Late Stage Capitalism retweetledi
Late Stage Capitalism
61.1K posts
Late Stage Capitalism
@GlobalCollapse
Physician. 30yr investor. Jan 2026 Top 5: 1. IREN/IRE 2. APP/APPX 3. DGXX 4. DMGGF/DMGI 5. ZMDTF Honorable Mention: CIFR, BITF, TSM Longshots: MIGI TGEN
USA Katılım Eylül 2011
736 Takip Edilen5.8K Takipçiler
Late Stage Capitalism retweetledi
Over $50 million into these $IREN 110 calls now. Something brewing?
Flow God@FL0WG0D
$20.7 million into these $IREN calls
English
@CKCapitalxx Unlikely to be 150-200M at a tier 2 level. If they upgrade to tier 3 that will be significant CAPEX
English
One of the reasons i love $VIVO is that they already controls three sites.
This is what they could be worth.
Norway is operational today. 41.5MW. $31 million in annualized revenue already running. Pending 40MW expansion that could take the site to over 80MW. At AI compute lease pricing that single facility could generate $150 to $200 million annually.
Finland is 291MW of secured powered land already controlled. At hyperscaler AI compute rates that pipeline represents $800 million to $1.5 billion in potential annual revenue at full buildout.
UAE is a 25MW sovereign AI data center platform already in place. At AI compute pricing that facility alone could generate $50 to $80 million annually once fully tenanted.
Revenue estimates on assets they already control today:
2026: $31M
2027: $45-60M
2028: $100-200M
2029: $300-500M
2030: $700M to $1B+
One site already cash flowing. Two more with hundreds of megawatts behind them.
All of this on a $80 million market cap.
CK Capital@CKCapitalxx
Opened a new long position in $VIVO today. This one took some digging. Let me walk you through the full picture. I originally was going to wait to put this post out but seems like the market just now found this gem this week as it’s up 50%. $78 million market cap. And I think this is just the beginning. On April 21 VivoPower closed a $41 million acquisition of an operational 41.5MW data center in Mo i Rana Norway. Powered by 100% renewable hydroelectric energy at below $0.035 per kilowatt hour. One of the cheapest power costs for data centers anywhere in Europe. Zero equity raised. Zero dilution. Paid at a disciplined 4x EBITDA multiple. The company flipped from EBITDA negative to EBITDA positive in a single transaction. $31 million in annualized revenue and $10 million in EBITDA. Not projected. Operational today. And then six days after closing they announced something most people missed. The facility prequalified 30MW into Statnett’s ancillary services markets. The Nordic grid pays data centers to hold flexible load capacity in reserve. $1.9 million in additional annualized EBITDA. Zero incremental capex. Pure margin on top of existing operations. Three independent revenue streams from the same single asset. AI tenant lease in active tender with strong hyperscaler inbound interest. Grid demand response already paying. Waste heat district heating in feasibility. One facility. Three revenue lines. None cannibalizing each other. Now here is what sits behind it. Norway has a pending 40MW expansion. If it clears the site goes from 41.5MW to over 80MW. When a hyperscaler signs a long term AI compute lease the revenue per watt jumps 3 to 5x over standard hosting rates. That single facility at full capacity could generate $150 to $200 million in annual revenue. Behind Norway: 291MW of secured powered land in Finland. A 25MW sovereign AI data center platform already in place in the UAE. Power-secured land in constrained markets that hyperscalers cannot easily replicate. Tembo EV spin on deck at a targeted $838 million valuation. Nasdaq already approved the ticker TEMB. When that closes it strips $8 million in annual overhead from a company generating $10 million in EBITDA. Here is what the market is almost entirely missing. Management published a formal 10 year strategic growth plan with specific targets. By 2029: 2GW of controlled power sites. $1 billion in annual revenues. $200 million in operating free cash flow. By 2033: 10GW of power sites. $3 billion in annual revenues. $1 billion in operating free cash flow. By 2036: 20GW spanning EU, GCC, ASEAN, and Africa. $5 billion plus in annual revenues. A $78 million company with management formally targeting $1 billion in revenue by 2029 on assets they are already controlling today. Board bought 2.65 million shares personally. Terminated the $180M dilution shelf. Former Microsoft Global AI leader on the advisory council. GCC sovereign family offices backing the financing. The risks are real. Management has a history of big promises. The hyperscaler lease is not signed yet. The float is tiny. But $78 million market cap. $31 million in current revenue. Three revenue streams from one facility. $838 million spin on deck. Management targeting $1 billion by 2029. The asymmetry here is one of the more interesting setups I have seen in the small cap space in a long time. Long $VIVO. Not financial advice. DYOR. $VIVO
English
Late Stage Capitalism retweetledi
$IREN closes $3 billion convertible notes offering.
