Google Cloud Next 25 - CEO Thomas Kurian addresses Trump tariff uncertainty and reassures EU customers of data sovereignty

As geopolitical tensions rise and regulatory landscapes shift, Google Cloud CEO Thomas Kurian appears confident about navigating increasingly complex international waters. In comments made today at the vendor’s annual Google Cloud Next conference in Las Vegas, Kurian offered customers and the market assurances regarding the ongoing turmoil of international tariffs under the Trump administration, as well as addressed speculation regarding the potential restrictions on US cloud services in the European Union.

Kurian spoke candidly - yet diplomatically - about the volatile tariff landscape that has emerged since Trump's return to office in January. The administration's approach to tariffs has been, to put it kindly, notably inconsistent - initially threatening 60% tariffs on Chinese goods, then moderating the stance after market backlash, only to reassert aggressive positions weeks later. This pattern has extended to European trade relations, with proposed 10% blanket tariffs on EU goods followed by shifting exceptions for what Trump has termed "friendly countries" - a designation that appears subject to frequent revision.

With a knowing smile on his face, Kurian said: 

I think you all see that the tariff discussion is an extremely dynamic one.

But he did go on to compare the ongoing tariff situation to navigating global crisis’ in recent decades:

Obviously, we have an executive leadership team that has worked globally for many, many years. We run a global supply chain. We run a global distribution network. And we have been through many cycles like this, whether it's 2001, 2008, or when the COVID crisis happened in 2020.

While Kurian seemed keen to lay out a ‘stay calm and keep moving’ stance in Google's ability to navigate the tariff landscape, he remained cautious about making definitive predictions:

We are very confident that our executive team will find a path through it... I think we have to wait and see how this political environment and regulation evolves.

Google Cloud’s distributed global infrastructure might provide some insulation against tariff impacts, where its geographic diversification should help insulate it (and its customers) against some levels of uncertainty, However, it goes without saying that the fundamental challenge for Google Cloud - and indeed all US tech companies - is developing long-term strategy in an environment where trade policy lacks predictability or consistency. Kurian’s diplomacy is wise considering that the safest bet regarding trade under Trump is that uncertainty is the only certainty (at present). 

Hedging against EU-US data transfer challenges

Beyond tariffs, Kurian addressed another looming challenge: speculation that there is a potential collapse of the EU-US data transfer framework. Recent developments have raised serious questions about whether US cloud services could face restrictions in Europe.

The Privacy and Civil Liberties Oversight Board (PCLOB), a critical component of the Transatlantic Data Privacy Framework (TADPF), has been weakened by Trump's removal of Democratic members. This has cast doubt on whether US cloud providers, if certain measures are removed, can continue operating in Europe without running afoul of GDPR and other EU privacy regulations.

However, again, Kurian offered EU-based customers assurances on any situation that may unfold:

In order to serve Europe on Europe's terms, we invested in sovereign solutions much earlier than competition. I think you will see the maturity of our sovereign solutions is yet another indication that we take European governments and their regulations and their needs [seriously].

When pressed on the possibility of an executive order that might prevent Google from providing services to the EU, Kurian outlined Google's technical safeguards:

Technologically, we have solutions to protect against [an executive order that would prevent us from providing services to the EU].

These solutions include giving customers complete control over data location and encryption. He added:

When you use any of our regions in Europe, you have sole control of the location of your data. You can encrypt it with your encryption key, as a customer. You can keep the encryption key away from Google in an off-site location you have. We don't have access to your environment,

If we get a request to hand over a customer's data, you can deny us access to it, because you have the encryption key, and we do not have the ability to even reach the encryption key.

For organizations with the highest security requirements, Google offers even more isolation:

We also offer something for people who want to put the workload in the cloud but are worried about long term survivability. That's what we mean by Google Distributed Cloud, air-gapped, where it runs fully detached. There is no connection to Google and there is no connection to the internet.

We also have sovereign partnerships with European organizations, where a European entity does a supervisory role on top of our operations, so they can trust and verify our operations. They can also be the custodian of the encryption keys.

Under Trump, it’s clear that organizations in the EU can’t rely on Biden-era security policies - the TADPF faces significant uncertainty. Biden’s executive orders on the matter offered reassurances to the EU that European data would be handled appropriately, especially in a post-Snowden world, but should this framework collapse, EU businesses would face difficult choices: migrate to European alternatives, risk substantial GDPR penalties, or partner with US providers that have developed genuine sovereign capabilities. And it’s important to note that executive orders aren’t law and Trump could easily overwrite what came before - and could be used as a bargaining chip as relations between the EU and the US become increasingly hostile. 

