Tech earnings reports show mega-cap companies thriving on AI

Tech investors want to know how much the industry leaders are boosting profitability now that they’re in cost-cutting mode.

But there is one area where they want to invest heavily. Artificial intelligence.

Alphabet, Microsoft, Amazon and Meta all reported quarterly results this week, updating Wall Street on efforts to improve efficiency amid growing economic concerns. When it comes to AI and the recent boom in so-called Large Language Models (LLMs) powering products like ChatGPT, megacap tech companies can’t afford to be left behind.

Generative AI programs use large amounts of data and processing power to produce human-looking output (blocks of text, snippets of code, or computer-generated images). It requires a specialized supercomputer, which is not cheap.

On this week’s earnings calls, tech CEOs delved into the potential of AI, whether they’re building their own models or rapidly integrating them into their products. A common theme was to highlight the high cost of building and running these applications.

Here’s what executives from Alphabet, Microsoft, Amazon and Meta told analysts:

Alphabet CEO Sundar Pichai sees intense pressure to offer AI products as sophisticated chatbots hit the market and recognize the threat facing the company’s core Google search engine exposed to The company recently declared an internal “Code Red”.

Pichai said on Tuesday’s earnings call that the company is “on track” toward its AI goals.

“We will continue to embrace advances in generative AI to improve search in thoughtful and purposeful ways,” said Pichai.

He said Google uses AI to improve ad conversion rates and reduce the amount of “toxic text” that goes into AI models. The company also integrates his two main AI teams, Brain and DeepMind.

Not only does it use its own homegrown chip to drive the model, Pichai says it uses processors from Nvidia, which uses the majority of its graphics chips for training and deploying cutting-edge AI.

Microsoft uses OpenAI’s GPT technology in its Bing search engine, Office, and Teams conferencing system.

CEO Satya Nadella says AI will ultimately drive revenue growth and is already seeing increased usage in the company’s apps. For example, he says that Bing downloads have quadrupled since Microsoft added chatbots. Through Microsoft’s integration with Bing he generated over 200 million images.

Nadella warned that building the massive data centers needed to run AI applications would require a lot of capital.

“We will continue to invest in cloud infrastructure, especially AI-related spending, to meet the growing demands of our customers’ transformation,” Nadella said. “And we expect the resulting revenue to increase over time.”

Amazon CEO Andy Jassy gave an unusually lengthy answer to analysts’ questions about the company’s generative AI plans on Thursday.

Jassy emphasized that the market is huge, saying Amazon is building its own LLM and designing data center chips for machine learning.

“These large-scale language models, generative AI capabilities, have been around for a while. But frankly, models weren’t attractive until about six to nine months ago,” says Jassy. said Mr. “They have gotten bigger and better so quickly that they really offer an amazing opportunity to transform virtually every customer experience that exists.”

Jassy also said Amazon’s scale allows it to be one of the few companies to build an LLM.

“There are a handful of companies that want to invest that time and money, and we’ll be one of them with Amazon,” Jassy said.

Unlike Microsoft and Google, Amazon’s focus is on selling access to its technology through its Amazon Web Services division. However, Jassy said Amazon will work on some applications, such as programs that help engineers write code.

“Every business within Amazon is built on a massive language model to reinvent the customer experience,” says Jassy. That includes voice assistant Alexa, he said.

After turning his attention to the Metaverse in late 2021, Meta CEO Mark Zuckerberg has tried to dispel the notion that his company is no longer focused on the Metaverse.

But he wanted investors to know that while Meta invested in Metaverse technology, it could also devote significant resources to AI.

Zuckerberg said that while the company has used machine learning to provide recommendations and has strong products such as Facebook’s news feed and advertising system, its new main focus is its generative foundation model.

“It’s been an amazing year of progress in this space, and the work we’re doing now will impact all of our apps and services,” Zuckerberg said.

The company has used the technology in a variety of products, including chat experiences in WhatsApp and Facebook Messenger, tools to create images for Facebook and Instagram posts, and eventually programs that spew full videos from short descriptions. He said he plans to work on it.

A concept he’s particularly excited about is “AI agents.” This often refers to an AI program that can carry out a goal.

“There is an opportunity to introduce AI agents to billions of people in a useful and meaningful way,” Zuckerberg said. One potential for AI agents is to handle customer service for businesses. Meta said.

Zuckerberg spoke about the company’s large investment in building data centers for AI applications. He said technology has been the “main driver” of Meta’s capex growth over the past few years.

“At the moment, we are not far behind building AI infrastructure,” said Zuckerberg.

This does not mean that Meta has stopped buying graphics processors. Zuckerberg said the company needs to “continue to invest,” but will do so after it launches its generative AI product and has a better understanding of the resources it needs.

clock: amazon big beat