convex 🐸 🦙

$IREN is the NVDA of AI Power! $CVX king maker $TSLA future proofing “IREN OG”

[Frans Bakker @FransBakker9812]

"there are thousands of people on our sites right now pouring concrete and racking GPUs."

$IREN CEO Dan Roberts speaks the truth.
This OnlyFrans image was captured on Sunday 7/26, and reflects a packed site at 8am.

Horizon 2 has GPUs, and fresh concrete is abundant.

Bullish!

[convex 🐸 🦙 @david00477369]

Last of my powder $IREN

[convex 🐸 🦙 @david00477369]

Jim, how far are non coders away from personal agents that remind us appointments, takers and daily tasks via text or some other medium. Right now I am using a mix of grok open claw and gpt (truly maxing my capabilities of AI, I need the easy button), with WhatsApp and email reminders/messaging but wanting a true personal agents. Haven’t tried perplexity due to cost but are they the closest to delivering a personal assistant type product? What timeline do you think this is mainstream and affordable? Mind sharing any breakthroughs on prompts that increase your productivity?

[IREN @IREN_Ltd]

We're giving hyperscalers, enterprises and AI developers the freedom to consume compute however it suits them, from bare metal through to a fully managed service.

Watch the full @RaiseSummit keynote panel: https://t.co/6n8sdQUpzE

[convex 🐸 🦙 @david00477369]

@_Sgr_A_Star New unit of economics ccpp bwahahaha.

[convex 🐸 🦙 @david00477369]

@bitcoinbutcher1 Respect. 🫡

[convex 🐸 🦙 @david00477369]

$IREN Australia ant 🐜

[robo @roberto45580514]

The australian press reports Anthropic plans to acquire at least 1.4GW of data centres to service their Australian growth plans.
They report Anthropic sent out an RFP asking those parties invited to the beauty parade to respond to a number of questions , including
Do they have access to finance , Have they a history of building 200MW clusters
What does the applicants land bank look like ( ie will it come into contact with residential power requirements)
Do they have the supply chain relationships to deliver a project of this size sequentially over the time frame envisioned . Anthropic reportedly wants 1 GW on line by end 2027 .

At least half a dozen parties were invited to meet anthropic and government officials
Iren amongst them and certainly ticks the boxes for requirements posted above.
Anthropic’s decisions expected/ reported to be made towards the end of this quarter

[convex 🐸 🦙 @david00477369]

$IREN

[NVIDIA AI Infrastructure @NVIDIAAIInfra]

Congratulations to @IREN_Ltd on achieving NVIDIA Exemplar Cloud status on NVIDIA HGX B300 for training workloads. This validation reflects close engineering collaboration and gives enterprises confidence that demanding AI training workloads can perform at scale on IREN’s AI Cloud.

[convex 🐸 🦙 @david00477369]

@TechniCameron ASTS

[convex 🐸 🦙 @david00477369]

@TechniCameron I’ll take some today. Great call

[convex 🐸 🦙 @david00477369]

@bitcoinbutcher1 @warriors Bump. Thanks for voicing sentiment. BB.

[XCap @XCapitalMgmt]

Love this @danroberts0101. Long time shareholder and will be for many years to come.

This feels like going on offense, but there are many ways to go on offense spending before revenue hits and it's not clear why we're on offense in this area, but not others, specifically:

*Timing & Guidance For Early 2027 Childress Air Cooled Capacity - Demand must be off the charts for the rest of air cooled capacity at Childress. Why haven't we ordered the GPUs? Does the $NVDA partnership speed those delivery timelines up? Could we move faster than early 2027 on the retros or is labor totally dedicated to Horizon? Why are analysts not modeling the full value of that site (both in terms of topline growth and the relatively low retrofit capex) - are they not getting guidance on those fronts?

*Desired Customer Base For Sweetwater As DSX Flagship - Sweetwater as the $NVDA flagship DSX campus is absolutely incredible. But what does that mean exactly? Are the advanced discussions (12-24+ months) still happening, or does DSX flagship mean you're now targeting a different customer base entirely? Laying out the vision for what customers that campus serves as the DSX flagship would be extremely compelling, particularly as the vision's clearly evolved from whole-of-site tenants in 2025 to what appears to be something much more dynamic now.

Know you're slammed, but hopefully the above will be laid out more clearly at August earnings. Love the brand recognition efforts, but there's no better magnet for eyeballs than potential economics and the questions above are looming heavy in the minds of all stakeholders.

Thank you to you and the team for all the hard work on behalf of shareholders. Excited for H2 and 2027 (and to buy an $IREN Curry jersey).

