Did you think Artificial Intelligence was just an abstract cloud of code and data? Think again. The war for AI is no longer being fought on whiteboards, but in the mud of civil engineering sites and ultra-high-voltage power grids. 🏗️⚡

In 2026, the intangible has never been so heavy. In this episode of The Last Men, we decode Mistral AI's radical industrial pivot. The European flagship, once a champion of "lightweight" models and renting servers, has just raised $830 million in debt to buy 13,800 Nvidia chips and build its own data center south of Paris.

Why borrow such an amount from traditional banks (BNP, HSBC) for computer hardware that is supposed to become obsolete in 18 months? Discover how computing power has become a rent comparable to oil, and why the geography of servers is now a massive geopolitical weapon against American extraterritorial laws (Cloud Act).

From the strategic acquisition of Koyeb (the "nervous system" to kill latency) to the staggering announcement of a 1.4 gigawatt AI mega-campus (the equivalent of a nuclear reactor!), dive into the new era of "Infrastructure Capital".

🎙️ Key Takeaways from the Episode:

* Mistral AI's Industrial Pivot: No more renting computing power from Microsoft or Google. Mistral is buying its own fleet of "heavy trucks" (13,800 Nvidia chips) to build its own garage (Data center in Bruyères-le-Châtel).

* The Debt Gamble ($830M): Why raise debt instead of equity? To avoid diluting capital. And why do banks accept? Because global demand is such that the chips will run at 100% capacity from the first second, generating astronomical profitability before they become obsolete.

* The Anti-Cloud Act Shield: The sovereignty issue. If a server belongs to an American company (even if it's located in Paris), US justice can demand access to the data. Mistral now sells the absolute guarantee that European state and banking secrets will never leak to the United States.

* The Acquisition of Koyeb (The Anti-Latency): Running one chip is easy. Making 13,800 chips communicate with each other without latency is an engineering nightmare. Koyeb is the essential software to orchestrate this massive flow.

* The 1.4 Gigawatt Mega-Campus: The shock announcement with the Emirati sovereign fund MGX. A titanic project requiring the power of a nuclear reactor, proving that AI has officially become heavy industry.

* The End of the "Garage Startup" Myth: The entry ticket for disruptive innovation is now measured in billions of dollars and gigawatts. Small structures can no longer keep up.

💡 Final Thought:

If the next generation of AI requires the energy of a nuclear reactor and billions of dollars of hardware confined in hyper-secure fortresses, can the decentralized and "Open Source" utopia really survive? Tomorrow's game masters will no longer be those who code the best algorithms, but those who hold the keys to the silicon fortresses and the switches to the electrical grid.

⏱️ Timestamps:

* 00:00 - Introduction: The myth of the Cloud vs. the reality of concrete.

* 01:50 - The financial operation: $830 Million for 13,800 Nvidia chips.

* 03:20 - Why Debt? Computing power as the new "oil rent".

* 06:05 - The geographic challenge and the US Cloud Act.

* 08:30 - Becoming "Full Stack": Controlling everything from the electrical outlet to the software.

* 09:20 - The timeline: The acquisition of Koyeb and the giant Campus with MGX.

* 10:45 - The latency nightmare: Why Koyeb is vital to synchronize the chips.

* 14:00 - The "Infrastructure Capital" era: The end of small AI startups.

* 16:20 - Conclusion: Will hardware centralization kill Open Source?

📈 Keywords:

Mistral AI, Nvidia, Digital Sovereignty, Cloud Act, Data Center, Artificial Intelligence, Koyeb, MGX, Tech Financing, The Last Men, Open Source, AI Infrastructure.

Hosted on Acast. See acast.com/privacy for more information.

What if the future of AI wasn't just about raw computational power, but about understanding and empathy?

In this episode, we delve into a groundbreaking shift in artificial intelligence, focusing on emotional intelligence as a key component. This pivot is spearheaded by Eric Zelickman, whose ambitious startup Humanzen is at the forefront of this movement. With a staggering billion-dollar raise at a $4 billion valuation, Zelickman's venture aims to redefine how AI interacts with humans, moving beyond the current models that often feel cold and disconnected.

Eric Zelickman is not just any researcher; he is a former PhD student at Stanford and a key figure in AI research. His experience includes being on the technical staff at XAI in 2024, and he was the lead author of a pivotal paper on language models. Zelickman has the technical expertise and vision that attract significant venture capital interest. His critique of current AI models centers on their lack of emotional intelligence and long-term contextual understanding, which he argues are crucial for effective human-AI collaboration.

The episode explores how Humanzen's empathetic AI model aims to address these shortcomings by building a profile that learns and grows with the user. This approach promises to foster better collaboration between humans and AI, which Zelickman believes is essential for tackling major global challenges like climate change and disease. The discussion raises critical questions about the future of AI, particularly how success in creating empathetic AI can be measured and what it means for human values and ambitions. As venture capitalists place massive bets on this new direction, the potential impact on technology and society is profound.

