In last week’s earnings calls, leaders from technology giants Microsoft, Meta, and Apple showed commitments to capex and little fear about the arrival of DeepSeek and its claims of a cheaper, more computationally resourceful AI model.
The China-based company DeepSeek, which recently took the number one spot in the App Store, has touted low-cost, efficient advances in inference, which is how an AI model makes predictions. (The model’s accuracy, however, has come into question, following penetration tests.)
While the arrival created an early stock-market stir for AI-powering chipmakers and energy producers, Big Tech leaders appeared unfazed by the competitor and see efficient AI, generally, as a win.
“When token prices fall, inference computing prices fall, that means people can consume more, and there will be more apps written,” Microsoft CEO Satya Nadella said during a Jan. 29 earnings presentation.
Microsoft. Nadella also shared that the DeepSeek’s R1 model launched on Microsoft’s GenAI-management platform on AI Foundry and GitHub “with automated red teaming, content safety integration, and security scanning,” and that developers will soon be able to run R1 models locally on Copilot+ PC machines.
Microsoft announced a quarterly capex of $22.6 billion, an increase from $20 billion last quarter. Nadella said the company had more than doubled its overall capacity in the last three years.
Meta. Meta, in its earnings call on Jan. 30, shared that its plans for an almost Manhattan-sized data center are still in motion. Capex reached $14.8 billion for the quarter, “driven by investments in servers, data centers, and network infrastructure,” according to CFO Susan Li.
DeepSeek uses a technique called “inference-time compute scaling,” which prevents the model from using its full computational power all at once.
Top insights for IT pros
From cybersecurity and big data to cloud computing, IT Brew covers the latest trends shaping business tech in our 4x weekly newsletter, virtual events with industry experts, and digital guides.
CEO Mark Zuckerberg, on the call, said he believes that “emerging” models will apply more compute at “inference time,” presumably when a model makes its predictions, to generate higher intelligence and better service.
“As a company that has a strong business model to support this, I think that’s generally an advantage that we’re now going to be able to provide a higher quality of service than others who don’t necessarily have the business model to support it on a sustainable basis,“ Zuckerberg said.
Meta recently announced a memory-like capability for its Meta AI assistant—a technology that Zuckerberg said he expects to be a “leading” one in 2025.
“The field continues to move quickly. There’s a lot to learn from releases from basically everyone who does something interesting, not just the ones over the last month,” he said on the call.
ServiceNow. Bill McDermott, the chairman and CEO of workflow-management platform ServiceNow, sees DeepSeek’s arrival as “fantastic news” as AI models quickly become commoditized. “The competitive differentiation will happen at our level in terms of the applications, the business processes, how you can help these companies run better, and then ultimately get a business outcome,” McDermott said during a Jan. 29 earnings meeting.
Apple. Apple CEO Tim Cook told CNBC that AI capabilities impacted iPhone usage, and that during the company’s last quarter, iPhone sales were stronger in countries featuring its AI system Apple Intelligence. Cook, on Apple’s Jan. 30 quarterly call, said little in response to a question about DeepSeek’s impact, but he remained relatively positive.
“In general, I think innovation that drives efficiency is a good thing.”