xAI’s space data centers, brain chips & big batteries

xAI reorganized into four product teams – focused on Grok (language models), code generation, media (image/video) and autonomous agents (“Macrohard”).

Co-founder turnover accelerated – at least six of 12 co-founders have left post-merger; Musk framed the move as necessary for “execution at scale.”

Orbital compute ambitions escalated – management floated plans for 100–200 GW of orbital power on the path to 1 TW of off-planet compute, plus a 2 GW “Colossus” data-center complex in Tennessee.

The combined “SpaceX + xAI” stack now looks like a multi-product AI studio welded to vertically integrated space and connectivity, not a single chatbot business.

Off-planet compute would shift constraints from terrestrial power and cooling to launch cadence, in-space manufacturing and orbital servicing.

Governance and key-man risk at a $1.25T private company will be central to how any eventual IPO is underwritten and priced.

Treat the Musk ecosystem as exposure to launch, satellites, broadband, AI models and payments/social in one capital stack, rather than a narrow bet.

Orbital data-center plans introduce new dependencies on insurers, in-space servicing, debris-mitigation rules and export controls.

Second-order winners and losers—in power, chips, launch competition and in-space infra—may matter as much as the headline listing.

CorTec reported a second Brain Interchange implant at Harborview Medical Center in Seattle under an FDA Investigational Device Exemption (IDE).

The first participant (implanted in 2025) has shown “notable neurological gains”, encouraging investigators to add more patients.

The system records neural activity and delivers targeted stimulation in real time, tuned for upper-limb motor recovery after stroke.

Moving from a single first-in-human case to multiple implants is a key threshold for any implantable neurotech platform.

CorTec is positioned in “neurorestoration” (stroke recovery) rather than only communication/typing BCIs, broadening its potential clinical footprint.

NIH backing + FDA IDE gives a regulated path toward reimbursement and guideline discussions if later data are strong.

Near term, diversified exposure likely sits in enabling components—electrodes, hermetic packaging, surgical tools and AI signal-processing IP.

Expect increasing overlap between therapeutic and assistive BCIs as platforms prove themselves in clinical rehab and then expand indications.

Human data velocity—safely scaling from 1–2 implants to dozens—is a more durable moat than any single headline.

EVE reports the Ruite New Energy Lingshou project, using 628Ah LFP cells, is now operational and claimed as an industry first for grid-scale storage.

The plant is 200 MW / 400 MWh, using 80 “Mr. Giant” battery containers and 40 converter cabins, reducing connection points vs. smaller-cell designs.

EVE has shipped similar systems overseas (including to Australia and Europe) and signed a 10 GWh supply agreement with Guowang Technology.

Ultra-large cells target lower levelized cost of storage, cutting hardware complexity, cabling and footprint, especially for 4-hour+ projects.

Larger cells also concentrate risk—a failure or thermal event can affect more capacity—putting more weight on safety engineering and insurance.

The project shows how quickly Chinese BESS manufacturers are iterating, with 600–1000+ Ah cells and >10 MWh containers becoming the new frontier.

Grid batteries are shifting from bespoke EPC projects to a productized infra category; fleet reliability and software may be the real differentiators.

Operators with strong monitoring, fire-safety and data-driven maintenance should be better positioned to handle ultra-large cells.

For portfolios already exposed to AI data-center power demand, these storage deployments are part of the same firm-capacity story.

RIE2030 raises research spending to around 1% of GDP annually, a 32% increase over the prior package, with quantum billed as an “early and deliberate bet.”

Singapore will host Quantinuum’s latest quantum computer, giving local researchers and startups direct access to state-of-the-art hardware.

A quantum startup co-founded by Nobel laureate John Martinis is partnering with Singaporean labs, deepening the local ecosystem.

Quantum is being treated as national infrastructure, similar to earlier bets on semiconductors and biomedicine.

On-shore access to cutting-edge systems shortens the loop between research, prototyping and commercial pilots in finance, logistics and materials.

As a high-trust, relatively neutral hub, Singapore is positioning itself as a key node in Asia-Pacific quantum infrastructure.

Exposure will often come via regional hubs where public R&D, talent and infrastructure are co-located, not only via hardware makers.

Quantum-adjacent B2B software—optimisation, security, developer tools—may see nearer-term traction than universal machines.

National strategies like RIE2030 are strong leading indicators of where labs, cloud regions and joint ventures are likely to cluster.