Trending Now - Google Cloud CEO Discusses AI Infrastructure and Customer Needs

WHO:

Techmeme - Thomas Kurian is the CEO of Google Cloud.

Stratechery - Sundar Pichai is the CEO of Google.

Stratechery - Demis Hassabis is the CEO of DeepMind.

Stratechery - Koray Kavukcuoglu is the CTO of DeepMind.

WHAT:

Techmeme - Google is introducing a new ultra-low latency storage solution called Rapid Storage, capable of 15 terabits per second with microsecond latency.

Techmeme - Google is launching two new TPUs: TPU 8t, optimized for training with 9,600 TPU chips and three times better performance than the current generation, and TPU 8i, optimized for inference with 1,152 chips and a new Collectives Engine for efficient calculation performance.

Techmeme - Google Cloud's Gemini Enterprise has grown 40% sequentially quarter-over-quarter, with a current token generation rate of 16 billion tokens per minute, up from 10 billion tokens in January 2026.

Techmeme - Google is committing to invest $10 billion now in cash at a $350 billion valuation in Anthropic, with an additional $30 billion potentially to follow if performance targets are met.

Techmeme - Cisco's President and Chief Product Officer Jeetu Patel stated, "The biggest impediment to scaled adoption in enterprises for business-critical tasks is establishing a sufficient amount of trust."

Techmeme - Cisco aims for 70% of its products to be built entirely by AI by the end of 2027, according to Jeetu Patel.

TechRadar - Google introduced three new agents at Google Cloud Next, which are set to be released soon.

Techmeme - Google is introducing two new TPUs: TPU 8t, which consists of 9,600 TPU chips and has three times better performance than the current generation, and TPU 8i, which consists of 1,152 chips and features a new Collectives Engine for efficient calculation performance.

Techmeme - Google is committing to invest $10 billion now in cash at a $350 billion valuation in Anthropic, with an additional $30 billion potentially to follow if performance targets are met.

Techmeme - Google Cloud's Threat Intelligence Agent can reduce investigation times from 30 minutes to 30 seconds.

Techmeme - Google Cloud's Knowledge Catalog is designed to enhance data processing and improve the accuracy of information retrieval by mapping data across various systems and applications.

Techmeme - Google is introducing a new ultra-low latency storage solution called Rapid Storage, capable of 15 terabits per second with microsecond latency.

Techmeme - Google is committing to invest $10 billion now in cash at a $350 billion valuation in Anthropic, with an additional $30 billion potentially to follow if performance targets are met.

Techmeme - Cisco aims for 70% of its products to be built entirely by AI by the end of 2027, according to Jeetu Patel.

TechRadar - Cisco reported that 85% of enterprises are running AI agent pilots, but only 5% have moved those agents into production, highlighting a significant trust gap in the industry.

TechRadar - Cisco's AI Defense Explorer Edition is a free, self-service red teaming tool introduced to help organizations test their agent workflows before they reach production.

TechRadar - Google introduced three new agents at Google Cloud Next, which are set to be released soon.

Techmeme - Google is committing to invest $10 billion now in cash at a $350 billion valuation in Anthropic, with an additional $30 billion potentially to follow if performance targets are met.

Techmeme - Google Cloud's Threat Intelligence Agent can reduce investigation times from 30 minutes to 30 seconds.

Techmeme - Cisco aims for 70% of its products to be built entirely by AI by the end of 2027, according to Jeetu Patel.

TechRadar - Google is introducing a new ultra-low latency storage solution called Rapid Storage, capable of 15 terabits per second with microsecond latency.

TechRadar - Google is launching two new TPUs: TPU 8t, optimized for training with 9,600 TPU chips and three times better performance than the current generation, and TPU 8i, optimized for inference with 1,152 chips and a new Collectives Engine for efficient calculation performance.

TechRadar - Google Cloud's Gemini Enterprise has grown 40% sequentially quarter-over-quarter, with a current token generation rate of 16 billion tokens per minute, up from 10 billion tokens in January 2026.

TechRadar - Google Cloud's Threat Intelligence Agent can reduce investigation times from 30 minutes to 30 seconds.

TechRadar - Google is introducing two new TPUs: TPU 8t, optimized for training with 9,600 TPU chips and three times better performance than the current generation, and TPU 8i, optimized for inference with 1,152 chips and a new Collectives Engine for efficient calculation performance.

