• Fiscal.ai
  • Posts
  • 🗞 Is This the Biggest Market Ever? $1.5T in Total Commitments

🗞 Is This the Biggest Market Ever? $1.5T in Total Commitments

Accelerating growth at unfathomable scale.

Written by: Ryan Henderson & Braden Dennis

Happy Sunday! 👋 

Today we’re taking a look at the astounding growth at scale for the 3 leading cloud giants.

Let’s dive in!

Platform Update

This week, Fiscal.ai launched Beats/Misses in the Estimates Tab 📈 

Now, right after earnings, you can see whether or not a company beat estimates across a variety of metrics (Revenue, EPS, Free Cash Flow, CapEx, and more).

This feature also includes up to 10 years and 40 quarters of historical beats and misses.

Featured Story

Is This the Biggest Market Ever?

Over the last several years, much has been made about “enterprise AI demand”. But what does that actually refer to?

Let’s give an example: In 2023, Duolingo introduced a new premium tier called Duolingo Max. The tier includes a feature called “explain my answer” that can audibly articulate to a user why their answer was wrong, much like a human teacher.

Underneath the hood, every time a subscriber taps that “explain my answer” button, Duolingo is spending less than one-tenth of a cent to ping OpenAI’s latest models for an answer.

If we take this analogy further downstream, in order to provide the best possible service to Duolingo, OpenAI has to continuously train its models. To do that, OpenAI partners with a public cloud provider like Microsoft Azure and buys compute capacity. To be more specific, OpenAI is renting the capacity of a bunch of AI-optimized servers loaded with semiconductors probably in some datacenter in the suburbs of DC. (Fun fact: 70% of internet traffic passes through a DC suburb called Loudoun County)

This is one example of an enterprise AI application. But if you zoom out, pretty much every company in the world is coming up with their own applications for AI. And this is, unsurprisingly, leading to a jaw-dropping number of orders for the public cloud vendors.

Big tech reported earnings this week, and one metric in particular stood out: Backlog.

Amazon Web Services, Microsoft Azure, and Google Cloud (aka “The Hyperscalers”) now have nearly $1.5 trillion in combined customer commitments. They have added $701 billion in new commitments over the last 6 months alone!

To be clear, these are not empty promises. In order to be counted as remaining performance obligations in an SEC filing, the commitments have to be “contractually binding and non-cancellable”.

While the duration of these commitments can be quite long (Anthropic recently signed a deal to spend $100 billion with AWS over the next 10 years), the Hyperscalers are still seeing accelerating growth in the short term as well.

In fact, each of the cloud providers reported an acceleration in revenue growth from Q4 to Q1.

Let’s take a look at each individually:

Compared to Azure and Google Cloud which tend to bundle other features and book commitments years in advance, AWS encourages a pay-as-you-go model. That’s why AWS is still the largest cloud vendor on a revenue basis, despite having the smallest backlog.

When Amazon reported earnings on Wednesday, the company surpassed analysts already rosy assumptions for growth. AWS, which is now a $150 billion ARR business, grew 28% YoY marking its fastest growth rate in 15 quarters.

One big differentiator for AWS is their custom chips business. While clouds like Azure and Oracle rely heavily on chips from other vendors, Amazon is able to save a ton thanks to its custom silicon chips called Trainium. Here’s a quote from the earnings call:

“We expect Trainium will save us tens of billions of dollars of CapEx each year and provide several hundred basis points of operating margin advantage versus relying on others' chips for inference.”

Outside of its cloud division, Amazon reported strong numbers across the board.

Though the company is still best known for its low-margin online stores business, Amazon reported record operating margins as it now generates more than 60% of its revenue from its other, higher-margin segments.

  • Subscriptions: 7% of Revenue

  • Advertising: 9% of Revenue

  • AWS: 21% of Revenue

  • 3P Seller Services: 23% of Revenue

The complete change in Mr. Market’s perception of Google over the last 12 months has been astounding.

In April of last year, Google was in a nearly 30% drawdown, trading at an EV/EBIT of 14.5x, and deemed an “AI loser” by most. Fast forward 12 months: Shares are up 165% off the April lows and Google is now knocking on the door of being the largest company in the world by market cap.

What happened?

Google is now viewed as the only true, full-stack AI company. They own the customer application layer, the LLM layer, the cloud layer, and even some of the chip business as well.

This full stack approach is helping Google win market share, especially for AI-focused workloads:

“We are unique in the market because of our vertically optimized AI stack... The fact that we own frontier models, own the silicon, really helps us stay ahead of the curve. I think we are the only provider in the market that offers all of these in a vertical stack.”

This demand is showing up most notably in Google Cloud’s backlog. Google has now grown its backlog by 398% compared to the same period a year ago.

Also, like AWS, Google Cloud has its own custom silicon business known as tensor processing units (TPUs), which is helping the division deliver significant margin expansion, despite the rising prices from most chip vendors.

Compared to Google and Amazon, Microsoft had a relatively quiet quarter, which feels a little weird to say for a business that generates $318 billion in annual revenue and is still growing 18% annually.

Unfortunely, Microsoft doesn’t split out Azure’s full revenue numbers directly so it’s difficult to make apples to apples comparisons. However, they do give the revenue growth rate.

Azure accelerated revenue growth this quarter to 40%, which sits squarely in the middle of AWS and Google Cloud. While this is still an astounding growth rate for a business that’s estimated to have more than $80 billion in annualized recurring revenue, investors seem a little more cautious with Microsoft than its peers.

Shares of Microsoft fell slightly after the earnings report and are now in a 24% drawdown from the latest highs. This seems partly attributable to Microsoft’s close relationship with OpenAI, which has been experiencing significant market share losses in recent months and accounts for more than 50% of Microsoft’s $633 billion in remaining performance obligations. 

That’s all for this week. 

If you have any questions about Fiscal.ai or any feedback for the newsletter, feel free to reply to this email!