Skip to content

Compute & Silicon

Chips, data centers, and the physical substrate the AI build-out runs on.

7 articles

  • A $20B Signal About Where AI Infrastructure Is Actually Going
    Compute & Silicon5 min read

    A $20B Signal About Where AI Infrastructure Is Actually Going

    NVIDIA’s $20 billion deal with Groq marks a pivotal shift in AI infrastructure, moving the focus from model training to large-scale inference efficiency . Rather than a traditional acquisition, NVIDIA executed a "strategic intervention" by licensing Groq’s core technology and absorbing its top engin

  • China’s EUV Breakthrough
    Compute & Silicon6 min read

    China’s EUV Breakthrough

    For most of the last decade, discussions about artificial intelligence have focused on models, talent, and data. Yet in late 2025, a different story began to reshape that narrative. China has reportedly built a prototype extreme ultraviolet lithography machine, the most critical industrial tool for

  • Software & Silicon
    Compute & Silicon10 min read

    Software & Silicon

    NVIDIA’s real advantage isn’t the GPU, it’s CUDA: a vertically integrated software stack that turns theoretical FLOPs into real, high-utilization performance and continuously improves throughput, latency, and scaling via software alone. In AI inference, CUDA is what drives tokens per watt-dollar dow

  • How Google Is Competing in the AI Compute Race
    Compute & Silicon3 min read

    How Google Is Competing in the AI Compute Race

    Google is scaling its TPU strategy, offering a cost-efficient alternative to Nvidia that is gaining traction across major AI developers. Gemini-powered search continues to grow, supported by Google’s TPU-backed infrastructure. The broader market is moving toward multiple winners as Google, Anthropic

  • The Compute Brokerage Era
    Compute & Silicon3 min read

    The Compute Brokerage Era

    Compute is becoming the next major liquidity market as GPU demand outpaces supply constrained by energy, siting, and fabrication. Bursty, latency-sensitive inference and agentic workloads break the old rent-a-server model. A new layer of compute brokers is emerging to allocate, schedule, and price c

  • AI Roundtripping: NVIDIA, OpenAI, Oracle and the Circular Financing Debate
    Compute & Silicon6 min read

    AI Roundtripping: NVIDIA, OpenAI, Oracle and the Circular Financing Debate

    A series of large, interlinked deals between NVIDIA, OpenAI, and Oracle has raised questions about circular financing in AI. Some view it as inflated growth built on mutual dependence, while others see it as a practical way to fund and scale the infrastructure behind today’s AI expansion.

VE Weekly

The week in tech, AI and early-stage capital.

One send a week. Whole pieces, not a digest of links out.

One email a week. Unsubscribe in one click. Your address is never shared.