steal.
@max_spero_ Max Spero on x
$10m for 34 years of data on how to run a budget airline into the ground
@smb_attorney @smb_attorney on x
Wow this is *terrible* Google is not just buying “operational data.” The package includes more than 175,000 employee records dating back to 1986, plus productivity metrics, behavior data, HR records, internal emails, and 500 million Microsoft Teams chats/collaboration logs.
@amanda_goodall Amanda Goodall on x
Nothing you wrote at work in 1997 was prepared for this plot twist.
@hammer_mt Mike Taylor on x
Your company data just isn't worth that much — $20 per worker per year.
@marell_evans Marell Evans on x
What a signal. There's a massive opportunity to build a company that systematically acquires valuable data assets from distressed businesses across industries and geographies, then repurposes them for the AI economy. As companies disappear, their data doesn't! Real cool💡
@regulatorynerd Matt Janiga on x
If you're going to shut down your startup, feels like you should at least contact Google Corp dev to see if they'd buy your enterprise data, too.
@highyieldharry @highyieldharry on x
Great work once again from the antitrust masterminds
@dejavucoder Sankalp on x
google is rl env maxxing
@p_remarks @p_remarks on x
Seems like an interesting precedent. Time to scour the shitco microcap database for valuable data plays
@abhijaymrana Abhijay Rana on x
Between this and the Mechanize acq, it seems like in-housing data efforts is Google's P0. Always funny when people immediately think “dead startup” == “bad data” though, as if the reason Spirit went down is because they didn't file their quarterly financials correctly.
@harveenchadha Harveen Singh Chadha on x
this is going to be the norm if you are a regulated enterprise industry you are sitting on a data mine that you can potentially turn into huge profits, frontier labs need actual data more than synthetic to make meaningful progress basically every company is a rl environment
@turnernovak Turner Novak on x
I hope Google is doing this to start an airline
@ryanlockai Ryan Lock on x
If you want to get a piece of this, we just helped a roofing company get a $450k offer for their De-Identified Data. These labs need your workflows to train their models and are paying high dollar right now ($100k-$1M) Dm me if you want to get a quote on your data.

Chronicles