Shares of Microsoft (MSFT 1.33%) fell 11% in January, according to data from S&P Global Market Intelligence.
Software stocks have been under pressure lately as investors fear the impact generative AI will have on their businesses. Additionally, other investors have been wondering whether artificial intelligence companies will be able to earn adequate returns on their massive capital spending.
While Microsoft reported a solid beat on its first-quarter earnings and also unveiled a groundbreaking self-designed inference chip, the stock still fell, as its "mild beat" wasn't enough to allay those multiple fears.
NASDAQ: MSFT
Key Data Points
Microsoft exceeds growth expectations, but also spending expectations
In its fiscal second quarter, Microsoft grew revenue by 17% and adjusted (non-GAAP) earnings per share by 24% to $4.14. Both figures beat expectations.
At first glance, all of Microsoft's key divisions showed strong growth, with Azure up 39%, the Microsoft 365 commercial cloud software up 15%, and its Dynamics enterprise resource planning software up 16%.
So, what exactly was the problem? It's hard to pin down, but it likely had to do with the even greater surge in capital expenditures. Capex was $37.5 billion in the quarter, including finance leases, up 66% from the prior year's $22.6 billion. That pushed free cash flow down to $5.9 billion, well below the company's $38.5 billion in net income.
In that light, investors may have wanted to see an accelerating growth rate in key segments. Azure's 39% growth was very high but down one percentage point from the prior quarter. Meanwhile, forward revenue guidance for next quarter is between $80.65 and $81.75 billion, which would amount to around 16% revenue growth, one percentage point of deceleration relative to Q2.
So while Microsoft's growth continued to be strong -- incredibly strong for a company of its size -- investors may have been hoping for even more revenue growth.
In addition, Microsoft reported a massive $625 billion in remaining performance for its Azure cloud unit. RPO is essentially future contracted revenues. But the problem investors may have had was that 45% of that total is a future commitment from OpenAI. In recent months, some have come to doubt OpenAI's ability to fulfill its massive $1.4 trillion in future spending obligations, given that OpenAI reportedly "only" made $20 billion in revenue last year, while also burning through billions in cash.
Image source: Getty Images.
But the sell-off seems like an overreaction
Microsoft management maintained that it is supply constrained, with demand exceeding its ability to serve customers, which is why it's accelerating its capital expenditures. Rising memory prices are also driving up costs. Furthermore, not all of its capital expenditures are going to the Azure business, as Microsoft is also infusing its own first-party software across 365, Dynamics, GitHub, and other products with AI diffusion.
So, overwhelming demand seems like a good problem to have.
Furthermore, in January, Microsoft also unveiled its new Maia 200 AI inference chip. In the blog post describing the chip, Microsoft disclosed that the new Maia 200 outperforms AWS' Trainium and Google's TPU on several metrics, including teraflops per second on FP4 and FP8 precision settings, as well as HBM (high-bandwidth memory) capacity. By deploying its own self-designed chip, Microsoft should be able to develop performant AI at lower relative costs going forward.
Given that Microsoft isn't just a software company but also one that builds the expensive AI infrastructure behind its own products, Microsoft shouldn't be grouped with other software stocks that AI agents may disrupt. And while OpenAI's competitiveness bears monitoring, at the moment, it is still very much an AI leader. Thus, Microsoft may be worth a look for investors on this pullback.
