Apr 9, 2025 · 54m · big-technology
Google Cloud CEO Thomas Kurian on AI Competition, Agents, And Tariffs
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this in-depth interview, Google Cloud CEO Thomas Kurian discusses how Google is scaling enterprise cloud and AI adoption through custom TPUs, multi-model ecosystems, and autonomous agents. He addresses competitive dynamics against AWS and Microsoft, inference unit economics, real-world customer ROI, and hardware supply chain resilience.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Alex holds 25.9% of the talking time here. How this is scored →
Alex as informed peer 5.8 Guest teaching 4.7 Guest disagreement 2.3 Alex pushing back 3.8
speaking balance: gold is Alex, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 9:35 Kurian dismisses Amazon's positioning
Kurian directly takes aim at cloud rivals, bluntly arguing that competitors only claim Google pushes its own models because their own proprietary models are non-existent or terrible.
Hardest push from Alex ▶ 47:15 Kantrowitz drills into tariff exposure on hardware componentsAfter Kurian attempts to dodge policy talk, Alex refuses to let the issue drop, quoting Gavin Baker on semiconductor and server import vulnerabilities and asking if Google Cloud costs will increase.
Biggest teaching moment ▶ 15:55 Kurian dissects model research vs training costsKurian corrects common industry confusion highlighted by Mustafa Suleyman's quote, clearly delineating between exploratory frontier skill research and actual downstream training and inference runs.
Alex holds their own ▶ 14:40 Kantrowitz challenges Microsoft and OpenAI's partnership dynamicsAlex demonstrates deep domain expertise by dissecting the structural differences between Microsoft's arm's-length OpenAI relationship and Google's integrated in-house DeepMind architecture.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Alex as informed peer | Guest teaching | Guest disagreement | Alex pushing back | Why |
|---|---|---|---|---|---|---|
| Analyzing Google Cloud's AI-Driven Growth Architecture | 5 | 4 | 1 | 2 | Alex prompts Kurian on GCP's recent ~30% growth surge and the exact role AI plays in driving adoption. Kurian methodically categorizes customer entry points into infrastructure (TPUs), models/databases, and packaged agents without confrontation. | |
| The Varied Impact of AI Across Cloud Customer Segments | 5 | 4 | 2 | 4 | Alex pushes past AI start-up edge cases to question whether broad enterprise cloud purchasing decisions are truly predicated on AI. Kurian tempers hype by distinguishing high-impact sectors like retail from traditional utilities where AI remains secondary. | |
| Google's Open Model Strategy vs Cloud Competitors | 6 | 3 | 5 | 5 | Alex channels Amazon's competitive critique that Google forces its own proprietary models on users. Kurian forcefully rejects this premise, highlighting 200+ hosted models and sharply quipping that rivals say that only if their own models are terrible. | |
| DeepMind Integration and Enterprise Product Synergies | 5 | 5 | 2 | 3 | Alex asks what architectural advantage Google gets from having DeepMind in-house compared to Microsoft's partnership with OpenAI. Kurian explains tight feedback loops between pre-training, inferencing infrastructure, and domain tuning in Mandiant cybersecurity and Wendy's drive-thru ordering. | |
| Deconstructing Model Training, Research, and Inference Costs | 7 | 6 | 3 | 5 | Alex brings up Microsoft AI CEO Mustafa Suleyman's argument that rivals can cheaply copy frontier models without spending billions on pre-training. Kurian corrects the market confusion by differentiating exploratory frontier research from actual training and serving inference costs. | |
| Reasoning Architectures and Computational Trade-Offs | 6 | 6 | 3 | 4 | Alex cites Jensen Huang's claim that reasoning compute costs 100x more and asks if Kurian's numbers match. Kurian contextualizes the claim, explaining that real enterprise deployments time-bound and cluster-limit reasoning calculations depending on latency constraints. | |
| Open Source Resilience, AgentSpace, and Enterprise Search | 6 | 5 | 2 | 3 | Alex queries whether the rise of open-source models like DeepSeek threatens to commoditize cloud vendors. Kurian draws a historical parallel to Kubernetes, arguing that orchestration, serving performance, and application layer tools like AgentSpace remain heavily differentiated. | |
| Google Workspace AI Strategy: Driving Daily Habit Formation | 5 | 4 | 1 | 2 | Alex asks why Google bundled Gemini directly into Workspace seats rather than monetizing it as a standalone add-on like Microsoft Copilot. Kurian explains the behavioral psychology of habit formation and feedback data flywheels drawn from their 2014 autocomplete rollout. | |
| Evaluating Enterprise AI ROI Against Consumer Skepticism | 6 | 6 | 3 | 5 | Alex challenges Kurian with media criticism labeling generative AI 'mid' and failing to deliver consumer value. Kurian responds with concrete enterprise metrics, citing AES cutting audit times from 14 days to one hour and Verizon reaching 96% accuracy. | |
| Operationalizing AI Agents and Multi-Agent Collaboration | 5 | 5 | 1 | 2 | Alex asks Kurian to demystify agent buzzwords and explain multi-agent orchestration. Kurian breaks down single versus multi-agent architecture using an automated mobile trade-in and retail appointment scheduling workflow. | |
| Supply Chain Resilience, Tariffs, and Hardware Infrastructure | 7 | 3 | 4 | 8 | Alex pushes hard on new hardware tariffs using investor Gavin Baker's thesis that tariffs will cripple US AI datacenters, pressing repeatedly on component costs. Kurian repeatedly declines to comment on policy and maintains tight discipline regarding confidential supply chain mitigations. | |
| Scaling Google Cloud's Enterprise Go-to-Market Engine | 6 | 5 | 1 | 2 | Alex asks how Kurian transformed Google Cloud from an engineering-heavy organization with weak sales into a $40B+ enterprise business. Kurian explains the multi-year rebuild of enterprise sales compensation, expanding the partner ecosystem to 100,000, and hybrid on-prem delivery. |
Statements from this episode (24)
Ford uses Google Cloud TPUs for digital wind tunnel simulations
“Ford Motor Company, for example, when they brought their, ah, they wanted to use our chips and our system called TPU, Tensor Processing Unit, to model air flow and wind tunnel simulation using Computers rather than physical wind tunnels, so they're doing that …”
Over two million developers build daily on Google Cloud's AI platform
“We have over two million developers building every day, every morning, every night using our AI platform”
Kurian: OpenAI is the only major model Google Cloud does not offer
“The only model we don't offer today is OpenAI, and that's not because we don't want to offer their model.”