The deal terms:
> IREN pays just 1% interest per year (super cheap loan).
> The loan must be paid back by 2033.
> Investors can swap their IOU for IREN stock, but only if the stock price climbs about 32.5% above today's price.
> The "capped call": IREN spent $201 million on a side deal to protect existing shareholders. Normally, when IOUs convert to stock, it creates more shares and waters down everyone else's ownership. This side deal cushions that effect up to a stock price of $110.30.
Bottom line - IREN raised a huge pile of cheap cash to grow their business, using a clever structure that delays dilution and bets on their stock going up.
I believe a deal is imminent. 🧐
English
@skeleGG_ 73 is also going to be a big resistance I think that's where the capped call on the last convertible ends
English
@GlobalCollapse 70$ gonna be the new floor soon. No dilution until 110$ on their latest offering.
English
Between this, the other dark pool buys, and the call purchases it really does seem like something is going on... the big money knows something. I don't usually pay much attention to this but the timing is suspicious to me. $IREN
𝒰𝓂𝒷𝒾𝓈𝒶𝓂@Umbisam
$IREN 🤔
English
@SylentTrade The problem I have with that logic is if big news is coming then why price the convertible before releasing the news?
English
$IREN just raised 3bil at a 1% rate and are limit dilution until $110. Best finance team in the sector and it’s not even close. Now the question you have to ask yourself is why raise 3bil from notes, 1 bil so far from atm, and already have 1.5bil in cash on hand when there is no major “news” coming. I expect a massive headline from IREN which will make the god candle we got from the nvidia pump look like peanuts in comparison.
English
$CBRS IPO price was $185 which is a P/E of 135 or so... at $320 its a PE of over 225. Even at 75% growth that seems nuts. In one month it will be close to $200.
English
@stocks4545 No reason for them to hit the ATM at the same time they close a $3B convertible note IMO. Unless they have more cooking than we know about the $3B convert plus the $2.6B cash on hand they have is more than enough for a while. The ATM should be a last resort.
English
@CWB_Research Awesome post, would love to know your thoughts now that the earnings are out and if you think everything played out as you thought... I dont understand the nuance as well as you clearly do so would love your insight on it.
English
$DLO reports Q1 2026 tonight. I expect a big volume beat, a small take rate miss, and margin pressure the market will misread. Here is the full picture.
Explanation of Recent Price Weakness
First, I would like to address the recent weak price action in dLocal. After reaching a multi-month high of $14.03 on May 6, 2026, $DLO stock is down 12%. I view the catalyst as MercadoLibre's earnings on May 7, 2026. Specifically, $MELI beat on revenue ($8.85B, up 49% YoY) but missed on EPS as margin compression from their credit book and explicit take rate reductions weighed on profitability. Management disclosed they had lowered take rates in certain e-commerce categories — a competitive move designed to defend market share against Shopee and Amazon in Brazil — and the market extrapolated that read across to dLocal, another high-growth, EM-focused business with declining take rates.
For reasons outlined below, I believe that dLocal, like MercadoLibre, could miss on EPS this afternoon (but for very different reasons). Hopefully the recent 12% decline in $DLO stock demonstrates that the market has largely priced in such as miss.
Earnings Preview
Consensus sits at $331–334M in revenue and $0.17 EPS. The options market has priced in a +/– 10% move. My read (seen in the attached table) is more constructive on volume than the Street — and more sober on margins. Neither conclusion changes the thesis.