TechRadar - Google Cloud's Threat Intelligence Agent can reduce investigation times from 30 minutes to 30 seconds.

VentureBeat - Cisco's President and Chief Product Officer Jeetu Patel stated, "The biggest impediment to scaled adoption in enterprises for business-critical tasks is establishing a sufficient amount of trust."

VentureBeat - Cisco aims for 70% of its products to be built entirely by AI by the end of 2027, according to Jeetu Patel.

Techmeme - Google is committing to invest $10 billion now in cash at a $350 billion valuation in Anthropic, with an additional $30 billion potentially to follow if performance targets are met.

VentureBeat - Cisco reported that 85% of enterprises are running AI agent pilots, but only 5% have moved those agents into production, highlighting a significant trust gap in the industry.

VentureBeat - Cisco's AI Defense Explorer Edition is a free, self-service red teaming tool introduced to help organizations test their agent workflows before they reach production.

VentureBeat - Cisco aims for 70% of its products to be built entirely by AI by the end of 2027, according to Jeetu Patel.

TechRadar - Google introduced three new agents at Google Cloud Next, which are set to be released soon.

TechRadar - Google is launching a new ultra-low latency storage solution called Rapid Storage, capable of 15 terabits per second with microsecond latency.

VentureBeat - Cisco reported that 85% of enterprises are running AI agent pilots, but only 5% have moved those agents into production, highlighting a significant trust gap in the industry.

Stratechery - Google Cloud's Gemini Enterprise has grown 40% sequentially quarter-over-quarter, with a current token generation rate of 16 billion tokens per minute, up from 10 billion tokens in January 2026.

Stratechery - Google Cloud is introducing two new TPUs: TPU 8t, optimized for training with 9,600 TPU chips and three times better performance than the current generation, and TPU 8i, optimized for inference with 1,152 chips and a new Collectives Engine for efficient calculation performance.

Stratechery - Google Cloud's Threat Intelligence Agent can reduce investigation times from 30 minutes to 30 seconds.

Stratechery - Google Cloud's Knowledge Catalog is designed to enhance data processing and improve the accuracy of information retrieval by mapping data across various systems and applications.

Stratechery - Google Cloud is launching a new ultra-low latency storage solution called Rapid Storage, capable of 15 terabits per second with microsecond latency.

Stratechery - Google Cloud's Threat Intelligence Agent, powered by Gemini, can reduce investigation times from 30 minutes to 30 seconds.

Stratechery - Google Cloud generates 16 billion tokens per minute, which is a 40% increase since the last quarter.

Stratechery - Google Cloud's Gemini Enterprise has grown 40% sequentially quarter-over-quarter.

Stratechery - The new TPU 8t is composed of 9,600 TPU chips and has three times better performance than the current generation.

Stratechery - The new TPU 8i consists of 1,152 chips and features a new Collectives Engine for efficient calculation performance.

Stratechery - Google Cloud is introducing a new ultra-low latency storage solution called Rapid Storage, capable of 15 terabits per second with microsecond latency.

Stratechery - Google Cloud generates 16 billion tokens per minute, up from 10 billion tokens in early 2026.

Stratechery - Google Cloud's Gemini Enterprise and agent platform are being introduced to customers.

Stratechery - Google Cloud's capex investment is half directed towards Google Cloud.

Stratechery - Google Cloud is running the same stack as Google itself.

Stratechery - The new TPU 8t is optimized for training and has three times better performance than the current generation.

Stratechery - The new TPU 8i is optimized for inference and has a new Collectives Engine for efficient calculation performance.

WHERE:

Stratechery - Google Cloud services are utilized by various companies including Citigroup, Bosch, eBay, Virgin Voyages, Walmart, and the Food and Drug Administration.

WHEN:

Techmeme - The interview with Thomas Kurian was conducted on April 15th, 2026.

TechRadar - The interview with Thomas Kurian was conducted on April 15th, 2026.

Stratechery - The interview with Thomas Kurian was conducted on April 15th, 2026.

Stratechery - The annual keynote at Google Cloud Next occurs in April each year.

WHY:

TechRadar - The introduction of new agents and the integration of AI with cybersecurity technologies is aimed at addressing the increasing speed of cyber threats.

