Spend money to make money
Google CEO Sundar Pichai arrives for the US Senate’s bipartisan Artificial Intelligence (AI) Forum at the US Capitol in Washington, DC, on September 13, 2023.
Mandel Ngan | AFP | Getty Images
“We are moving from talking about artificial intelligence to implementing artificial intelligence at scale,” Microsoft CEO Satya Nadella said at his company’s earnings call on Tuesday. “By integrating AI into every layer of our technology, we are winning new customers and helping realize new benefits and productivity gains.”
Last year saw the beginning of the generative AI boom, as companies raced to include increasingly sophisticated chatbots and assistants across mainstream products. Nvidia was the big money maker. Its graphics processing units, or GPUs, are at the heart of the big language models created by OpenAI, Alphabet, Meta, and a growing array of heavily funded startups all vying for a slice of the generative AI pie.
As 2024 approaches, and CEOs outline their plans for continued investment in AI, they are making their strategies more clear to investors. One key priority area, based on the latest earnings calls, is AI-as-a-service models, or large AI models that customers can use and customize to their needs. Another way is to invest in AI “agents,” a term often used to describe tools ranging from chatbots to programming assistants and other productivity tools.
Overall, the executives emphasized the idea that AI is no longer just a toy or a concept for research labs. It’s here for real.
At larger companies, two huge areas of investment are AI initiatives and the cloud infrastructure needed to support massive workloads. To that end, cost reductions will continue in other areas, a message that has become familiar in recent quarters.
Meta CEO Mark Zuckerberg on Thursday emphasized the company’s ongoing efforts in artificial intelligence along with broader cost cuts.
Meta founder and CEO Mark Zuckerberg speaks during a Meta Connect event at Meta headquarters in Menlo Park, California on September 27, 2023.
Josh Edelson | AFP | Getty Images
“2023 was a ‘Year of Efficiency’ focused on making Meta a stronger technology company and improving our business to give us the stability to deliver on our ambitious, long-term vision for AI and transformation,” Zuckerberg said on the earnings call.
Nadella told investors that Microsoft is committed to expanding investment in its AI and cloud efforts, even if that means looking more closely at expenses in other divisions, with “disciplined cost management across every team.”
Microsoft CFO Amy Hood emphasized “consistency in redirecting our workforce toward the AI-first work we do without adding a material number of people to the workforce,” and said the company will continue to prioritize investment in AI as “ The thing that happens.” To shape the next decade.”
It was a similar theme at Alphabet, where Sundar Pichai spoke of his company’s “focus and discipline” as it prioritizes scaling AI for search, YouTube, Google Cloud and beyond. Investing in infrastructure such as data centers is “key to achieving our big AI ambitions,” he said, adding that the company has cut non-priority projects and invested in automating certain processes.
“We continue to invest responsibly in our data centers and compute to support this new wave of growth in AI-powered services for us and our customers,” Pichai said. “You’ve heard me talk about our efforts to constantly re-engineer our cost base and improve our speed and efficiency. This work continues.”
Within Google Cloud, Pichai said the company will cut expenses by reallocating resources to the most important projects, slowing the pace of hiring, improving technical infrastructure and using artificial intelligence to streamline operations across Alphabet. Capital expenditures, which reached $11 billion in the fourth quarter, were largely due to investment in infrastructure, servers and data centers, he said.
Alphabet CFO Ruth Porat confirmed that the company expects full-year capital expenditures for 2024 to be “significantly larger than 2023,” as it continues to invest heavily in AI and the “long-term opportunity” presented by AI applications within DeepMind. Advancement of cloud and other systems.
Generative AI “will ultimately drive tens of billions of dollars in revenue for Amazon over the next several years,” Amazon CEO Andy Jassy said on an earnings call this week.
AI will remain a big investment area for the company, leading to an increase in capex this year as Amazon pours more money into LLMs, other generative AI projects, and needed infrastructure. Jassy emphasized Amazon’s efforts in AI chips, naming clients such as Anthropic, Airbnb, Hugging Face, Qualtrics, and Snap.
