How cloud infrastructure has become the backbone of corporate artificial intelligence

Oliver Parker, Google Cloud’s global vice president of generative AI, said on Thursday, 24th, that corporate adoption of artificial intelligence relies on cloud infrastructure, with scale and security being key priorities over the next 12 months. The executive highlighted the integration of chips, infrastructure, models, and agents as a competitive advantage for Google.

While the tech industry celebrates recent advances in language models and the race for data centers, the foundation that supports all of this—cloud infrastructure—often remains in the background.

For Parker, Google Cloud’s global vice president of generative AI, this foundation has never stopped being essential; it has simply become less visible. “If you look at the current state of AI and the large language models, especially in companies adopting this capability, you’re actually accelerating cloud technology from below to make AI work,” Parker said during the Google Cloud Summit Brazil in São Paulo.

He compared the current moment to the arrival of cloud computing 10 or 15 years ago, when companies like Amazon, Microsoft, and Google itself began migrating existing applications to the cloud, making them easier to run and use.

“You’re taking an existing technology and enabling people to consume it in a much easier way,” he said. Today, according to Parker, AI follows the same logic. For him, generative AI systems “are only as good as the data platform” that supports them—and that platform is ultimately built on cloud technology.

Market perspective: AI providers and cloud hyperscalers

Parker divides the current market into two increasingly overlapping categories. On one side are the “intelligence providers”—companies like Google, Anthropic, and OpenAI, focused on language models. On the other side are the cloud infrastructure “hyperscalers”—Amazon, Microsoft, and Google, again.

Google, he said, is currently the only company operating fully on both fronts. “Our AI business is better because of our cloud business, and our cloud business is becoming better because of our AI business,” he said.

Vertical integration—from chip to application layer, through infrastructure, models, and agents—is cited by Parker as the company’s main competitive advantage. “People talk about the cloud as if it were yesterday’s thing, and they talk about AI as if it were only today, but if you look at all the AI providers, they all need cloud providers to support them at scale,” he said.

What has changed—and what hasn’t—in technology procurement

With over 15 years of selling cloud products before taking over the AI division, Parker sees both disruptions and continuities in how companies buy technology today.

The first change, he said, is speed. “I’ve never seen anything move so fast in terms of technological capability. From one week to the next, you can see a leap in frontier intelligence,” he said.

The second is the very nature of the technology. Unlike the internet or cloud computing, which Parker describes as “fundamental technological capabilities that enable things to happen,” generative AI would be “almost like a living system, that breathes and is always improving.”

What hasn’t changed, he said, is companies’ requirements. “Corporate customers still want to understand how to make the technology work. They need a supplier they can trust, who offers the right security and who helps with implementation,” he said.

According to Parker, the difference is that this implementation now has shorter deadlines. “The customer isn’t going to wait a year for a project. They want to see value in three weeks,” he said.

This reason led Google Cloud to expand its team of “Forward-Deployed Engineers”—engineers who work integrated with customers’ own technical teams—an announcement made by the company on Thursday, 24th, during the Summit.

Outlook for Brazil and the next 12 months

Regarding the growing interest of technology companies in Brazil, Parker recalled that Google and Alphabet have had an operation in the country for a long time—the company arrived in Brazil in 2005, with an engineering center in Belo Horizonte—but that the cloud unit only began operating locally in 2018.

“I think we saw the opportunity in Brazil, as well as the commitment to the region over a long period, unlike other new entrants in the market,” he said, referring to the more recent competition from other clouds in the country.

Marcel Silva, head of sales and AI market strategy for Latin America, pointed to three factors that make Brazil particularly favorable to AI adoption. The first is demographic: “Our population is very young, very digitally native. We adopt technology very, very quickly.”

The second is the smaller presence of legacy systems in Brazilian companies, compared to markets such as Europe and the United States, which reduces the complexity of integrating new technologies. The third, and most important, is the very nature of generative AI. “It’s a very inclusive technology, because you don’t need a lot of technical knowledge to use it—you interact through natural language, it’s much easier than 20 or 25 years ago, when you had to know how to program.”

Priorities for the next 12 months: scale with security

Over the next 12 months, the clearest sign that corporate adoption of AI has matured, according to Parker, will not come from a more advanced model, but from two much more practical things: scale and security.

When asked what he would consider a sign of genuine progress in this period, he pointed to these two axes, which apply both to Brazil and globally.

The first is scale: companies that already have successful point-use cases—in finance, back office, human resources—now need to take these initiatives across the entire organization.

The second, and most critical, is security. “With everything that’s happening around cyberattacks, AI agents, and security posture, I think this is going to be very important,” he said. “It’s scale, coupled with security. That’s how I think about the next 12 months”—a point that gains special weight, he said, in regulated sectors such as finance.

“Freedom of choice” as a business strategy

For Parker, when helping a company choose its AI strategy, the ideal is to use what he called a philosophy of “freedom of choice.” Google Cloud offers infrastructure, models like Gemini, and security tools, but does not force the customer to use only its own products at each layer. “Our main philosophy is that we will offer our customers freedom of choice. They will be able to use Gemini, they will be able to use various other tools and different models.”

This flexibility allows the company to close deals at different points in the chain: just infrastructure, for those who already have their own models; the complete platform, for those who want a variety of tools; or just the ready-made agents, for those looking for a more immediate solution.

“We can be partners at the silicon level, or we can be partners up there, at the top of the technology stack,” he said. “Not many people can say: I can offer you the infrastructure to build your own AI, or you can build on my managed platform, or you can just use my ready-made agents.”

Frequently Asked Questions

Why does artificial intelligence depend on cloud infrastructure?

Generative AI models and systems need cloud infrastructure to operate at scale. According to Oliver Parker, global vice president of generative AI at Google Cloud, these systems are only as efficient as the data platform that supports them.

What is Google’s competitive advantage in the AI market?

According to Oliver Parker, Google operates both as a language model provider and as a cloud infrastructure hyperscaler. The executive highlights the integration of chips, infrastructure, models, and agents as the company’s main competitive advantage.

What has changed in the way companies contract AI projects?

Companies now demand results within shorter deadlines, says Oliver Parker. According to the executive, corporate customers do not want to wait a year for a project and expect to see value in about three weeks.

What are Google Cloud’s Forward-Deployed Engineers?

Forward-Deployed Engineers are engineers who work integrated with Google Cloud customers’ technical teams. According to Oliver Parker, the demand for faster implementations led the cloud unit to expand this team.

Why does Brazil favor the adoption of artificial intelligence?

According to Marcel Silva, head of sales and AI market strategy for Latin America, Brazil has a young and digitally native population, companies with fewer legacy systems, and rapid technology adoption. Silva also says that interaction through natural language makes generative AI more accessible.

What will be the priorities of corporate AI over the next 12 months?

Scale and security will be the main priorities, according to Oliver Parker. The executive says companies will need to expand successful use cases across the entire organization and strengthen protection against risks related to cyberattacks and AI agents.