Research notes and analysis from the Futurum practice areas.
16 results
Brendan Burke, Research Director at Futurum, shares insights from AI Infra Summit 2026, where power-capped data centers pushed vendors toward optics, near-memory silicon, and physical AI ahead of a 2028 co-packaged optics ramp.
Brendan Burke, Research Director at Futurum, shares his insights on AMD's MLPerf Inference 6.1 results and why ROCm's measured rate of improvement under a six-week release cadence is now AMD's most important competitive number.
Brendan Burke, Research Director at Futurum, analyzes how Arm's Agentic AI strategy is expanding by launching CSS for Mobile 2, Neoverse CSS N4, and Arm Total Design for robotics.
Brendan Burke, Research Director at Futurum, examines d-Matrix’s integration of Raptor inference XPUs into NVIDIA MGX racks via NVLink Fusion, testing if 3D-DRAM architecture outperforms Groq’s SRAM-based LPX.
Brendan Burke, Research Director at Futurum, shares insights on IBM’s native Arm mainframe design and the execution tests that remain before deployment.
Brendan Burke, Research Director at Futurum, shares his insights on how NVIDIA Groq 3 LPX strengthens Vera Rubin and what cloud providers must prove before faster tokens support premium pricing.
Brendan Burke, Research Director at Futurum, shares insights on NVIDIA Vera CPU adoption across Grok and Starmind and the evidence required to validate orbital AI computing.
Qualcomm announced a multi-generation collaboration with AWS covering customized silicon for AI inference and optical connectivity solutions extending up to 1.6T.
Brendan Burke, Research Director at Futurum, analyzes QumulusAI’s Q2 FY 2026 earnings, focusing on direct AI compute demand, GPU fleet expansion, and capacity execution.
Brendan Burke, Research Director at Futurum, shares his insights on the Cerebras CS-4 launch and why power delivery and networking now decide whether wafer scale can serve trillion parameter models at scale.
Brendan Burke, Research Director at Futurum, shares his insights on NVIDIA’s exclusive guarantee of OpenAI’s Ohio data center at PORTS-Pike and why credit support is becoming the competitive weapon no rival chipmaker can match.
Brendan Burke, Research Director at Futurum, shares his insights on the $240 million IBM and Together AI agreement and why an anchor-tenant inference cluster moves IBM Cloud to the landlord side of the GPU lease.
Brendan Burke, Research Director at Futurum, shares his insights on 100 days that put Blaize, CoreWeave, NVIDIA, Nokia, Firmus, and Indosat on Indonesia’s sovereign AI map, and the generation-plus-distribution model that can expand globally.
Brendan Burke, Research Director at Futurum, shares his insights on NVIDIA’s $500 billion financing platforms and how DSX reference designs turn AI compute into collateral that banks, insurers, and pension funds can underwrite.
Brendan Burke, Research Director at Futurum, examines NVIDIA’s FMS 2026 storage push and whether openness extends NVIDIA’s platform control to the data path.
Brendan Burke, Research Director at Futurum, analyzes AMD’s Q2 FY 2026 earnings, focusing on Data Center growth, EPYC demand, Helios momentum, and the next phase of AI infrastructure execution.