Stratechery - The introduction of new agents and the integration of AI with cybersecurity technologies is aimed at addressing the increasing speed of cyber threats.

Stratechery - Google Cloud aims to address the growing demand for AI capabilities in sectors like financial services and high-performance computing.

Stratechery - The shift to AI agents is driven by the need for efficiency improvements and productivity enhancements in organizations.

Stratechery - Google Cloud aims to protect critical information from cyber threats by integrating AI and cybersecurity technologies.

HOW:

Stratechery - Google Cloud's models, including Gemini, are designed to maintain long-running memory and interact with various tools and systems.

Stratechery - Google Cloud's Knowledge Catalog is designed to enhance data processing and improve the accuracy of information retrieval by mapping data across various systems and applications, allowing for better integration with existing enterprise data.

Stratechery - Google Cloud's Knowledge Catalog is designed to enhance data processing and improve the accuracy of information retrieval by mapping data across various systems and applications.

Stratechery - Google Cloud uses a world-class agent platform supported by advanced infrastructure to automate tasks and process flows.

Stratechery - Google Cloud's models, including Gemini, are designed to maintain long-running memory and interact with various tools and systems.

Stratechery - Google Cloud employs a Knowledge Catalog to enhance data processing and improve the accuracy of information retrieval.

(UPDATED April 25, 2026 05:33 UTC)

Google is making headlines with the launch of its new ultra-low latency storage solution called Rapid Storage, which boasts an impressive capacity of 15 terabits per second and microsecond latency. Additionally, the tech giant is introducing two new Tensor Processing Units, the TPU 8t for training and the TPU 8i for inference, both designed to significantly enhance performance in their respective areas. In a major investment move, Google is committing $10 billion in cash to Anthropic, with the potential for an additional $30 billion based on performance targets, while also reporting a 40% growth in its Gemini Enterprise service quarter-over-quarter.

(UPDATED April 25, 2026 03:53 UTC)

Google has announced the introduction of two new TPUs, the TPU 8t, which boasts three times better performance than the current generation with 9,600 chips, and the TPU 8i, featuring a new Collectives Engine with 1,152 chips. In addition, Google Cloud's Gemini Enterprise has seen impressive growth, increasing 40% quarter-over-quarter, while the company is also committing to invest $10 billion in Anthropic, with the potential for an additional $30 billion based on performance targets. Furthermore, Google Cloud's new Threat Intelligence Agent can significantly cut investigation times from 30 minutes to just 30 seconds, showcasing the company's ongoing commitment to enhancing data processing and retrieval accuracy.

(UPDATED April 25, 2026 01:21 UTC)

Google has announced a groundbreaking ultra-low latency storage solution called Rapid Storage, which can handle an impressive 15 terabits per second with microsecond latency. Additionally, the tech giant is launching two new TPUs, the TPU 8t for training and the TPU 8i for inference, both designed to significantly enhance performance in their respective areas. In related news, Cisco revealed that while 85% of enterprises are piloting AI agents, only 5% have fully deployed them, indicating a major trust issue in the industry, and they aim for 70% of their products to be AI-built by the end of 2027.

(UPDATED April 24, 2026 23:09 UTC)

Google has unveiled two new Tensor Processing Units, the TPU 8t and TPU 8i, with the former boasting three times the performance of the current generation thanks to its 9,600 chips, while the latter features a new Collectives Engine with 1,152 chips. In addition, Google Cloud's Gemini Enterprise has seen impressive growth, increasing by 40% quarter-over-quarter and now generating 16 billion tokens per minute. Meanwhile, Cisco's President Jeetu Patel highlighted the importance of trust in enterprise adoption and announced plans for 70% of their products to be AI-built by the end of 2027.

(UPDATED April 24, 2026 21:13 UTC)

Google has unveiled an ultra-low latency storage solution called Rapid Storage, which can handle an impressive 15 terabits per second with microsecond latency. Additionally, the tech giant is launching two new TPUs: the TPU 8t, designed for training with 9,600 chips and three times the performance of the previous generation, and the TPU 8i, optimized for inference with 1,152 chips and a new Collectives Engine. In a recent interview with Thomas Kurian April 15th, it was highlighted that Google Cloud's Gemini Enterprise has seen a 40% growth quarter-over-quarter, now generating 16 billion tokens per minute, while the new Threat Intelligence Agent can drastically cut investigation times from 30 minutes to just 30 seconds, addressing the rapid evolution of cyber threats.