Hosted on Acast. See acast.com/privacy for more information.

Are we witnessing the dawn of a new era in technology, or are we on the brink of a financial bubble?

This episode of the Deep Dive delves into the high-stakes world of AI and finance, specifically focusing on OpenAI's ambitious IPO goals. With a potential $1 trillion valuation on the horizon, OpenAI is setting the stage for what could be the largest technology IPO in history. But with such lofty ambitions come significant challenges and questions about sustainability and market confidence.

this episode is not a single individual but rather a collection of insights from industry experts and financial analysts who have been closely monitoring OpenAI's journey. The episode draws on a variety of sources to paint a comprehensive picture of the current landscape, highlighting the tension between OpenAI's mission to develop safe and beneficial artificial general intelligence (AGI) and the immense financial pressures it faces.

Throughout the episode, we explore the core conflict between OpenAI's grand vision and the reality of its competition. Despite spending billions on research and development, OpenAI's flagship model, ChatGPT5, recently underperformed in a specialized task, losing 66% in an autonomous crypto trading competition. In stark contrast, a much cheaper and specialized rival, Deep Seq, achieved a positive return. This disparity raises questions about the feasibility of a $1 trillion valuation and whether smaller, more focused AI models could shift the financial landscape. As OpenAI navigates these challenges, the episode invites listeners to consider the broader implications for the future of AI investment.

Hosted on Acast. See acast.com/privacy for more information.

$1 billion to build AI models with emotional intelligence

Have you ever wondered what it would be like if AI could truly understand and respond to your emotions? In today's episode of The Deep Dive, we're exploring a groundbreaking AI funding story that shifts the focus from sheer computational power to emotional intelligence. We're delving into the mission of a new startup, Humans in Darrett, which aims to create AI that is not just smart, but emotionally intelligent. This episode challenges the conventional notion of AI as cold and machine-like, and instead, envisions a future where AI can empathize and collaborate effectively with humans.

Leading this innovative venture is Eric Zelickman, a former Stanford PhD student and a key figure in the AI community. Zelickman has a background in logic and reasoning, having contributed significantly to advancements in AI language models. His work introduced the concept of a "scratch pad" for internal reasoning, which helped improve AI's logical flow. Now, he's making a bold leap into the realm of emotions, raising a staggering $1 billion at a $4 billion valuation to bring his vision to life. Investors are not just betting on the idea of emotionally intelligent AI; they are backing Zelickman's expertise and track record in the field.

This episode unpacks Zelickman’s critique of current AI models, which he argues are too detached and fail to maintain emotional continuity in conversations. Humans in Darrett aims to address this by developing AI that remembers emotional contexts and can empathize with users. The implications of such technology are profound, extending beyond improved chatbots to potentially revolutionizing fields like cancer research, education, and diplomacy. By fostering better human collaboration through empathy, AI could become a powerful ally in tackling complex global challenges. The episode leaves us pondering whether the $4 billion valuation is justified and what the future holds if AI can genuinely understand human emotions and motivations.

Hosted on Acast. See acast.com/privacy for more information.

Have you ever wondered what it would take for AI to truly remember you?

In this episode of "The Deep Dive," we explore a groundbreaking development in artificial intelligence: solving the long-term memory problem. Imagine interacting with an AI that doesn't just forget your previous conversations but remembers and builds on them, much like a human would. This is the vision behind a new startup called Super Memory, which has recently secured $2.6 million in seed funding from industry giants like Google AI chief Jeff Dean and Cloudflare CTO Dane Necht. The involvement of such prominent figures indicates the significance of tackling memory persistence in AI, a challenge that has long been a stumbling block for large language models.

Meet Dravya Shah, the 19-year-old prodigy behind Super Memory.

Originally from Mumbai, Dravya Shah's journey into the tech world began with a simple bot that formatted tweets into screenshots, which he sold to a company called Hypefury. This entrepreneurial spirit led him to the United States for university, and eventually to a pivotal role at Cloudflare. It was there that Shah's idea for Super Memory took shape, with encouragement from advisors like Dane Necht. His innovative approach to AI memory caught the attention of tech leaders, prompting them to support his vision of creating a universal memory API that could revolutionize how AI interacts with data.

Revolutionizing AI with a universal memory API.

Super Memory's technology is designed to manage and utilize vast amounts of unstructured data—ranging from files and documents to chat logs and multimedia content—by extracting key insights and building a comprehensive knowledge graph. This approach not only enhances the AI's ability to recall information but also improves performance by structuring data into a relational map, allowing for more efficient and context-aware retrieval. As a result, applications can achieve deeper personalization and understanding. With backing from investors and early customers like a fintech company and an AI video editor, Super Memory is poised to lead a transformation in AI, turning forgetful tools into personalized partners that genuinely remember and understand user interactions over time. This shift could redefine the future of AI, making it an integral part of our daily lives.

Hosted on Acast. See acast.com/privacy for more information.

What does it take to power the future of AI?