[convex 🐸 🦙 @david00477369]

@XCapitalMgmt @TMTLongShort Ender and Bean

[convex 🐸 🦙 @david00477369]

$ASTS

[TechniCam @TechniCameron]

$ASTS reclaiming the 50WMA at the apex of this weekly descending wedge. No we really want to see a daily close above the 200DMA and 50DMA next:

[convex 🐸 🦙 @david00477369]

@dannycheng2022 Position size trumps timing

[convex 🐸 🦙 @david00477369]

I’m in

[Dulce @litigious_dulce]

$NUAI — Second bite at the data-center-colocation trade
Who this is for: investors who traded the 2024–2025 public data-center cohort ($WULF, $APLD, $HUT, $CIFR, et al.) and already understand how these names re-rate on a first hyperscale lease. The argument below is simply that NUAI is the same setup, one stage earlier, into a demand backdrop that's now proven rather than speculative. The key points are:

· The high-conviction bets this cohort made on data center stocks in 2025 paid off massively, and they were placed when AI demand was far more speculative than it is now. Those names have since re-rated and the thesis is validated. NUAI is the chance to run the same playbook again — pre-lease, power-rich, hyperscaler in the room — except into a confirmed demand backdrop rather than a hoped-for one. The setup is clean and squarely inside a risk profile you already operate in.

· You don't need to redeploy out of the maturing names to capture the convexity here; a modest position can still be highly rewarding given the asymmetry. Small sizing is also better trade construction: it lets you dollar-cost-average through volatility, keeps dry powder for the July tape, and — underrated — keeps you emotionally detached enough to execute well when the chart is whipping around. Position size is risk management and behavioral management.

---------------------------------------------------
For investors who already own the operating cohort at valuations that arguably discount execution, New Era Energy & Digital offers exposure to the pre-contract phase of the same playbook. The relevant question isn't whether NUAI is "good" — it's whether the risk-adjusted payoff justifies a small, ring-fenced allocation. I think it does, with discipline around sizing and entry.

The setup. TCDC: ~650 MW gross secured in Ector County, TX, with 1+ GW potential. NUAI now holds 100% post the Sharon AI buyout. Per Northland ($11 PT), the site is in advanced commercial discussions with a top-four hyperscaler. The investable thesis therefore rests on the following considerations:

1. Migration to a fully islanded, behind-the-meter campus. BTM generation removes the ERCOT batch study, interconnection queue risk, ratepayer politics, and regulatory timeline — converting the dominant schedule risk into procurement/construction risk, which is more controllable. June 11 Odessa Development Corp commentary corroborated tenant urgency and the dedicated-generation path. In a market clearing on speed-to-power, this is the right architecture.

2. A credible Phase 1 power source. Northland's diligence ties Phase 1 (~207 MW) to Vistra/Luminant generation adjacent to CIFR's Odessa site, whose PPA rolls off ~July 2027 — synchronizing with NUAI's targeted delivery. Existing substation infrastructure underpins the "no significant electrical infrastructure required" claim.

3. Counterparty quality. Stream (Apollo-controlled, $40B) as developer/operator, Apollo the presumptive institutional equity arranging ~80% project debt, and Macquarie's $290M project facility. The JV was, per the analyst, hyperscaler-directed — a meaningful signal about deal seriousness.

4. Personnel as a stage-of-development tell. CDO Evan Pierce (5+ GW deployed; prior sole-BTM delivery to the same hyperscaler) and GC Michael Johnson (ex-CoreWeave, ex-Switch). Execution and transaction-closing hires, not narrative hires.

5. Legal overhang cleared. Agreement for full dismissal of the New Mexico action announced May 28 removes a discrete tail risk and unblocks legacy-asset monetization.

6. Catalyst path is time-boxed. The Macquarie facility requires lease execution by Oct 8, 2026; a ~14-month build for pre-YE27 energization implies a signature by early Q4. Lender covenant, tenant timeline, and PPA roll-off converge on the same window. You're not underwriting an open-ended story — there's a forcing function and a defined catalyst date.