The number I want to focus on going into tonight is TPV. Q4 2025 was the strongest quarter in dLocal's history at $13.1B — up 70% YoY and 26% sequentially from Q3's $10.4B. A print of $12B in Q1 would represent a natural seasonal step-down of roughly 8% QoQ, which is consistent with Q1 historically being softer than Q4 across the payments industry as consumer and merchant volumes normalise after the holiday period. On a YoY basis however, $12B against Q1 2025's implied base of roughly $8.5B would represent approximately +41% YoY growth — comfortably ahead of the Street's $9.5–10B range and tracking well ahead of the pace required to hit the full-year 50–60% TPV guidance. The market is anchoring too conservatively on sequential deceleration without fully accounting for the step-change in merchant rollouts that drove the Q4 surge. Those merchants do not go away in January.
Revenue will beat for the same reason, but the beat will be partially offset by a take rate that continues to drift lower than consensus models. I expect take rate to come in around 2.40–2.44% versus the Street's ~2.46%. This is not a warning sign — it is the deliberate, documented output of dLocal's volume discount strategy, and it is worth understanding precisely because it is the lens through which the market will interpret this print negatively.
Here is how dLocal's pricing actually works. When a global merchant — say a major streaming platform or ride-hailing company — scales its volume with dLocal across 30 markets, they reach contractual thresholds that trigger lower per-transaction rates. The merchant pays less per dollar processed; dLocal processes dramatically more dollars. The take rate percentage falls; the gross profit dollars grow. This is not margin erosion — it is a volume flywheel that every scaled payments business in the world runs. Adyen does it. Stripe does it. The difference at dLocal is that the markets they operate in are growing so fast that the volume growth more than compensates for the rate compression, often by a wide margin. The correct way to read a take rate miss alongside a TPV beat is not "pricing is under pressure" — it is "the merchant mix is shifting toward the exact high-volume accounts that make this business compoundingly valuable." Gross profit dollars are the metric that matters. If those come in ahead of the revenue beat, the quarter is strong regardless of what the take rate line says.
Operating margins will disappoint, and the market may react poorly to that on first read. My expectation of 17–18% sits below the Street's 19–20%, and below the 20.1% dLocal printed for the full year 2025. But this is precisely what management guided. The 2025 hiring cycle was heavily backloaded into H2, meaning the full annualised cost of those hires hits the income statement in H1 2026 comparisons for the first time. The cost base has not expanded — the quarter in which it becomes visible in the YoY comparison has simply arrived. This is the OpEx overhang management flagged explicitly on the Q4 call. It is temporary, structural, and already accounted for in my DLO model.
Why an EPS Does not Weaken my Conviction
The operating leverage story in dLocal does not begin until H2 2026. Management said so. The Q3 2026 earnings report — expected in November — is the first moment when the annualisation overhang fully clears and the underlying operating efficiency of a business growing 50%+ on a fixed cost base becomes visible in reported numbers. Until then, every quarter is going to look like a mixed bag to anyone running a short time horizon: strong volume, lower take rate, margin pressure.
This is why I continue to review Q1 and Q2 2026 as an accumulation window for dLocal. I added shares last week, increasing dLocal to a 17% weight in my portfolio. At current levels I will continue adding and am willing to increase my position in dLocal to a max 25% portfolio weight.
English
@james_j_otis It started before the converts... and that hedging was all done when the converts closed, the day $IREN dropped $4. So I dont agree...
English
@GlobalCollapse This is hedging against converts! "the big money knows something." ... no they do not, just managing greeks.
English
@Billions777888 Nah that doesnt Jive. Horizon 6-10 would be 300MW power, 200MW IT load. Plus Horizon 6-10 is basically tabled if they are converting existing BTC mining ops to air cooled cloud.. that means they wont be building Horizon 6-10.
English
$IREN I know this sounds completely crazy but this is a code that a deal is incoming. Mike is on the IREN board so he can't just say it directly
English
@CKCapitalxx Oops... was a good play a month ago but sucks to be anyone that just bought it in the last 3 days. It'll recover though, still cheap at these levels. I sold some $6-7 June puts in the last hour.
English
Tomorrow is going to be a big day.
$CBRS IPOs tomorrow and $DGXX is my favorite proxy.
