Hello and welcome to Protocol Enterprise! Today: while consumer demand for semiconductors plummets, cloud tech providers can’t get enough chips, the third-quarter cloud provider market share numbers are in, and what might come next in the evolving U.S. policy toward AI in China.
Hyperscalers hungry for silicon
The semiconductor industry is experiencing whiplash at the moment: Within the span of weeks, consumer chip sales have gone from boom to bust, perhaps one of the fastest shifts in recent memory.
Earnings season thus far has reinforced that trend, and reports from companies such as Texas Instruments suggest the slump may be expanding into other markets, such as industrial applications. But not so for the data center: Large U.S. hyperscalers such as AWS, Google, Meta, and Microsoft issued quarterly report cards this week, and the verdict was positive overall even if their revenue isn’t growing as fast as it once was.
- Cloud computing revenue growth is slowing, but after years of 30% to 40% growth, it was obvious it was going to happen, according to Maribel Lopez of Lopez Research.
- “There is a major rationalization happening in the marketplace — ’we have five cloud vendors, how is that possible? Shouldn’t we have two?’” Lopez said, echoing a common conversation she has with companies now.
But slowing revenue growth hasn’t translated to massively scaled back data center expansion plans from the big cloud providers just yet. Jefferies chip analyst Mark Lipacis wrote in a client note that the commentary from Google, Meta, and Microsoft was positive for data center chip companies such as Nvidia, AMD, Marvell, and Broadcom.
- Meta, in particular, is boosting its spending on new infrastructure by 12% to $36.5 billion at the midpoint of the company’s guidance.
- In its earnings call this week, CFO David Wehner said that Meta’s AI applications require much more expensive server and networking equipment and were a big contributor to the increase in spending plans.
- Bernstein chip analyst Stacy Rasgon noted in an interview that Meta’s data center plans were under close scrutiny because nearly all of its revenue is from advertising, which has been hurt as the economy teeters. Of the huge tech companies, it would have been more likely to make cuts.
The other hyperscalers talked about expanding their data center spending too.
- Google executives discussed spending more on technical infrastructure, and servers are the largest component.
- Azure’s revenue missed Wall Street expectations slightly, but Microsoft executives said that capital spending will rise sequentially based on strong customer signals.
- Amazon issued good news too: Executives said it planned a year-over-year increase of $10 billion in tech infrastructure to support AWS’ rapid growth.
Even if there’s continued strength among the hyperscalers, Intel doesn’t appear poised to take advantage. Nothing can sum up the cost of years of technical mistakes at Intel better than its latest data center and AI quarterly report card: The once-flush segment has ceased to turn a profit, and its revenue fell 27% to $4.2 billion from just a year ago.
- According to Jefferies data, as recently as five years ago Intel commanded nearly 100% of server processor sales.
- As recently as the third quarter, Nvidia and AMD now take about half of the total quarterly sales of CPUs and GPUs to data center customers, and the veteran Intel brings home roughly half too.
- Big wins for AMD and Nvidia represent a huge problem for Intel in the longer run.
- Referring to the company’s consumer business, Dylan Patel, chief analyst at SemiAnalysis, quipped, “I don’t care about a dollar of client earnings, give me 10 cents of data center earnings.”
A MESSAGE FROM CAPITAL ONE SOFTWARE
Many business leaders aren’t sure where to begin when it comes to migrating to the cloud. To help organizations adapt to this revolution, Capital One launched Capital One Software, a new enterprise B2B software business focused on providing cloud and data management solutions.
Google Cloud makes a move
Google Cloud inched up in the market share rankings for cloud providers this week after reporting 38% revenue growth for Google Cloud Platform services, Google Workspace collaboration tools, and other enterprise services.
The third-largest cloud provider, Google now has 11% share, up from 10% in the last quarter, according to a report yesterday from Synergy Research Group. AWS’s and Microsoft’s market shares remained steady in the top two spots at 34% and 21%, respectively.
Overall, enterprise spending on cloud infrastructure services totaled more than $57 billion in the third quarter, an increase of more than $11 billion from the same period last year, according to Synergy. And according to figures reported by the Big Three in their quarterly earnings, there’s a lot more in the pipeline.
Alphabet’s revenue backlog, which represents customer contract commitments for future services and is primarily related to Google Cloud, grew to $52.4 billion as of the end of September, from $51.2 billion in June, according to a quarterly report filed with the SEC. The cloud provider expects to recognize about half of that amount as revenue in the next 24 months.
Amazon’s revenue backlog, which is primarily related to AWS, grew to $104.3 billion from $100.1 billion from the end of June through September. The average remaining life of its long-term contracts is 3.8 years and AWS said it was unable to determine the pace at which it would recognize that revenue because “the amount and timing of revenue recognition is largely driven by customer usage, which can extend beyond the original contractual term.”
Microsoft reported $183 billion in remaining performance obligations. Of that, $180 billion was related to its commercial revenue, which includes Azure and other cloud services, Office 365 Commercial, the commercial portion of LinkedIn, Dynamics 365, and other commercial cloud properties. That number was down from $189 billion reported in June, but the company expects to recognize approximately 45% of that revenue in the next 12 months.
— Donna Goodison (email | twitter)A U.S. AI reserve force?
For some companies operating in AI, data, or emerging technology fields — particularly those without any ties to China — the latest news out of Washington must look promising.
The U.S. Commerce and Defense Departments and White House National Security office are working as a tag team delivering offensive blows to China’s technology industry while signaling more protections for industry, including more government fusion with private sector tech.
“We have a whole China assessment going on. I meet with my staff once a week and say, ‘Okay, what do we do next?'” said Alan Estevez, undersecretary of commerce for industry and security at the U.S. Department of Commerce Bureau of Industry and Security in a talk yesterday. What’s next, as also signaled by the White House and Defense Department, includes efforts to rein in China’s AI, biotech, and quantum capabilities, he said.
It remains to be seen what a regulatory approach to curbing AI’s open-source-centric research and development ecosystem looks like in practice.
Meanwhile, a new national defense strategy document from the Defense Department explains that the U.S. will follow industry’s lead on AI, autonomous tech, and human-machine interfaces with military relevance.
And it aims to boost opportunities for private sector integration. “We will increase the availability of fellowships, internships, and rotational assignments — including in the private sector — to grow the skills of our workforce, provide a broad range of experiences, create collaboration opportunities, and carry best practices back to the Department,” the plan stated.
Around the enterprise
VMware patched a severe vulnerability in its Cloud Foundation and NSX Manager products after exploit code was released earlier this week.
KubeCon North America took place this week in Detroit Rock City, and Datanami had an overview of some of the developments around its flagship project, Kubernetes.
A MESSAGE FROM CAPITAL ONE SOFTWARE
The flexibility of the cloud helps companies like Capital One unlock access to their data with performance that can scale instantly. But this flexibility and scale can also create a unique challenge for organizations and users who are not proficient in cloud optimization.
Thanks for reading — see you Monday!