Radar · Beta
Who supplies AI compute, where the physical chokepoints are, and what's shifting in the fab, datacentre and power layers beneath the model layer.
incumbents vs new entrants · ranked by mentions in compute/infrastructure coverage
fab capacity · HBM · packaging · grid · export controls
If implemented as described, the arrangement could add domestic semiconductor packaging supply capacity in the United States; the report does not establish production volumes or the timing of output.
If confirmed, customer validation could be a step toward Samsung supplying high-bandwidth memory for AI accelerators, with implications for the AI-memory supply chain. Sens.ai has not identified independent confirmation of the reported validation.
Coming soon
A timeline of announced compute capacity commitments — new fab, datacentre and cluster buildouts, tracked by announcement date.
Waiting for: A larger population of items carrying both a computeCategory and a verified announcement date; the fields exist but are not live yet (migration not applied).
Coming soon
Announced AI-related investment broken down by compute category — accelerators, datacentres, fabs, networking, energy, memory, cloud capacity.
Waiting for: The AI desk's own investment extraction (FR-613, shipped this sprint) to accumulate enough rows to chart — today's 155 investments rows are still almost entirely defence-desk.
Coming soon
The split between compute/infrastructure coverage that serves training workloads versus inference/deployment workloads, over time.
Waiting for: workloadPhase (FR-614, shipped this sprint) to be live and populated — its migration is written but not applied, so no item carries it yet.
Coming soon
Which countries are building, exporting, or restricting AI compute capacity — a measured pillar, not a hand-curated one.
Waiting for: Enough geo-tagged compute/infrastructure coverage to measure a posture per country rather than curate one — deliberately deferred rather than shipped as a guess (see the roadmap's excluded-scope note).
If the collaboration yields manufacturable advances, it could help address constraints in AI-chip performance and packaging, but this short report gives no measurable outcome. Sens.ai has not identified independent confirmation of the collaboration details beyond this report.
The discussion will focus on energy, grid connections, data centers, cooling, and related systems that the article says are needed to scale AI. Its claims about constraints and opportunities are prospective framing for a conference session, not evidence of a specific market change.
If delivered, the commitment could expand U.S. access to advanced compute and research support in strategically relevant areas including quantum computing, energy, and healthcare. These are company-announced plans, and the release does not detail the allocation schedule or independently establish resulting capabilities.
The article argues that system-level integration, data movement and energy management are becoming important competitive factors in the AI era, but provides no quantified evidence of resulting performance or market changes.
If pursued, the financing could secure substantial near-term AI compute while the reported Terafab initiative signals an ambition to lessen supplier dependence. The financing remains reported as sought, not completed, and the source gives limited detail on either effort.
The order suggests Hanmi may supply equipment for capacity serving AI server substrates, but the source does not specify the equipment, delivery schedule or resulting capacity.
If this coordination results in expanded or better-integrated infrastructure, it could support China’s ability to meet electricity and connectivity needs for data centres. The excerpt gives no implementation details or scale.
If the forecast reflects sustained demand, it would indicate that AI infrastructure spending is extending to a broader set of chip and connectivity components; the excerpt does not establish the scale or timing.
