Latest Interviews
Showing 31–45 of 231 transcripts.
Clear all filters- RAISE Summit17 min
AI for Chip Design & Chip Design for AI | Ricursive Intelligence | RAISE Summit 2026
Anna Goldie, Azalia Mirhoseini, Azalea Amir-Hosseini
Founded by former Google Brain engineers Anna Goldie and Azalea Amir-Hosseini, Recursive Intelligence deploys its AlphaChip reinforcement learning system to replace months of manual chip design with superhuman layout generation that runs 1,000 times faster than commercial tools. The company is currently in Phase One, accelerating physical design for major manufacturers like NVIDIA and AMD, while its roadmap targets a "design-less" future where companies specify workloads to receive co-optimized hardware instantly. Ultimately, Recursive aims to close a recursive self-improving loop by eventually building its own chips and models, promising up to 20x performance gains through microarchitectural customization.
- RAISE Summit19 min
The Agentic Web is Next: Building the Infrastructure for AI Agents on the Web | RAISE Summit 2026
Founded by Co-CEO Abhishek, Yotori deploys its autonomous Navigator N1.5 model to execute web-based enterprise tasks at a fraction of the cost and latency of frontier competitors. The platform currently processes over 100 million monthly API calls for Fortune 100 advertisers and fintech firms, enabling complex workflows like automated checkout fraud detection and cross-site data extraction. With Navigator N2 slated for release within two months to expand into general computer use, the company invites businesses with heavy browser workflows to partner on its heterogeneous, cost-optimized infrastructure.
- RAISE Summit21 min
It Takes AI to Break AI: Securing the Agentic Attack Surface | Alice | RAISE Summit 2026
Alice Lead of Product and Engineering Avi Golan explains how enterprises face escalating security and safety risks as AI evolves from static chatbots to autonomous agents with tool-calling capabilities. He details a specialized Agentic AI Red Teaming methodology that moves beyond traditional static scanning to continuously simulate adversarial attacks, ensuring robust threat modeling before production deployment. This approach addresses the unique vulnerabilities of non-deterministic models, such as reasoning manipulation and indirect prompt injection, which have previously led to significant operational failures and legal liabilities across major brands.
- RAISE Summit17 min
Perfect Search, Built for AI Agents | Will Bryk, Exa | RAISE Summit 2026
Exa is rapidly scaling its semantic search infrastructure to power the next generation of AI agents, which are projected to outnumber human web searches within the year. By leveraging deep learning embeddings and a new marketplace for proprietary data, the platform enables precise, low-latency queries that reduce LLM costs while aiming for 99.9999% decision accuracy. This technology drives the company's mission to transition the digital landscape from keyword-based recommendations to "perfect search," targeting the 2028 U.S. presidential election as a pivotal milestone for enhancing societal decision-making.
- RAISE Summit19 min
Hot Takes & Hard Truths: Building Growth That Compounds | Clay x CapitalG | RAISE Summit 2026
Clay repositions itself as an AI go-to-market platform that transitions sales operations into self-learning revenue engines by integrating third-party data signals to execute scalable growth campaigns. Co-founder Varun leads a strategic shift toward "marketing alpha," utilizing unique tactics like custom film production and rapid, data-driven prospecting to secure high returns on investment for clients such as Canva and Anthropic. The company further distinguishes its value through a new pricing model based on workflow outcomes rather than data tokens, while emphasizing environmental fit in talent acquisition to maximize organizational performance.
- RAISE Summit40 min
Compute as Capital: Redefining AI's New Currency | WEKA, Akamai, TensorWave & More | RAISE 2026
Val Bercovici, Carmen Li, Parimal Pandya, Sam Tabar, Piotr Tomasik, Robin Wauters
With global AI infrastructure spending projected to reach $7 trillion by 2030, industry leaders identify power access and specialized talent as the primary constraints limiting supply. Experts from White Fiber, Weka, and Silicon Data argue that compute has evolved into a high-value asset class driven by unique market dynamics, including appreciating H-100 GPU values and emerging financial instruments for compute volatility. While current value capture remains anchored in the physical supply chain of chips and energy, the sector anticipates a structural shift toward application-layer returns by 2031 alongside advancements in virtual supercluster technologies and regional market expansion.
- RAISE Summit42 min
The Data Problem: What AI Actually Runs On | RAISE Summit 2026
Matt Fitzpatrick, Philippe Mizrahi, Noam Salinger, Molly Presley, Travers Nisbet, Petrina Anne Steele, Petrina Steele, Emily Fortuna
Enterprise AI adoption remains critically hampered by organizational inertia and legacy infrastructure, with only 5% to 8% of projects reaching production while global compute allocation shifts decisively toward 80% inference needs. Leaders must address fragmented data silos and misaligned workflows by implementing global namespaces and human-in-the-loop validation protocols to ensure agents meet strict liability standards. As constraints evolve from pure data quality to energy supply and physical infrastructure, successful transformation will depend on re-engineering business processes rather than merely overlaying new models onto broken systems.
- RAISE Summit18 min
From Metal to Model: Why Operations Win the AI Infrastructure Race | Mirantis | RAISE Summit 2026
Mirantis is launching Cordoned AI, a metal-to-model platform designed to address the shift toward inference-dominated workloads where hardware commoditization makes operational full-stack ownership the primary competitive advantage. The event highlights how rising failure rates in massive GPU clusters necessitate advanced observability and rigorous process procedures to reduce downtime and control costs. By enabling dynamic management of both CPU and GPU resources through an open architecture, the solution aims to support the diverse latency and throughput requirements of modern AI infrastructure.