7. Valuation bridge. Recent HPC leases clear $140–$190/kW/month. Phase 1 alone (133 MW IT, 15-yr NNN, ~45% NUAI economics) supports analyst floor work of ~$3.55–$8.16/share; Phase 1 + half of Phase 2, ~$6.80–$15.64. The $11 PT credits 283 of ~933 MW potential. The analytical crux: ~$6.25 today embeds legacy-E&P-plus-option value; a signed Big-4 lease re-bases the multiple to contracted-infrastructure comps (private marks ~25x NOI; public REIT/infra peers ~19–29x EBITDA). And critically, every figure above is TCDC-only — it credits a fraction of one campus and assigns zero to the rest: remaining TCDC phases toward 1+ GW, the ~7 GW New Mexico site, and management's stated multi-site pipeline. The first lease isn't just a TCDC re-rate; it's the proof-of-execution that turns a single-asset story into a repeatable platform, and in this cohort deal flow tends to accelerate sharply once the anchor signature de-risks the developer. The floor math understates the option value precisely because it stops at TCDC.

8. Financing / dilution. The market reflexively discounts micro-cap developers for dilution risk, but the playbook this whole cohort converged on (e.g., project-level non-recourse debt) is built to fund the build without repeatedly hitting the corporate equity line. NUAI is pointed the same way: Macquarie's facility sits at the TCDC SPV, the Stream/Apollo JV is expected to carry ~80% project debt plus institutional equity, and NUAI's land contribution offsets part of its cash equity. The dilutive cleanup (SharonAI overhang, going-concern) is largely behind them, which strongly implies funding from here can be minimally dilutive at the parent — the difference between compounding equity value per share and treadmilling it.

9. Flow catalyst. NUAI is on FTSE Russell's preliminary additions list for the Russell 3000 and Russell Microcap indexes, effective at the open June 29. In a thin, ~84%-float micro-cap, mechanical index buying can move price independent of fundamentals — and the post-event flow vacuum is a known reversal risk.

10. EV, not conviction. Illustratively — subjective inputs, not a model — assign ~45–55% to a lease landing within ~12 months, a re-rate toward the lower-mid of the floor range on success, and a substantial (not total) drawdown on failure/dilution/slippage. Even at conservative odds the convexity is positive: the downside is partially floored by legacy assets and site value, while the upside is a multiple of spot. The asymmetry is the reason to hold it — and the wide variance is the reason to keep it small.

---------------------------------------------------
How much of a deal is already in the price? Strip NUAI to its deal-agnostic floor. Assume TCDC land at $100M, a deliberately low ~$15M on the legacy E&P assets (below even the analyst's ~$19M low case), and ~$60M cash per the June report, less modest drawn project debt. That's roughly $150M of identifiable asset value — ~$1.20–1.50/share on ~125M fully diluted shares. Everything above that is the market paying for commercialization it hasn't seen yet. At ~$6.18 / ~$595M, that means ~$150M is asset-backed and ~$445M (~75% of the cap) is pure deal optionality. This is not a stock where a lease is "priced in" — it's a stock where little else is.

Now treat the lease as the binary it is. If Phase 1 signs and NUAI re-rates to the $11 PT, and if it fails back to the ~$1.35 floor, the current ~$6.15 solves to an implied ~50% probability of a Phase 1 deal — a coin flip. That's the first insight: at today's price you're paying roughly fair odds for the first leg, nothing more.

The second leg is where the asymmetry lives. A Phase 1 signature doesn't just validate 133 MW — it makes leasing the rest of TCDC effectively inevitable, because the same hyperscaler, the same islanded-power solution, and the same JV machinery are now proven and in place. The analyst's own sensitivity table reflects this: 283 MW supports ~$11, but ~433 MW supports ~$17, and fuller TCDC credit runs higher — so a ~$20 post-signing fair valueis reasonable once the market extrapolates from "one phase" to "the campus." Re-run the binary against $20 instead of $11, and the current price implies only a ~26% probability of success — i.e., the market is under-pricing the conditional outcome. Put differently: if you handicap the lease at ~50/50 and believe the signed-state is worth ~$20, fair value today is ~0.5 × $20 + 0.5 × $1.35 ≈ $10.70, ~75% above spot.

And ~50% is almost certainly too low as a real-world estimate — it's just what the tape implies. Per the June 15 Northland note, the hyperscaler is actively prioritizing TCDC as other Stream developments slip, is pushing for pre-YE27 delivery, and is steering the project toward the islanded BTM design that strips out the interconnection risk — that's a counterparty leaning in, not kicking tires. But the more important point for handicapping is that the bet does not rest on this specific lease. The binary above treats "deal" as one tenant signing one contract; in reality NUAI has redundancy on both sides of the table. On the demand side, every hyperscaler is power-starved, so a powered, shovel-ready Permian campus with secured generation is a scarce, fungible asset — if the current counterparty stalls, the site doesn't lose its value, it just changes buyers. On the supply/execution side, NUAI isn't even locked to a single developer: neighbors like CIFR and APLD operate adjacent Odessa capacity, understand the power topology intimately, and carry their own hyperscaler relationships — any of them could plausibly step in as developer/operator and accelerate a deal rather than restart one. So the failure case isn't "no deal ever," it's "a slower or differently-structured deal," which is a far smaller haircut to the floor than the binary assumes.

