Google talks up AI after disappointing financial results

Google talks up AI after disappointing financial results

'I feel comfortable with all the investments we've made in making sure we can develop AI responsibly,' says Pichai

Google's rough performance in advertising didn't mask gains made in the more partner-friendly areas of artificial intelligence and cloud, the vendor's parent company Alphabet reported Thursday during its latest quarterly earnings call.

The company missed its earnings per share estimates by around 15%, and revenue growth was below expectation at 7%.

CEO Sundar Pichai told analysts on the call - which covered earnings for the fourth quarter of Alphabet's fiscal year, which ended 31st December, that AI is "such a foundational technology" and to not count Google out as a leader in the emerging market.

Google is "investing not just in terms of research, but actually getting it all production-scale ready," Pichai said on the call. "We do see secular opportunities ahead, both in terms of putting these APIs out, making sure every developer, every organisation in the world can use it."

Google's AI investments

Google has flexed its AI muscle with deployments to its flagship Search offering, its AI-powered Google Lens technology for smartphone cameras and its Vertex AI machine learning training and deployment platform, Pichai said.

Google has "been investing for a while, and it's clear that the market is ready" for AI, Pichai said. Serving costs need to improve, but the field is still in "very, very early days."

"Consumers are interested in trying out new experiences," he said. "I feel comfortable with all the investments we've made in making sure we can develop AI responsibly. And we'll be careful. We'll be launching … labs products in certain cases, beta features in certain cases. And just slowly scaling up from there."

He continued: "We are committed to putting out experiences - both in terms of new products and experiences actually bringing direct LLM (large language models) experiences in Search and making APIs available for developers and enterprises."

Shots at Microsoft?

Pichai's comments on Google fighting for AI supremacy come after Microsoft detailed its own bets on AI during its latest quarterly earnings call, held in January.

Microsoft has made early splashes with its investment in OpenAI - the organisation behind the ChatGPT program for generating text content and the Dall-E 2 program for generating images - not to mention the GitHub Copilot tool for programmers, which Microsoft CEO Satya Nadella deemed "the first at-scale AI product built for this era" with more than a million users to date.

Without naming Microsoft or OpenAI, Pichai said that Google research from 2017 and its diffusion models are "the basis of many of the generative AI applications you're starting to see today.

"Our talented researchers, infrastructure and technology make us extremely well-positioned as AI reaches an inflection point," he said.

The "coming few months" will see more information from Google around increasing access to its Language Model for Dialogue Applications (LaMDA) and its Pathways Language Model (PaLM).

"In the coming weeks and months, we‘ll make these language models available starting with LaMDA so that people can engage directly with them," he said. "This will help us continue to get feedback, test and safely improve them."

Google also has industry-specific AI offerings in manufacturing, life sciences and retail, with others to come, Pichai said. And users of Google Workspace have access to AI-powered features such as smart canvas and smart compose.

Google is working to bring LLMs used for AI to its Gmail and Docs applications, he said. "We are just at the beginning of our AI journey, and the best is yet to come."

Cloud's path to profitability

While Google's AI investments are still a way from making a splash, its cloud segment saw sizable revenue for the quarter - even though the segment is still not profitable.

Google Cloud brought in $7.3 billion during the quarter, an increase of about 30% year over year. The unit reported an operating loss of $480 million for the quarter, but that is about half the loss experienced a year prior.

Alphabet did not break out revenues for Google Cloud Platform (GCP), but Ruth Porat, chief financial officer for Alphabet, told analysts on the call that GCP revenue was greater than Google Cloud, "reflecting strength in both infrastructure and platform services."

Akin to comments made by Microsoft executives during its quarterly earnings, Porat said that GCP saw slower consumption growth in the fourth quarter "as customers optimise GCP cost reflecting the macro backdrop."

Alphabet did not break out revenues for Google Workspaces - the enterprise applications suite that competes with the likes of Microsoft's Office 365 - but Porat said that the segment delivered "strong results" that "were driven by increases in both seats and average revenue per seat."

Porat said that Google Cloud is still unprofitable because Google is "investing ahead of our revenues given the growth in the opportunity overall and the desire to ensure that we're equipped, able to support customers across segments around the globe."

The vendor is working to bring cloud to profitability - and to improve Alphabet's margins overall - in part by re-examining spend with suppliers and vendors, looking at costs associated with its hardware and Pixel offerings and adopting AI and automation internally to save on costs.

Alphabet is trying to lessen the cost of compute, data centres, servers and supply chain to keep margins up. Plus, Google will save money as it sheds physical office space and slows down on hiring, Porat said.

"In cloud, we've remained very focused on the path to profitability," she said. "That's a revenue and margin driver. And then with other bets, we‘re solely focused on investing sustainably."

When asked if Google has reached the end of low growth in its ad business, Porat stopped short of saying the worst is over.

"We're not going to predict the global environment," she said. "We did say the challenging backdrop is ongoing. … But we're very focused on what we can control. And I think most important and what we're really excited about here is innovation to help advertisers overall and our cost reengineering that really gets us to align this long-term, sustainable value creation.

Layoffs cost Alphabet billions

On Alphabet's headcount, the company expects the recently announced layoffs of about 12,000 people to cost between $1.9 billion and $2.3 billion in severance and related costs. Google will pay most of that during the first quarter of 2023.

"I'm grateful to the Googlers leaving us for all of their contributions and their hard work to help people and businesses everywhere," Pichai said.

Alphabet finished the year with about 190,000 employees, a 20% increase from the headcount at the end of 2021.

Porat said that Google will still hire "in priority areas with a particular focus on top engineering and technical talent as well as on the global footprint of our talent."

Reducing its office space will cost the company $500 million during the quarter, with additional charges possible later in the year.

Other updates

Pichai told analysts on the call that integration of recent cybersecurity acquisition Mandiant continues. Mandiant has more than 1,800 customers, he said.

He called the 2022 Pixel 6a, 7 and 7 Pro "the best-selling generation of phones we've ever launched, and we gained share in every market we operate in this year."

Alphabet brought in $76 billion in consolidated revenue for the quarter, an increase of 7% year over year, ignoring foreign exchange.

The bulk of Alphabet's revenue still comes from online advertising, particularly from Google Search. The ad business brought in $59 billion during the quarter, down about 4% year over year.

For the full year in 2022, Alphabet brought in $283 billion, up 14% year over year ignoring foreign exchange.

Alphabet traded at $103.80 a share Thursday afternoon, down about 4% from market close.

This article first appeared on Computing's sister site CRN.