Apple CEO Tim Cook has cited generative AI as an important investment area for his company, teasing an announcement later this year.
“As we look to the future, we will continue to invest in these and other technologies that will shape the future,” Cook said during a call with analysts. “This includes artificial intelligence where we continue to spend a tremendous amount of time and effort, and we are excited to share details of our continued work in this area later this year.”
“Let me just say that I think there is a huge opportunity for Apple with the new generation of AI and artificial intelligence, without going into too much detail and getting ahead of myself,” Cook added.
While investors want to see investments in AI by companies that play a key role in providing infrastructure, they also want to know where and how the money is being made.
Enterprise customers are looking to use existing models that they can customize and build upon, Jassy said, pointing to Amazon’s Bedrock as a key focus.
“What we’re seeing is that customers want choice,” Jassy said. “They don’t want just one model to rule the world. They want different models for different applications. And they want to try all the different sized models because they produce different cost structures and different latency characteristics.”
Andy Jassy on stage at the 2022 New York Times DealBook in New York City, November 30, 2022.
thos robinson | Getty Images
Nadella pointed to Microsoft Azure as the dominant “model as a service” offering, emphasizing that customers do not have to manage the underlying infrastructure but have access to a range of large and small language models, including some from Cohere, Meta, and Mistral, and the options are open. Source. A third of Azure AI’s 53,000 customers joined in the past 12 months, Nadella said.
Alphabet executives highlighted Vertex AI, a Google product that offers more than 130 generative AI models for use by developers and enterprise clients like Samsung and Shutterstock.
Gossip isn’t limited to master’s degree holders and chatbots. Many tech executives have spoken about the importance of AI agents, or AI-powered productivity tools, to get things done.
Eventually, AI agents could take the form of scheduling a group hangout by scanning everyone’s calendar for conflicts, booking travel and activities, buying gifts for loved ones, or doing a specific job function like outside sales. However, tools are currently largely limited to tasks such as summarizing, creating to-do lists, or helping with writing code.
Nadella is bullish on AI agents, pointing to Microsoft’s Copilot Assistant as an example of a “cutting-edge” AI application in terms of productivity benefits and a successful business model.
“You’ll start to see people thinking of these tools as productivity enhancers,” Nadella said. “I see this as a new vector for us in what I will call the next phase of knowledge work and front-line work, even in their productivity and how we engage.”
Just before Amazon’s earnings, the company announced Rufus, a creative AI-powered shopping assistant trained on the company’s product catalog, customer reviews, user Q&A pages, and the broader web.
“The question of how we think about Gen AI in our consumer business: We’re building dozens of generative AI applications across the company,” Jassy said on the call. “Each of our companies has multiple AI applications that we are building. They are all at different stages, many have been launched and others are in development.”
Meta will also partly focus on building a useful AI agent, Zuckerberg said on his company’s call.
“Going forward, the main goal will be to build the most popular and most advanced AI products and services,” Zuckerberg said. “And if we succeed, everyone who uses our services will have a world-class AI assistant to help get things done.”
Alphabet executives praised Google’s Duet AI software, or “bundled AI agents” for Google Workspace and Google Cloud, designed to boost productivity and complete simple tasks. Within Google Cloud, Duet AI helps software developers at companies like Wayfair and GE, and cybersecurity analysts at Spotify and Pfizer, Pichai said. Duet AI will soon integrate Gemini, Alphabet’s LLM software that powers its Bard chatbot, he added.
Pichai wants to introduce an AI agent that can complete more and more tasks on behalf of the user, including within Google Search, though he said there is “a lot of implementation to come.”
“We will again use generative AI there, especially with our more advanced and cool models,” Pichai said. This “allows us to act as a proxy over time, if I think ahead and perhaps go beyond answers and follow up with users further.”
He watches: Megacaps return capital
(Tags for translation) Generative AI