Lyft picks up $300 million bill for Amazon AWS: Cloud contracts of the future

Lyft & Amazon AWS

Lyft, probably the best-known Uber rival, recently filed its IPO (Initial Public Offering) documents with the US Securities and Exchange Commission (SEC), and the S-1 Registration Document gives an insight into a company’s finances and throws up a couple of interesting points for ITAM.

Lyft talk about their relationship with Amazon AWS in the documents and page 102 shows that, in March 2018, Lyft signed an agreement with Amazon to spend “an aggregate of at least $150 million” on AWS services up to June 2021.  That’s a pretty significant contract – $150 million over 39 months = $3.8 million per month, which is almost $46 million in a full year.

However, in January 2019 the contract was amended; under the new agreement, Lyft committed to spending $300 million up to June 2021. That’s double the amount of spend in 75% of the original timeframe (29 rather than 39 months), which equates to a little over $10 million per month – $120+ million in a full year.

The contract states a minimum of $80 million spend in each of the three years and we see on page 38:

“If we fail to meet the minimum purchase commitment during any year, we may be required to pay the difference, which could adversely affect our financial condition and results of operations”

Lyft must have a particularly strong view of their planned cloud use for the next few years, but even so – it’s interesting to see an organisation commit themselves to such a large amount of guaranteed spend, particularly in the cloud.

We can see the potential downside for Lyft, but where’s the upside? If they use more AWS than expected, more than that $300 million (or $80 million a year) – what, if any, benefit/reward do they receive? I’ve assumed they get better than usual pricing from the outset of the contract but where’s their incentive to over-achieve? Perhaps there’s tiered pricing so the “per unit” cost decreases if spend hits X?

Negotiate Next

I’d also be interested to learn what, if anything, has been negotiated for after this contract. It comes to agreement renewal in a couple of years, where the customer is $300 million deeper into AWS – using more services and more reliant on it even than now – and oh…the price is significantly higher. What’s going to happen then? Exiting a cloud service is rarely easy but following a period of using as much cloud as possible, it will be almost impossible I’d imagine. Are Amazon biding their time, waiting to make up any discounts on this contract with price increases on the next?

From Amazon’s perspective, almost regardless of the “unit price” discounts they’ve given, it’s a smart move. Not only are they guaranteed a sizeable amount of income, the chances of Lyft moving to a competitor are almost 0% for the next 2 years. No matter how attractive another cloud looks, with great pricing and awesome features…the benefits are unlikely to be worth enough to balance out the remainder of Lyft’s contract at that point in time.

Unless…

La Liga y La nube

This puts me in mind of the concept of “buyout” clauses in the contracts of sports stars, particularly prevalent in Spanish football. Aimed at stopping teams from blocking player moves, they’re a contractual standard in Spain.

Although a club must include such a clause, there is no restriction on the size of the buy-out amount – which is how Ronaldo had a €1 billion clause at Real Madrid, and Messi with a mere €700 million in his Barcelona contract.

As you can see, if Team A has a player they don’t want to lose, they set the buy-out clause at such a high amount they hope it will dissuade potential buyers; if it doesn’t and the player is bought, Team A will earn enough money to cushion the fact they’ve just lost a star player!

A good example was the transfer of Neymar from Barcelona to Paris Saint Germain (PSG), where the French club triggered the Brazilian’s buyout clause of €222 million. PSG wanted Neymar so much, they were willing to pay such an over-inflated (even in the world of football) figure.

Could we see something similar in the world of cloud contracts?

Let’s say Microsoft are confident Lyft’s business will continue to grow healthily for the next 7 years, and estimate they’ll spend at least $900 million on cloud services over that period. What if they “bought out” Lyft from their AWS contract – effectively spending $300 million to pay off Amazon in return for locking Lyft into an Azure contract that guarantees $900 million spend – so $600 million “profit”* for Microsoft? (*Not all profit but you know what I mean!).

While there would be other considerations, primarily the difficulty of moving from one public cloud provider to another, it could be an interesting tool for the cloud providers.

References

Lyft SEC S-1 Registration Document – https://www.sec.gov/Archives/edgar/data/1759509/000119312519059849/d633517ds1.htm

Neymar Transfer – https://www.sports.legal/2017/08/the-neymar-transfer-an-analysis-of-buy-out-clauses-part-1/