newsfilter.io

Tony Kim

Showing 18 of 8 transcripts.

  1. Sourcery with Molly O'Shea24 min

    Lumentum CEO: How Lasers Are Transforming AI Data Centers

    Michael Hurlston, Tony Kim, Molly O'Shea

    Under CEO Michael Hurlston, Lumentum has transformed from a diversified manufacturer into a pure-play data center optical leader, tripling revenue and driving its stock price up to 12-fold by capitalizing on the industry's rapid shift from copper to fiber connectivity. This transition, necessitated by the physical limitations of copper in high-speed environments, is straining the specialized Indium Phosphide supply chain due to long lead times for dedicated fabrication facilities. Facing tens of millions in new hyperscaler orders, the company is simultaneously expanding into space-based internet infrastructure and navigating geopolitical dynamics that favor U.S. domestic suppliers over Chinese competitors.

  2. RAISE Summit18 min

    The Invisible Backbone of AI: Why Light Powers Intelligence | Lumentum | RAISE Summit 2026

    Michael Hurlston, Tony Kim

    Lumentum outlines a four-stage data center evolution from kilometer-scale interconnects to millimeter-level optical PCB integration, a shift driven by copper's inability to support 1.6 terabit speeds. This transition promises a 100 to 1,000-fold expansion in optical lanes, with scale-up rack integration expected by mid-2027 and commercial scale-in adoption projected by 2029. To meet the demand for hundreds of millions of lasers, the company has expanded Indium Phosphide wafer fabrication while strategic investments from NVIDIA secure supply for these critical semiconductor components.

  3. Sourcery with Molly O'Shea1h 9m

    BlackRock's Tony Kim on AI's Next Winners? Chips, Memory, Robotics & Quantum

    Tony Kim, Molly O'Shea

    Kim outlines a fundamental shift in the global technology sector where a $30 trillion hardware-centric economy is driven by a 10,000x increase in compute density and severe memory shortages known as the "Rampocalypse." Leading institutions are responding through silicon co-design and gigawatt-scale energy infrastructure while BlackRock allocates capital toward immediate compute needs and long-term bets on quantum and orbital data centers. The narrative concludes with strategic observations on China's robotics acceleration and the restructuring of traditional enterprises into AI-native entities through proprietary token flow architectures.

  4. RAISE Summit17 min

    Fireside Chat with Sid Sheth, Founder & CEO of d-Matrix | RAISE Summit 2026

    Sid Sheth, Tony Kim

    D-Matrix, founded in 2019, has transitioned its SRAM-based Corsair accelerator into mass production while partnering with NVIDIA to deploy heterogeneous compute solutions for the inference-focused agentic era. Moving beyond current SRAM limitations, the company is developing a 3D stacked In-Memory Compute architecture scheduled for a 2027 tape-out, which aims to simultaneously optimize latency and throughput for high-interactivity workloads. This strategic trajectory envisions a market where specialized inference accelerators coexist with HBM-based GPUs, handling distinct layers of the AI pipeline from heavy pre-fill computations to memory-intensive decoding.

  5. RAISE Summit19 min

    Scaling AI Infrastructure for the Agentic Era | Charlie Kawwas, Broadcom | RAISE Summit 2026

    Charlie Kawwas, Tony Kim

    Frontier artificial intelligence laboratories are increasingly partnering with Broadcom to deploy purpose-built XPU architectures that replace general-purpose hardware, enabling custom designs that eliminate efficiency losses and proprietary ecosystem taxes. This strategic shift supports a massive scaling trajectory where Broadcom projects over $100 billion in revenue by 2027 while delivering 16 times the current compute capacity through modular chiplet technology. As major clients contract for gigawatt-scale power expansions, the collaboration prioritizes rapid time-to-market for specialized "agentic AI" workloads while addressing supply chain constraints through new advanced packaging facilities.

  6. Milken Institute36 min

    The AI Investment Cycle: Platforms, Infrastructure, and Markets | Global Conference 2026

    Yun-Hee Kim, Dennis Gada, Tony Kim, Lily Liu, Scott Rubner

    This event analyzes the current trillion-dollar trajectory of global AI capital expenditure, which is driving a structural market shift away from traditional SaaS models toward a three-layer stack dominated by compute and intelligence providers. Key figures from Citadel, BlackRock, and the Solana Foundation discuss the emergence of machine-to-machine financial rails, the geopolitical bifurcation of sovereign AI, and the impending influx of trillion-dollar AI valuations into public markets. The discourse concludes by projecting long-term economic rewiring that will necessitate new frameworks for human capital, cognitive preservation, and decentralized ownership models.

  7. Goldman Sachs11 min

    Copper: AI Hype or Supply Squeeze?

    Adam Crook, Tony Kim

    A commodities strategist analyzes the divergence between gold's resilient safe-haven performance during the Iran conflict and the skepticism regarding its near-term upside caused by shifting macro rates. The trader assesses silver's limited upside potential due to its tight correlation with gold, while maintaining a neutral stance on copper despite AI-driven demand narratives that conflict with current global surplus levels. Finally, the outlook for aluminum projects a temporary physical deficit and price appreciation through summer before a surplus is expected to emerge following the resumption of Middle Eastern production later in the decade.

  8. RAISE Summit30 min

    RAISE Summit 2025: The AI Evolution Open Source, Fast Inference, and the Agentic Revolution

    Tony Kim, Thomas Wolf, Rodrigo Liang, Arjun Kharpal

    The AI sector is rapidly pivoting from foundational model development to agentic inference and production, driven by a strategic race for full-stack superiority where proprietary models currently retain a six-to-twelve-month performance lead. Enterprises are increasingly adopting open-source alternatives to ensure data privacy and cost efficiency, prompting a hardware shift focused on energy-efficient inference rather than raw training power. While consumer assistants and automated SaaS applications are expected to mature within two years, the industry faces critical bottlenecks regarding gigawatt-scale energy demands and the potential for data security breaches.