Signal & Scale 🕸️

$WYFI @WhiteFiber_ has now accumulated enough missed or still-unconfirmed execution milestones that investors have to start applying a larger management/execution discount to the stock. NOT GOOD. The Aug 12 $WYFI earnings release had a specific management target: full contracted run-rate billing across the 40 MW later in August. September 14 $WYFI presentation: It explicitly says NC-1 initial billing had commenced, but “full revenue contribution is expected during Q3.” It also describes NC-1 project-level financing as expected. This matters because: August target → missed → reset to Q3 → Q3 ended → still no public confirmation from the company. And this is no longer one isolated milestone. The table is getting too long. I remain constructive on the underlying $WYFI opportunity. But management needs to start underpromising and overdelivering. Execution credibility matters. Please & Thank you!

Dec 31, 2024 close → Sep 25, 2026 close (last ~21 months) $IREN +349% $9.82 → $44.13 $CIFR +282% $4.64 → $17.73 $KEEL +155% $1.49 → $3.80 $CLSK +51% $9.21 → $13.95 $WYFI +21% IPO close $16.22 → $19.62 $GLXY +19% $20.39* → $24.24 $BTDR -47% $21.67 → $11.51 Owning power was not enough. The market paid very differently for execution, customer proof, financing, & the speed of monetization.

$WYFI launched WhiteFiber Continuum today. 🗞️ The interesting part is what it does to one of AI infra’s hardest constraints: contiguous power ⚡ Massive +ve impact for: - Krambu 🤝 WhiteFiber - NC-1, 2, 3 billing rate WhiteFiber has demonstrated distributed GPU infrastructure across two data centers 83 km apart, connected over dark fiber, with: 111.2 Tbps aggregate bandwidth 0.9 ms round-trip latency The architecture combines WhiteFiber infrastructure with DriveNets AI Fabric, WEKA storage and high-capacity fiber to make geographically separate GPU infrastructure behave more like one logical cluster. That matters because power increasingly exists in fragments. 30 MW here. 40 MW there. 50 MW somewhere else. If those sites can operate as one logical compute resource, smaller power islands become much more useful to large AI customers. That could materially expand the universe of infrastructure $WYFI can monetize. It strengthens the part of the thesis I find most interesting: operated capacity may scale faster than owned infrastructure 📍 The upcoming $WYFI catalysts remain: - NC-1 full 40 MW run-rate billing - NC-1 permanent project financing - NC-2 / NC-3 leases - First meaningful Continuum customer Continuum could expand what counts as usable infrastructure. But, NC-1 permanent financing tells us how fast $WYFI can compound it.

AI capability: what should we actually benchmark

$KEEL wen deal?

Ad for Bitcoin & Ethereum

The problem with SBF (FTX), Dario (Anthropic), and the rest of EA crew is this 👇 Your child may simply be in their way to help save 75M 🦐 on earth and 🐛🪲 in distant galaxies. After all, they get to choose. This is why we Bitcoin and open source.

Crypto Winter → Crypto Spring

$META's AI capex is $CLSK's contracted revenue. How has the hyperscaler performed against the company now building infrastructure for it?

📍 The market is still underestimating how big the next phase of AI is. 25% of GDP is not thinking big enough. We need more data centers. Every MW. Every chip. Every byte of HBM. The first wave gave us access to knowledge. Then came reasoning and judgment. Now we are moving into agentic labor, where AI can actually execute objectives. The next step is management, where AI coordinates agents, workflows and eventually entire systems.