Microsoft CEO Satya Nadella
Photo: Microsoft

Microsoft: What post-pandemic cloud slowdown?

Protocol Enterprise

Hello and welcome to Protocol Enterprise! Today: Microsoft’s cloud growth story keeps on growing, no-code AI tools need guardrails, and Silicon Valley goes back to the garage.

Spin up

So much for solving their imminent supply-chain disasters: Data, analytics and AI execs said they spend 54% of their time working on marketing and customer engagement functions, according to the 2021 Europe and U.S. Data, Analytics and Artificial Intelligence Executive Organization and Compensation Survey from Heidrick and Struggles. Supply chain? It gets just 9% of their time.

Microsoft’s cloud muscle

We’re all running out of ways to describe Microsoft’s financial performance over the last several years. The company continues to put up strong growth numbers against huge performances a year ago, even if sometimes it feels like McKayla Maroney is running the after-hours stock trading desks.

Microsoft reported revenue of $51.7 billion during its second fiscal quarter, the last calendar quarter of 2021. That’s a 20% increase compared to the prior year, and as usual, the overall results were driven by the performance of its enterprise cloud services.

  • The closely watched “Microsoft Cloud” revenue number, which includes commercial Office 365 users, Azure, Dynamics 365 and parts of LinkedIn, came in at $22.1 billion, up 32% compared to the same period last year.
  • A year after taking control of Microsoft in 2014, CEO Satya Nadella set a $20-billion-a-year revenue goal for Microsoft’s enterprise cloud services business. It’s now recording that much revenue a quarter.

Microsoft investors were initially disappointed that the company didn’t have more to offer amid a week during which tech stocks have been hammered by investor concerns that the pandemic boom is coming to an end.

  • Azure growth appeared to be the culprit, slowing from 50% growth in the last quarter to 46% growth in the current quarter.
  • That was in line with the consensus Wall Street estimates for Azure growth, but apparently some investors were expecting a better performance and reacted accordingly.
  • However, after CFO Amy Hood projected that Azure growth would increase in the current quarter ending in March two hours after the release of the initial results, those after-hours bears turned into bulls.
  • The stock market is weird.

The long-term picture for Microsoft’s enterprise business remains quite strong. Even if the pandemic accelerated a lot of transitions to cloud computing, the overall trend won’t go away as life starts to approach something resembling normal.

  • The company has $147 billion in “commercial remaining performance obligations,” or business Microsoft has booked but has not yet delivered, on its books.
  • That’s a 31% increase compared to the same period a year ago, and a sign that Microsoft (like other cloud companies) continues to sign customers to long-term deals.
  • And don’t forget that Microsoft raised prices on Office 365 commercial customers this year, which likely has yet to cycle through its big installed base but will start to show up in future results.

Microsoft is the first of the Big Three cloud companies to report earnings this season; stay tuned for Amazon and Google’s results next week.

— Tom Krazit (email | twitter)


Club Revenue on Nasdaq digs into the strategies driving revenue growth at the highest performing companies. Tune in as Clari’s CMO Cornelius Willis interviews innovative revenue leaders to learn their tactics for building sales teams that drive unmatched success for their customers.

Learn more

No-code AI is democracy-lite

Software companies such as C3 AI talk up their no-code AI tools as a way to “democratize” AI, giving “citizen data scientists” the power to use technology that was once the sole domain of the coding elite. But considering the risks associated with Easy-Bake AI, it turns out even those companies place lots of restrictions on non-coders.

Not only are non-coders often limited to off-the-shelf machine-learning algorithms, companies including Databricks say users without formal machine-learning training may be prevented from putting models into production without oversight from data scientists.

"This is almost always true if the model is going to be customer-facing as opposed to an internal tool,” said Kasey Uhlenhuth, staff product manager at Databricks.

Ultimately, some companies including C3 AI encourage more machine-learning education.

“We basically have removed the coding from the equation, but you still need to train the citizen data scientists on what can you do with AI and machine learning,” said Ed Abbo, president and chief technology officer at C3 AI. With power comes responsibility.

— Kate Kaye (email | twitter)

Upcoming at Protocol

The process of delivering retail services changed forever in the wake of the pandemic. But how will shopping continue to evolve now that retailers around the world have embraced digital technology at a dizzying speed?

Join Protocol’s David Pierce on Feb. 1 at 9am PT for our virtual event, “How tech is making sure shopping will never be the same.” David will be joined by Allison Barr Allen, co-founder & COO of Fast; Faisal Masud, CEO of Fabric; and Jeremy King, senior vice president and head of engineering at Pinterest. RSVP here.

DIY chips — in the garage?

The same month that Intel announced its new Ohio megafab, a 22-year-old from New Jersey was building his own (somewhat smaller) version of a factory in a garage. Sam Zeloof, a Carnegie Mellon student studying electrical engineering, taught himself how to build chips from YouTube videos and engineering textbooks from the 1960s, according to WIRED.

To fabricate his computer chips, Zeloof scoured eBay for old chip gear, read up on patents and built his own machinery. He repurposed an old projector into a photolithography machine and substituted an HVAC system for X-acto blades and tape.

In the process he managed to do the seemingly impossible: turn hand-cut squares of polysilicon into actual working chips. Not everyone can turn used materials into computer chips without multimillion-dollar budgets, but Zeloof still believes he can make garage chipmaking more accessible.

Besides, his own journey from individual transistors to full-blown chips was inspired by another DIY-er, so who’s to say the rest of us can’t bring our own chips too?

— Aisha Counts (email | twitter)

Around the enterprise

Nvidia seems prepared to give up on its proposed $40 billion acquisition of Arm after widespread regulatory pushback. Arm could go public this year instead.

The global chip shortage won’t ease until the middle of 2022, according to the U.S. Department of Commerce, but you already knew if you read Max Cherney’s report on the topic earlier this month.


Everyone wants to IPO—but how do you really get it done, in a way that moves markets and inspires investors? You need the type of confidence and growth that starts from within. WalkMe’s CFO, Andrew Casey, shares how he rethought quotas, metrics, and what drives his teams, so WalkMe could go public—and go big.

Learn more

Thanks for reading — see you tomorrow!

Recent Issues