However, Kurian doesn’t seem worried. He said:

As geopolitical environments and national regulations evolve, people evolve to meet those requirements.

Agent orchestration in the enterprise

Beyond navigating regulatory challenges, as we outlined yesterday, Google Cloud is making a strategic bet on AI agent orchestration as a core differentiator in the enterprise market. Kurian detailed the company's vision for multi-agent systems that position Google Cloud as an essential orchestrator of diverse enterprise AI ecosystems.

Central to this strategy is Google's Agentspace, a platform designed to address agent-to-agent integration challenges in the enterprise. Kurian said today:

Agentspace came out of our observations that many organizations spend a lot of time looking for information and trying to get tasks done. I have SharePoint, I have Office, I have my Oracle applications, I have Salesforce, and I'm spending all my time trying to find information about my supply chain, my customers, and my internal benefits system.

Agentspace offers three core capabilities, according to Kurian:

Number one, you can use Agentspace to have Google search for your company, access these systems and make them searchable and discoverable. Number two, you can integrate that into a conversational AI system, so that the AI can help you do research on this information, summarize it and give you answers to complicated questions. And third, you can then ask an agent to do tasks on your behalf.

To facilitate adoption, Google has developed extensive integration capabilities. He also said that whilst good progress has been made around building a coalition around a certain set of standards, there are more companies signing up in the pipeline:

People then asked us, can you provide connectors to all these systems so it's easy to discover them? We already have 100 connectors live. There's another 300 under development. That makes it easy for people to adopt the technology without having to rip or replace anything you have.

The agent-to-agent protocol is a key component of Google's strategy, enabling different AI systems to work together seamlessly:

We designed it to give it to the community, and it's supported by lots of partners. And I think that is the approach, I think in the end, that will be successful. It's very similar to what we did with Kubernetes years ago.

Addressing AI's environmental impact

Kurian also addressed growing concerns about AI's environmental footprint, where he was keen to highlight Google's efforts to reduce energy consumption while improving model performance:

First and foremost, we have done a lot of work these last two years to hugely reduce the amount of cost required for both the training and inferencing of models. If you look at just inferencing since January 2024, not even 2023, the cost of inferencing has dropped by more than 20 times.

These comments challenge the assumption that AI's growth will necessarily lead to proportional increases in energy consumption. The public perception is that AI will lead to negative outcomes for the environment; but Kurian had a different take:

A lot of times, people take the size of a model and the cost of the model and then, say, if the usage is going to grow 20 times, you will need 20 times more energy - without realizing that within even a 12 month period, we have reduced it by 20 times, so you don't need as much energy.

We introduced water cooling many years ago for our processors. The new Trillium version that we announced yesterday is a water cooled system. But our earlier systems are also water cooled. Those provide another 35% to 40% more energy efficiency when you're actually serving models. And just to give you an example, we have more than seven times the water cooled AI systems as the rest of the world combined.

Beyond hardware optimizations, Google Cloud is also investing in more sustainable energy sources. Kurian said: 

We are investing in new AI energy technologies. For example, recently we introduced, and we are working with a bunch of organizations for nuclear power. If you look at several of the locations in Europe, we have sustainable energy sources: wind, hydroelectric, solar.

There are several regions, even with the increased AI demand, that are higher than 95% sustainable energy. For us, it's really important that people see AI as being a technology that can also drive both efficiency in the consumption of energy and also create new forms of energy because of the investments that we're making.

My take

Kurian's message at Google Cloud Next 25 could be summed up as "Keep calm and cloud on," but the subtext reveals how technology companies are carefully hedging their bets in an unpredictable geopolitical landscape. His diplomatic answers about Trump's tariff whiplash acknowledge the chaos while essentially saying, "We've seen worse, we'll figure it out."

The comparison to previous crises was telling though, highlighting the scale of impact these trade wars could have on organizations. And despite Kurian’s steady hand, it’s fair to say that no one can predict which way the Trump trade winds will blow next week, let alone next quarter. For customers, this means planning for flexibility rather than certainty.

And on Kurian’s comments regarding the potential collapse of the EU-US data relationship, the company’s early investments in sovereignty solutions now look less like regulatory compliance and more like strategic foresight.

With all that being said, Google Cloud has been clear this week that it’s leading with customers front of mind. Its approach to agentic AI addresses the interoperability and ‘garden wall’ challenges faced by organizations as they make decisions regarding vendor choices, whilst offering buyers choice and protection in an increasingly turbulent economic and political environment. That bodes well for the company, in my opinion.