Latest Interviews
Showing 61–75 of 129 interview transcripts.
Clear all filters- RAISE Summit18 min
Cheap Tokens, Expensive Mistakes: The Real Economics of AI at Scale | RAISE Summit 2026
Liran Zvibel, Dylan Patel, Karen Kwok
Major AI operators are pivoting from pure pricing power to structural cost efficiency, leveraging advanced KV cache offloading and optimized storage architectures to slash inference expenses. Recent benchmarks reveal that NAND-based memory solutions and strategic hardware combinations, such as high-capacity AMD GPUs, can outperform standard setups by up to 20 times in multi-turn agentic workflows while delivering 7x higher throughput. This shift allows emerging providers and frontier labs to significantly expand gross margins, with companies like Anthropic and OpenAI now projecting substantial operating profits driven by these infrastructure innovations rather than model weights alone.
- RAISE Summit19 min
The Mindset Shift Needed to Thrive Alongside AI | Marco Argenti, Goldman Sachs | RAISE Summit 2026
Goldman Sachs has achieved a 20% net engineering productivity gain by shifting from spec-driven to prototype-driven development, where business professionals utilize AI tools to build functional models that engineers then scale. This transformation is underpinned by strict governance frameworks that treat AI output as junior input requiring rigorous code review, alongside a leadership culture where executives actively model adoption while retaining final strategic decision-making authority. Consequently, junior staff now focus on client engagement and apprenticeship rather than mechanical analysis, as internal copilots handle the bulk of data processing and initial research tasks.
- RAISE Summit18 min
Keeping AI Honest: The AI Wave Seen From the Inside | Olivier Pomel, Datadog | RAISE Summit 2026
Datadog reported fiscal revenue projections reaching $4.3 billion, driven by a 30% year-over-year growth rate fueled by rising AI adoption across its diverse customer base. CEO Olivier Limousin highlighted the company's strategic pivot from passive observability to active autonomy through the acquisition of Adaptive ML, aiming to build proprietary models that are faster and cheaper than frontier alternatives. While current systems resolve 40% of issues independently, leadership maintains a culture of rigorous caution to ensure near-perfect accuracy before pursuing full self-healing capabilities.
- RAISE Summit18 min
Betting on the AI Application Layer | Grant Lee, Gamma | RAISE Summit 2026
Gamma, an AI-driven visual communication platform with over 100 million users and a $2.1 billion valuation, is expanding its international footprint by establishing a London office to support its 80% non-US user base. Led by Grant Lee, the company is pivoting from simple slide generation to enterprise-grade brand alignment while adopting a lean organizational structure that prioritizes cohesive user experiences over proprietary model ownership. Lee further advocates for a multi-model strategy to navigate regulatory shifts and positions the current market as a transition from bolting AI onto workflows to fundamentally reinventing product architecture.
- RAISE Summit28 min
Fireside Chat with Yann LeCun, Executive Chairman of AMI Labs | RAISE Summit 2026
Yann LeCun, Tom Mackenzie, Chet Haase, Francois Beaufort, Romain Guyett
Yann LeCun founded AMI Labs to develop JEPA-based World Models that overcome the physical reasoning limitations of current Large Language Models by predicting abstract states rather than discrete tokens. This strategic departure from Meta, driven by incompatible visions for Artificial General Intelligence and business focus, positions the Paris-based entity to lead global industrial applications like Level 5 autonomy and domestic robotics. To ensure geopolitical neutrality and preserve data sovereignty, LeCun is also spearheading Project Tapestry, a distributed initiative aggregating parameters from diverse international contributors without requiring raw data sharing.
- RAISE Summit22 min
OpenAI x Cerebras: Sachin Katti & Andrew Feldman in Conversation | RAISE Summit 2026
Sachin Katti, Andrew Feldman, Henri Delahaye
In December 2024, OpenAI and Cerebras finalized a historic $20 billion partnership to deploy GPT-5.6 on Cerebras infrastructure, enabling a record-breaking inference speed of 750 tokens per second while expanding data center capacity across Europe to address sovereign AI demands. This collaboration has accelerated OpenAI's internal operations, establishing Codex as the default interface for all employees and driving a monthly release cadence of frontier models through AI-assisted research workflows. Looking forward, the alliance prioritizes enterprise adoption over token volume, anticipating that the rapid integration of AI agents will fundamentally rewire organizational structures and redefine productivity metrics within the next year.
- RAISE Summit19 min
Winning Travel's AI Race | Ariel Cohen, Navan and Molly O'Shea, Sourcery | RAISE Summit 2026
Navan leverages its "Navan Cognition" orchestration platform to deploy hybrid human-AI workflows that handled 60% of complex travel issues with human-comparable satisfaction while driving 50% year-over-year Gross Booking Value growth. This strategy enabled the company to achieve profitability and positive cash flow by combining AI efficiency with a "supervisory model" where human agents oversee automated agents for high-stakes executive travel. Ariel Cohen positions Navan to capture market share from legacy competitors by replacing the traditional SaaS model with performance-based revenue and a blended business-leisure booking experience.
