AI COMPUTE•NVIDIA•GPU CLOUD•DATA CENTERS•SEMICONDUCTORS•ENERGY•FUNDING•M&A•AI COMPUTE•NVIDIA•GPU CLOUD•DATA CENTERS•SEMICONDUCTORS•ENERGY•FUNDING•M&A•

GPU Data Hub Daily Brief

The AI Infrastructure Brief

The most important developments across AI compute, GPUs, data centers, chips and energy.

29 September 2026 edition

AMD Agrees to Acquire World Labs for $8.2 Billion as Infrastructure Capital Surges

AMD has struck a definitive $8.2 billion deal to acquire Fei-Fei Li's World Labs to bolster its spatial AI and model engineering efforts, while Samsung committed $1 billion to KKR- and Nvidia-backed infrastructure provider Helix. At the same time, physical power limits and regulatory shifts are complicating data center expansions and accelerating internal silicon developments.

TOP STORY

AMD to Acquire Fei-Fei Li's World Labs in $8.2 Billion Transaction

AMD has entered into an agreement to acquire spatial intelligence startup World Labs in an all-stock deal valued at approximately $8.2 billion. Founder Dr. Fei-Fei Li will join the chipmaker as chief scientist and executive vice president to enhance AMD's internal AI model engineering capabilities.

DEALS & FUNDING

Samsung Group Injects $1 Billion into Nvidia-Backed Helix

Samsung Electronics and five affiliate entities are deploying $1 billion into Helix, an AI infrastructure firm established by private equity firm KKR. The capital expansion follows earlier support from Nvidia as Samsung aims to scale its global hardware presence.

Inference Provider Modal Labs Nears $750 Million Round at $15.75 Billion Valuation

AI inference startup Modal Labs is reportedly finalizing a $750 million financing round valuing the company at $15.75 billion. The prospective valuation represents a threefold increase over a four-month span.

SEMICONDUCTORS

OpenAI Clarifies Scope for Custom Jalapeño Inference ASIC

OpenAI confirmed that its custom Jalapeño AI inference ASIC is primarily intended to handle the company's internal workloads. However, leadership stated that broader external distribution in the future has not been entirely ruled out.

Synopsys Unveils Autopilot and AgentEngineer for Autonomous Chip Design

Synopsys introduced its Autopilot platform alongside the AgentEngineer suite, deploying autonomous AI agents to automate semiconductor engineering workflows from silicon to system architecture.

DATA CENTERS

Power Constraints Supplant Silicon Supply as Primary Data Center Bottleneck

Physical electrical capacity and distribution limitations have replaced chip availability as the top scaling constraint for AI hyperscale facilities, shifting performance optimization toward data-movement energy efficiency.

Goodman Withdraws Application for 90MW Sydney Campus

Property developer Goodman has withdrawn its development application for a planned 90-megawatt data center campus in Sydney, citing changes in the local regulatory and policy environment.

CLOUD

Meta Recruits MongoDB CEO CJ Desai to Lead New Enterprise Unit

Meta has hired MongoDB chief executive CJ Desai to head its newly formed Meta Enterprise Platform division, expanding access to its Muse model family and business agent software ecosystem.

What to watch next

Watch for AMD's integration of World Labs' spatial computing team under Fei-Fei Li and whether it challenges Nvidia's software stack moat. Monitor how severe data center power bottlenecks impact near-term hyperscale campus approvals, particularly in constrained global metro hubs. Finally, track whether frontier labs follow OpenAI's lead in developing dedicated in-house inference ASICs to rein in compute expenses.

GPU Data Hub Daily

Stay Ahead of the AI Infrastructure Economy

The most important GPU, AI, data-center, semiconductor and cloud developments delivered directly to your inbox.

By subscribing you consent to receive the daily briefing. Unsubscribe at any time. See our privacy policy.