How much compute power is truly necessary to unlock AI's full potential? This episode delves into a groundbreaking partnership that could redefine the AI landscape. We're exploring the colossal deal between AMD and OpenAI, involving billions of dollars and, intriguingly, gigawatts of power. As we peel back the layers of this agreement, we uncover the staggering scale of AMD's commitment to supply 6 gigawatts of power—enough to energize millions of homes. This episode challenges listeners to ponder the immense infrastructure required to fuel the ambitions of AI giants.

Meet the key players behind the scenes

Our guest today is an industry insider with deep insights into the tech world, particularly in the realm of artificial intelligence and computing power. They bring a wealth of knowledge about AMD's latest technological advancements and OpenAI's strategic maneuvers in securing resources from multiple tech giants. With a background in both hardware and software development, our guest provides a nuanced perspective on how these partnerships are shaping the future of AI and the potential risks and rewards involved.

A high-stakes gamble in the AI arms race

The episode provides a high-level overview of the AMD-OpenAI partnership, focusing on the technical and financial intricacies. AMD is betting big with its next-gen Instinct Mi450 chips, scheduled for deployment in 2026, and OpenAI is poised to become a major shareholder, contingent on AMD meeting ambitious performance targets. This deal is part of OpenAI's broader strategy, which includes commitments from Nvidia and Broadcom, and a massive Stargate initiative for data center expansion. The discussion highlights the intense competition and collaboration required to meet the astronomical compute demands of AI's future, as OpenAI diversifies its suppliers to avoid dependency.

Hosted on Acast. See acast.com/privacy for more information.

What would you create if you had the power to build your own AI agent with just a few clicks?

In this episode of the Deep Dive podcast, we explore OpenAI's groundbreaking announcements from their Dev Day 2025. The central theme of the event was "vibe coding," a concept that promises to revolutionize how we interact with technology by transforming ChatGPT from a simple chatbot into a comprehensive front-end development tool. With over 800 million weekly users, this evolution aims to make cutting-edge technology more accessible and visually immediate. Our mission today is to delve into the three main components driving this innovative idea: the new apps SDK, the Agent Kit API, and a significant change in content policy regarding mature experiences.

today the conversation revolves around the latest initiatives from OpenAI, a leading AI research organization known for its development of advanced artificial intelligence models. The discussion highlights OpenAI's strategic moves to partner with major companies like booking.com, Canva, Coursera, and Spotify, among others, showcasing the broad impact and potential of their new tools. These partnerships are set to enhance user experience by embedding third-party apps directly into ChatGPT, allowing seamless interaction without leaving the chat interface.

The episode provides a high-level overview of OpenAI's new offerings, focusing on their potential to democratize AI creation through vibe coding. The apps SDK enables deep integration of third-party apps into ChatGPT, while the Agent Kit API allows users to build customized AI agents with ease, using a visual interface akin to Canva. However, the introduction of mature content raises important safety and governance questions. OpenAI plans to implement robust age verification tools before launching any 18+ features, highlighting the delicate balance between innovation and responsibility. As listeners ponder the possibilities of these advancements, they are left with a thought-provoking question: what AI agent would you create to simplify your life?

Hosted on Acast. See acast.com/privacy for more information.

Are we inadvertently fueling climate change with our digital habits?

In this episode of "The Deep Dive," we explore a pressing paradox: the significant and growing energy and carbon footprint of AI and the data centers it requires. While digital technology is often heralded as a tool for decarbonizing the economy, its rapid expansion is proving unsustainable. In 2020, the digital sector alone accounted for nearly 4% of global greenhouse gas emissions, a figure that is projected to grow alarmingly if current trends continue. The conversation centers on the potential surge in data center energy consumption, which could see a dramatic increase from 530 terawatt hours in 2023 to almost 1500 TWh by 2030, with corresponding emissions potentially doubling France's national emissions.

the episode draws on insights from various sources to shed light on the topic. The discussion revolves around the implications of AI, particularly generative AI, on energy consumption. The host emphasizes the staggering adoption rates of technologies like ChatGPT, which reached 100 million users in just two months, illustrating the rapid pace at which these technologies are being integrated into daily life. This swift adoption raises questions about the sustainability of AI's energy demands, especially during the inference phase, which is projected to constitute a significant portion of AI's electricity consumption by 2030.

The clash between AI growth and climate goals

The episode delves into the conflict between the booming growth of generative AI and global climate objectives. As AI technologies expand, the demand for energy, often met by fossil fuels, is increasing faster than clean energy solutions can be deployed. This has led to the development of numerous gas-fired power plants specifically to power data centers. The local impact is particularly acute in places like Ireland, where data centers consume over 20% of the country's electricity, prompting a moratorium on new connections. Similarly, in France, data center energy demands threaten to overshadow the power needed for the country's energy transition goals. The episode concludes with a sobering reminder of the long-term consequences of delaying stricter efficiency standards and carbon budgets for the sector, potentially locking in significant emissions for decades.

Hosted on Acast. See acast.com/privacy for more information.