- RAISE Summit10 min
Matt Hicks, Red Hat ｜ RAISE Summit 2025
At the RAIDS Conference 2025 in Paris, Red Hat unveiled its strategic shift toward sovereign AI infrastructure, introducing products like the Red Hat Inference Server and OpenShift Lightspeed to optimize open-source model deployment across enterprise environments. By partnering with major hardware and cloud providers including NVIDIA and AMD, the company aims to resolve critical GPU utilization bottlenecks and control token costs, thereby enabling scalable agentic workloads that transition from experimentation to production. This approach positions software innovation, specifically through tools like VLLM and LLMD, as the primary driver for maximizing efficiency in a landscape where hardware performance is plateauing.
- RAISE Summit8 min
Vipul Ved Prakash, Together AI ｜ RAISE Summit 2025
Vipul Ved Prakash, John Furrier, Aviv Paul
Co-founded by Paul Ved Prakash, Together AI operates as an AI acceleration cloud that recently deployed a high-density Grace Blackwell 200 cluster in Memphis to deliver 1.4 petaflops of computing power for large-scale model training. The platform hosts over 200 open-source models and drives significant growth among "AI-native" customers by integrating specialized infrastructure for thermal management and dynamic workload orchestration. As the market shifts toward "Agentech," the company addresses generative AI limitations through test-time compute mechanisms and parameter-efficient inference techniques that enable complex, multi-step reasoning workflows.
- RAISE Summit8 min
Vipul Ved Prakash, Together AI ｜ RAISE Summit 2025 1
Vipul Ved Prakash, John Furrier, Aviv Paul
Together AI, an "AI acceleration cloud" founded three years ago, is deploying specialized supercomputer clusters like the new Memphis Grace Blackwell 200 to facilitate large-scale generative AI training and inference. Serving 80% "AI-native" clients in sectors from robotics to healthcare, the platform distinguishes itself through hardware-software co-optimization that enables real-time reasoning and supports a library of 200 open-source models. As the industry shifts from basic inference to multi-step agent technologies, Together AI aims to bridge the enterprise gap for integrated stacks by dynamically balancing high-density workloads across its unified cloud infrastructure.
- RAISE Summit16 min
Prashanth Chandrasekar, Stack Overflow ｜ Raise Summit 2025
Prashanth Chandrasekar, John Furrier, Prashant Thir
Stack Overflow CEO Prashant Thir highlights a critical market shift where high AI adoption clashes with declining trust, driving a demand for human-in-the-loop validation to counteract generic content and hallucinations. The company leverages its official data licensing agreements with major AI labs to boost model accuracy by 30–40% while simultaneously expanding enterprise offerings like Uber's Genie into dynamic knowledge intelligence layers. This strategy combines a growing revenue stream from AI reinforcement learning and data licensing with a community vision to transform the platform into a "trust network" that blends expert curation with AI speed.
- RAISE Summit8 min
Mike Mattacola, Coreweave ｜ Raise Summit 2025
Held in Paris, the RAISE Summit 2025 convened industry leaders to address the critical intersections of AI infrastructure expansion, sovereignty, and energy constraints. CoreWeave leveraged its strategic partnership with Dell and NVIDIA to outpace competitors by deploying H100, H200, and GB200 hardware generations first, while managing massive scaling complexities through joint troubleshooting and agile operational models. The event highlighted a structural industry shift from general-purpose IT to specialized AI engineering, characterized by high-density rack consolidation and a dramatic decline in x86 market share as physical data center requirements drive immediate procurement and deployment strategies.
- RAISE Summit10 min
Alison Wagonfeld, Google Cloud ｜ RAISE Summit 2025
Alison Wagonfeld, John Furrier
At the 2025 RAISE Summit in Paris, Google Cloud detailed its transition to a full-stack AI-first strategy that integrates DeepMind research directly into products like BigQuery while leveraging partners such as Databricks and hardware collaborations with Nvidia. The company championed open interoperability standards, specifically the Model Context Protocol and Agent-to-Agent protocol, to enable secure cross-platform agent interactions and mitigate vendor lock-in for over 500 enterprise customers. This approach addresses a market shift toward replatformization and internal AI competency centers, allowing organizations to extract measurable ROI through application-layer business logic rather than focusing solely on hardware advancements.
- RAISE Summit18 min
Benoit Dageville, Snowflake ｜ RAISE Summit 2025
Benoit Dageville, John Furrier
Snowflake has advanced its platform by establishing unstructured data as a first-class citizen through new AI SQL capabilities and the official launch of Semantic Views, which embed business logic to ensure AI query accuracy exceeds 95%. This evolution supports a three-phase integration of generative AI, ranging from data ingestion and natural language interaction to a future agentic layer that democratizes access for non-technical users while maintaining strict governance via a dual proprietary and open-source cataloging strategy. By distinguishing itself through a dedicated semantic layer and objective performance benchmarks, the company positions this shift as a foundational revolution that simplifies complex data relationships for enterprises without compromising security or cost efficiency.
- RAISE Summit14 min
Roman Chernin, Nebius ｜ Raise Summit 2025
Roman Chernin, John Furrier, Robin Sheeran
Nebius, a publicly traded full-stack AI cloud provider founded one year ago, has expanded its infrastructure to seven global locations and targets operating tens of thousands of GPUs by year-end. The company differentiates itself through proprietary liquid-cooled engineering and a pay-per-consumption model that accommodates diverse workloads ranging from real-time inference to massive training clusters. Nebius is pursuing a tenfold revenue growth to one billion Russian Rubles while pivoting its customer base toward established enterprises in life sciences and FinTech to support the transition to "AI factories."