(UPDATED April 24, 2026 19:17 UTC)

Google is making waves in the tech world by introducing two new Tensor Processing Units, the TPU 8t and TPU 8i, which promise significant performance boosts for training and inference tasks. In addition, Google Cloud's Threat Intelligence Agent is set to revolutionize security investigations by cutting down response times from 30 minutes to just 30 seconds. Meanwhile, Cisco's President Jeetu Patel highlighted the challenge of building trust for enterprise adoption and announced the company's ambitious goal of having 70% of its products developed entirely by AI by the end of 2027.

(UPDATED April 24, 2026 17:19 UTC)

Google is making a significant move by committing to invest $10 billion in Anthropic, which is currently valued at $350 billion, with the potential for an additional $30 billion if performance targets are met. Meanwhile, Cisco has revealed that while 85% of enterprises are piloting AI agents, only 5% have fully implemented them, pointing to a major trust gap in the industry, as noted by their President and Chief Product Officer Jeetu Patel. To address this issue, Cisco has introduced a free tool called the AI Defense Explorer Edition, designed to help organizations test their AI workflows before they go into production, with a goal of having 70% of their products built entirely by AI by the end of 2027.

(UPDATED April 24, 2026 15:19 UTC)

Google has unveiled three new agents at Google Cloud Next, with plans for their release in the near future. The tech giant also introduced an ultra-low latency storage solution called Rapid Storage, which boasts an impressive capacity of 15 terabits per second and microsecond latency. Meanwhile, Cisco's latest report reveals that while 85% of enterprises are testing AI agent pilots, only 5% have successfully implemented them, indicating a considerable trust gap in the industry.

(UPDATED April 24, 2026 13:20 UTC)

Google Cloud's Gemini Enterprise has seen impressive growth, increasing 40% quarter-over-quarter with a current token generation rate of 16 billion tokens per minute, up from 10 billion in January. The company is also launching two new TPUs: the TPU 8t, which boasts three times the performance of the current generation with 9,600 chips, and the TPU 8i, optimized for inference with 1,152 chips and a new Collectives Engine for enhanced calculation efficiency. Additionally, Google Cloud's Threat Intelligence Agent can significantly cut investigation times from 30 minutes to just 30 seconds, while the Knowledge Catalog aims to improve data processing and information retrieval accuracy across various systems.

(UPDATED April 24, 2026 11:18 UTC)

Google Cloud has announced impressive growth for its Gemini Enterprise, with a sequential increase of 40% quarter-over-quarter and a current token generation rate of 16 billion tokens per minute. The company is also introducing two new TPUs, the TPU 8t and TPU 8i, which promise significant performance improvements for training and inference, along with a new ultra-low latency storage solution called Rapid Storage, capable of 15 terabits per second. Additionally, Google Cloud's Threat Intelligence Agent, powered by Gemini, can dramatically cut investigation times from 30 minutes to just 30 seconds, addressing the urgent need for faster responses to cyber threats.

(UPDATED April 24, 2026 07:31 UTC)

Google Cloud has announced impressive advancements, generating 16 billion tokens per minute, which marks a 40% increase since the last quarter, while its Gemini Enterprise has also seen a 40% growth quarter-over-quarter. The company is rolling out new hardware, including the TPU 8t with 9,600 chips that boasts three times the performance of its predecessor, and the TPU 8i featuring 1,152 chips along with a new Collectives Engine for enhanced calculation efficiency. Additionally, Google Cloud is introducing Rapid Storage, an ultra-low latency solution capable of 15 terabits per second, aiming to meet the rising demand for AI capabilities across various sectors, supported by their Knowledge Catalog to improve data processing and retrieval accuracy.

(UPDATED April 24, 2026 05:51 UTC)

Google Cloud is making significant strides in the tech landscape, generating 16 billion tokens per minute and introducing its Gemini Enterprise and agent platform to customers, as highlighted in an interview with CEO Thomas Kurian conducted April 15th. The company is also launching new TPU models optimized for training and inference, which promise to enhance performance and efficiency for organizations using its services, including major players like Citigroup and Walmart. With a focus on integrating AI and cybersecurity technologies, Google Cloud aims to protect critical information while driving productivity improvements through its advanced infrastructure and world-class agent platform.