Category / T-CAT / HARDWARE
Hardware
Hardware / T-2026-6532
Meta's MTIA 300 is live, but the chip strategy is about leverage, not independence
Meta's MTIA 300 is already deployed, with three more chips on a six-month cadence. The real story is how custom ASICs complement, not replace, the GPU fleet.
SourceMeta rolls out in-house AI chips weeks after massive Nvidia, AMD deals
Hardware / T-2026-8146
Intel's Computex 2026 Bet: The CPU Returns as Agentic AI's Orchestrator
Intel's Computex 2026 announcements, from Xeon 6+ to rackscale AI with SambaNova, argue that agentic inference changes the compute balance of power.
SourceIntel Announces New AI Innovations at Computex — Chip to Rackscale AI ...
Hardware / T-2026-5689
Data Center Hardware in July 2026: The Stack Widens Beyond the GPU
June's top data-center hardware stories show the AI race broadening beyond accelerators to networking, memory, CPUs, and orchestration.
SourceData Center Hardware Highlights: July 2026
Hardware / T-2026-9151
The 370kW Rack: Why Power and Cooling Define AI Infrastructure in 2026
AI rack densities are forcing a rebuild of the data center industry around liquid cooling, grid-scale power, and modular deployment
SourceFuture of Data Centers in 2026: AI, Energy & Cooling Innovations
Hardware / T-2026-0097
Lambda's Balaban on AI Compute 2026: The Real Bottleneck is Shell, Not Silicon
Lambda's Stephen Balaban argues AI compute's real constraint is land and power, not chips, in a new podcast conversation.
SourceAI Compute 2026 with Stephen Balaban of Lambda - LinkedIn
Hardware / T-2026-8334
NVIDIA's RTX Spark Brings a Petaflop of AI Compute to the Windows PC
NVIDIA launches RTX Spark, a 1 petaflop AI superchip for Windows PCs, developed with MediaTek and optimized with Microsoft. First devices arrive fall 2026.
SourceNVIDIA RTX Spark AI Chip Windows PCs | The Silicon Review
Hardware / T-2026-0802
Nvidia's Computex 2026: Jensen Huang Declares War on Every Layer of Computing
Nvidia's Computex 2026 keynote announced six chips, a PC platform, a 500B open model, and a robot. The strategy: own every layer of the AI economy.
SourceJensen Huang Just Announced 6 New Chips, a PC, a Robot ... - FourWeekMBA
Hardware / T-2026-1752
NVIDIA's RTX Spark: The AI Inference Chip That Hides in Plain Sight
NVIDIA unveils RTX Spark, a single-chip fusion of AI and RTX graphics for Windows PCs, signaling a strategic bet on local inference and the edge.
SourceNVIDIA: World Leader in Artificial Intelligence Computing
Hardware / T-2026-8905
Nebius: the quiet European compute play building a multi-site GPU empire
Nebius details six data centre sites spanning Finland, New Jersey, Missouri, the UK, Iceland, and France, with thousands of Blackwell and H200 GPUs for AI workloads.
SourceLearn About Our Innovative Data Centres & Hardware - Nebius
Hardware / T-2026-9560
Nvidia, AMD, Google: The AI Chip Race by the Numbers
A data-driven look at the AI chip market from 2020 to 2030, where Nvidia holds 80% share, AMD triples its slice, and Google's TPUs power half its training.
SourceAI Chips in 2020-2030: How Nvidia, AMD, and Google Are Dominating (Key Stats) | PatentPC
Hardware / T-2026-7710
Intel Bets the Rack on Agentic Inference at Computex 2026
Intel's Computex 2026 announcements shift focus from training to inference, betting on CPU-dense racks and disaggregated architectures for agentic AI.
SourceIntel Announces New AI Innovations at Computex
Hardware / T-2026-8000
The Inference Hardware Bazaar: NVIDIA Still Leads, but the Alternatives Are Real
An enterprise guide to the LLM inference hardware landscape shows NVIDIA at 92% GPU market share, but specialized chips from Cerebras, FuriosaAI, and others are winning real…
SourceLLM Inference Hardware: An Enterprise Guide to Key Players