The takeaway: the market is pricing a coin-flip to $11, but the true odds of somecommercialization are higher than 50% — the tenant is already prioritizing the site, and even a broken deal likely reroutes to another power-hungry hyperscaler or a neighboring developer rather than to zero. That means you're getting two things underpriced at once: the probability of a signature andthe $11→$20 second-order re-rate once the rest of TCDC becomes obvious. Your edge isn't betting the lease happens; it's that both the odds and the conditional payoff are higher than the tape is crediting.

Bottom line. Risk is a function of two variables, not one: the probability of an outcome and the price you pay for it. The mature names are genuinely "safer" on probability — APLD, CIFR, HUT have already signed large hyperscale offtakes, so the binary NUAI still faces is, for them, resolved. But that safety is fully priced. TeraWulf is already ~$14.5B, Applied Digital ~$13B, Cipher ~$12B. For any of them to 5x means underwriting a $60–75B large-cap infrastructure outcome — possible, but a heavy lift, and the explosive part of that move is behind you.

NUAI is the inverse: lower probability, dramatically cheaper price — and that price is what makes a 5x still fully warranted here. At ~650M, a 5x is only ~$4B. That's not aggressive; it's roughly where the market values the *pre-contract* version of this exact story today. Fermi ( FRMI), a grid-independent AI-power campus developer with no binding tenant yet, carries a ~$6B market cap — and it IPO'd at roughly $16B with zero revenue or operating history. So ~$4B for NUAI doesn't require it to out-execute the cohort; it requires NUAI to simply be recognized as a credible developer with a hyperscaler lease — at which point it would still sit below pre-deal Fermi and at roughly a quarter of the $12–15B names that have already signed. A signed Big-4 lease re-rates the stock to $15+ as a conservative initial marker (floor-valuation math), with the path to the full 5x opening as the broader platform — remaining TCDC phases, New Mexico, follow-on sites — gets credited. The asymmetry is the entire point: you're paying micro-cap price for an asset the market repeatedly values in the billions before the contract even prints.

That configuration — institutional-grade validators (Apollo/Stream, Macquarie), a valuation still embedding mostly option value, and a setup this audience has personally traded to profit — is rare, and it only sharpens as the Oct 2026 lease window narrows from "someday" to "this quarter." The move is further amplified by mechanics this cohort knows cold: heavy short interest, anticipation of follow-on deals, and passive/index demand into the June 29 Russell add.

---------------------------------------------------
One practical point on building a position in NUAI. A low cost basis isn't just nice math — it's the mechanism that lets you actually stay in the trade long enough to collect the payoff. These names do not re-rate in a straight line; volatility is the toll, not the detour. Say the lease prints and NUAI runs to $20. It will, at some point, draw down to $14 — a ~30% pullback is not the exception in this cohort, it's the near-certainty. The entire difference in how you experience that move is your basis.

If you formed the bulk of your position in the $5–8 range, a slide from $20 to $14 is a wobble inside a triple. You're still up ~2x; you hold easily, and you probably add. If you chased the breakout and bought at $20, that exact same $14 print is a 30% loss staring at you in red — and that's precisely the moment investors capitulate, selling the bottom of a pullback in a stock that's still working. Same security, same price, opposite outcome — decided entirely by where you got in. Loss aversion is real, and it does its worst damage to high-basis holders. So accumulate small and DCA in the $5–8 range, keep dry powder for the seasonally soft July tape, and build the position before the move rather than into it. The low basis is what converts a correct thesis into a return you actually keep, instead of a great call you fumbled on a routine drawdown.

[convex 🐸 🦙 @david00477369]

@TheBigDegen Today we are not mates. Tomorrow back to TEAM $IREN

[convex 🐸 🦙 @david00477369]

If I’m being honest, I was never a fan of the company due to their CEO being outspoken on politics. He is there for business decisions not political drama. However, this ain’t a hobby. We are here to make money so I went in in February with Commons and March with Calls.

That said, they are fundamentally a good company with a great future outlook. They have played it well. I’m here for the ride.

Largest position by far is IREN

[convex 🐸 🦙 @david00477369]

@TechniCameron Substrate looking good too

[convex 🐸 🦙 @david00477369]

@dannycheng2022 @cantonmeow Where is resistance? Correction coming off such a 🚀?