- RAISE Summit19 min
The Token-per-Watt Era: What's Next for AI Hyperscalers | Nebius x Accel | RAISE Summit 2026
Roman Chernin, Philippe Botteri
Nebius, an AI-specialized cloud provider led by Chief Business Officer Roman Tchernin, marks its second anniversary as an independent entity with shares rising tenfold following its IPO. The company targets the projected $11 trillion global investment in AI infrastructure by offering a vertically integrated stack ranging from bare metal for hyperscalers to managed inference for AI-native startups. Nebius anticipates that true enterprise adoption beyond current exceptions like fintech firms remains unrealized and aims to future-proof its platform by shifting pricing models toward billing for solved tasks rather than raw compute hours.
- RAISE Summit20 min
Fireside Chat with Matan Grinberg, Founder & CEO of Factory | RAISE Summit 2026
Founded in 2023 by ex-physics PhD Matan Grinberg, Factory operates as a model-agnostic router that enables enterprises to deploy fully on-premises autonomous software agents directly within existing developer workflows. The company rapidly scaled from ten to ten thousand users within months by offering a cost-effective alternative to vendor lock-in, notably accelerating legacy COBOL migration and allowing sovereign data deployments for regulated industries. Backed by investors including Sequoia and Avant Capital, Factory currently automates both software engineering tasks and emerging non-software productivity use cases while predicting a future dominance of open models over frontier closed ones.
- RAISE Summit19 min
From Megawatts to Gigawatts: Building the AI Factories | Lenovo x NVIDIA | RAISE Summit 2026
John Donovan, David Hogan, Roxanne Varza
John of Lenovo and David of NVIDIA are advancing a deep engineering partnership to construct gigawatt-scale "AI factories" that integrate hardware, power infrastructure, and software optimization to address a projected global capacity doubling. Their four-stage deployment framework, which spans from conceptual design to post-launch model orchestration, specifically targets the European market's urgent need to close a 12-month GPU access gap while navigating complex regulatory and energy constraints. By combining Lenovo's global manufacturing scale with NVIDIA's regional expertise, the initiative aims to transform Europe into a sovereign AI hub capable of leveraging unique domain data for next-generation robotics and industrial applications.
- RAISE Summit42 min
2035 Decoded: Navigating the Decade Ahead | Dust, Glean & More | RAISE Summit 2026
Gabriel Hubert, Eve Bodnia, Arvind Jain, Nico Laqua, Raphaelle D'Ornano, Chris Blundell
Projected as a 10-to-15-year strategic journey, the panel outlines how enterprises must prioritize context graphs and data ownership over specific models to navigate the divergence between the 125 projected S&P 500 winners and the majority of firms unable to transition to agentic workflows. While Large Language Models dominate creative tasks, the future architecture will integrate Energy-Based Models for critical reasoning, demanding talent with advanced systems thinking to manage non-deterministic risks in sectors like finance and robotics. Long-term competitive advantage will ultimately accrue to organizations that retain control over system specifications and leverage human oversight to compound value through defensible data moats.
- RAISE Summit19 min
AI Transformation Playbook: Practical Lessons from our AI-First Journey | Make | RAISE Summit 2026
Visual workflow automation scale-up Make transitioned to an AI-first entity over two and a half years by implementing a three-phase strategy that combined a dedicated "Samurai" transformation team with a mandatory quarterly "AI Bar" for all 300 employees. This structured approach increased internal AI agent usage from 16% to 96% and generated 23 successful projects in a single year, achieving a 70% success rate compared to the industry average of 5–10%. The initiative shifted organizational focus from debating adoption to executing context management, resulting in 73% of projects directly impacting core business KPIs across every department.
- RAISE Summit20 min
Infrastructure-as-Prompt: Turning Intent Into Infrastructure | emma | RAISE Summit 2026
emma, Dmitry Panenkov, Dimitri
EMMA, a Luxembourg-headquartered sovereign cloud provider backed by investors including BlackRock, launched its "Infrastructure as Prompt" platform in September to automate cloud management through natural language interfaces. The system functions as a unified control layer that minimizes data egress costs and reduces deployment time to fifteen minutes by selecting optimal compute instances across hyperscalers and on-premises environments. As European sovereign cloud spending remains a small fraction of the global market, EMMA's approach aims to redefine digital sovereignty by shifting control from geographic location to governance and access management while offering a proprietary fleet of accelerators ranging from enterprise to consumer-grade models.
- RAISE Summit20 min
Self-Driving Production: AI Wrote Your Code. AI Should Fix It, Too. | Traversal | RAISE Summit 2026
Founded by an MIT PhD and Columbia professor, Traversal deploys an AI-driven site reliability engineer that distinguishes itself from traditional correlation-based tools by utilizing a proprietary causal search engine to identify root causes across petabytes of daily telemetry. This platform delivers L5 autonomy in production environments, evidenced by enterprise deployments at American Express that reduced incident resolution times by 58% and Pepsi initiatives that preemptively halted customer-facing outages through historical alert analysis. By integrating a unique production world model, the system autonomously executes complex remediation steps such as traffic failover and code generation, effectively transforming maintenance workflows from reactive war rooms to proactive self-healing operations.