Google Cloud sales reps receive no extra pay for selling DeepMind models
“Our field is not compensated any differently. Our partner ecosystem is able to use all the models in the platform, and most importantly, we have very large Anthropic customers running on GCP.”
Kurian: Competitors attack Google because their own AI models are terrible
“If you don't have your own model, or you have a model of your own, but it's terrible, naturally you're going to say something like that.”
Kurian questions Microsoft's AI contributions beyond supplying OpenAI with GPUs
“Whether that's how much of credit goes to Microsoft outside of providing them a bunch of GPUs, time will tell.”
Kurian: Google Cloud releases DeepMind models to developers within hours
“In fact, we're staging models out to the developer ecosystem within a matter of a few hours after they're finally built.”
Kurian: Google Search, Cloud, and YouTube share the same inference stack
“And one benefit we have at Google is all our services, whether that's search or us or YouTube, this inferencing of the same stack and same model series. So the model learns very quickly from all that reinforcement learning feedback and gets better and better.”
Kurian: Long-term AI economics depend primarily on inference cost, not training
“First and foremost, in the long run, if AI really scales, the cost you really want to care about is inference cost, because that's what's integrated into serving, and any company that wants to recover the cost of training has to have a large scale inference fo…”
Kurian: AI model pre-training will continue to see diminishing returns
“There are gains to be had. I don't think they will be at the same ratio as earlier because just, you know, there's always a lot of diminishing returns at some point. I don't think we are at the point where there are no more gains, but I think we won't see the …”
Kurian: Google Cloud reduced AI inference costs 20x in 2024
“Like if you look at just 2024, we've reduced the cost of inferencing, and you can see it in our prices of the models by a factor of 20 times.”
Kurian: AgentSpace is Google Cloud's fastest-growing product ever
“Yes.
We're very proud of it, yes.”
Kurian: Google bundled Gemini into Workspace instead of selling separate subscriptions
“We made Gemini part of Google Workspace. Rather than requiring somebody to buy a separate subscription.”
Kurian: Separate AI add-on subscriptions prevent daily habit formation for employees
“It requires people to change the way they work, and we want to drive daily usage of AI, and because it change, needs to change the way they work, you want them to get used to using it. If, hey, this group of users in a company gets it, that group of users is n…”
Kurian: Google's AI Assistant for Verizon Call Centers Achieves 96% Accuracy
“We've helped them build something called a personal research assistant, so that if I am a call center person, and you call me saying, here is my set of issues, and we can, how long does it take to research that information and put it back in front of you, so t…”
AES cut end-of-quarter audits from 14 days to one hour using AI
“A company called AES. It's a energy utility. It's an energy company. It builds you know, and delivers energy different parts of the world. It used to take them 14 days to run their end of quarter audit. They do it in one hour now.”
Kurian: Google built an open agent-to-agent protocol and dev kit
“So what we've done at Google is build an agent development kit, which has an API through which you can, one, create agents. We provide you a tool set to do it. We provide you a set of tools that these agents can use, but we also have an open agent to agent pro…”
Kurian: No enterprise software company grew faster than Google Cloud since 2019
“And to grow from the scale we were in 2019 to where we are now, No other enterprise software company has grown that fast, and that's a credit to our sales organization.”
Kurian: Google Cloud avoided building professional services to attract partners
“We made a decision early on. We're not going to have a big professional services organization specifically so that we can attract the partner community.”
Kurian: Google Cloud grew from 1,000 partners in 2019 to 100,000 today
“In 2019, we had about a thousand partners. Today we have a 100,000.”
Kurian: Cloud adoption will definitely surpass 50% of IT workloads
“We definitely see it getting north of 50%.”