Most people are going to try to chase Cerebras on day one. The IPO was oversubscribed, raised from $115-$125 to $150-$160, upsized to 30 million shares, and Polymarket has it closing above $50 billion on day one.
Getting in at a good price on the IPO itself is nearly impossible for retail.
So instead of chasing it I am watching $DGXX.
Digi Power X signed a 10 year $1.1 billion colocation agreement directly with Cerebras to build and operate their AI data center campus in Columbiana Alabama.
Up to $2.5 billion in total expansion options. Phase 1 operational by December 15 2026.
A $50 billion AI chip company going public tomorrow. The company building their infrastructure has a $550 million market cap.
When $CBRS starts trading tomorrow morning and every financial headline is about the hottest AI IPO of 2026 the market is going to start mapping who actually builds what Cerebras runs on. That connection gets made fast.
The re-rate for $DGXX does not happen gradually. It happens in real time.
Tomorrow is going to be interesting.
$DGXX $CBRS
English
Late Stage Capitalism retweetledi
$IREN, $NBIS, $CRWV, $CIFR
It is estimated that roughly 30% to 50% of data center projects planned for 2026 in the United States are facing cancellation or significant delays.
And I expect this to continue.
Farm Girl Carrie 👩🌾@FarmGirlCarrie
This is what local accountability looks like: In Festus, Missouri, a town of about 14,000 people, the city council quietly approved a $6 billion Ai data center to be built on 360 acres just north of Highway 67. Residents say they were never properly heard. Meetings were held in private. Documents were released too late. A week after the approval, the town held a regular election. Voter turnout jumped 129 percent. Every single council member who had voted yes lost in a landslide. A 70-year-old first-time candidate beat an 8-year incumbent by 40 percentage points. Now a recall petition is circulating to remove the mayor as well. The lawsuit against the city is already filed. Has your local government ever been held accountable like this? 🔥
English
@DustinHuntwn The $DGXX number is because they pay for the power, it isnt a passthrough like most. The rate $DGXX got is in line with every other deal for the most part once you factor that in, so dont expect anything higher for $KEEL. I own both.
English
$KEEL If Sharon gets $230 per kW per month like DGXX, that’s $304 million in annual revenue. BlackRock looks very likely to sign on and appears willing to pay even more than a hyperscaler. So we’re talking two enterprise deals at the smaller sites plus one hyperscaler deal at Panther Creek. That easily pushes total revenue over a billion dollars a year, with most of it dropping straight to profit.
The current 600-share valuation looks high compared to peers, but these contracts should more than justify it. Ben likes the stock because he sees it trading at least $50 within a year or two.
Like if you STILL believe $KEEL .
You can follow me,I'll be updating you with the latest info on KEEL and more related content.
Dustin@DustinHuntwn
$KEEL Ben is one of the smartest ceos around. He actually puts in the work unlike Jason Les and Fred Thiel who steal shareholder money and do nothing. Ben said location is most important factor for these tenets. No doubt they will sign 3 leases that are A+++ this year. He confirmed what he has been saying all along with the cooler temperatures and closer to huge hubs in the NE and W coast with fiber optic cables. The leases signed in Texas are for AI training models and not actually using the AI applications. He also said the demand was so high at Mose Lake they didn't even need to offer gpus as a service, which now confirms to me that they should be able to beat DGXX $230 kw per month deal. That's exactly why they threw out buying their own chips. The demand is so high that they didn't need to do what Iren did. Think about that. Like if you AGREE. Finally,you can follow me,I'll be updating you with the latest info on KEEL and more related content.
English
Late Stage Capitalism retweetledi
Opened a new long position in $VIVO today.
This one took some digging. Let me walk you through the full picture.
I originally was going to wait to put this post out but seems like the market just now found this gem this week as it’s up 50%.
$78 million market cap. And I think this is just the beginning.
On April 21 VivoPower closed a $41 million acquisition of an operational 41.5MW data center in Mo i Rana Norway. Powered by 100% renewable hydroelectric energy at below $0.035 per kilowatt hour.
One of the cheapest power costs for data centers anywhere in Europe. Zero equity raised. Zero dilution. Paid at a disciplined 4x EBITDA multiple.
The company flipped from EBITDA negative to EBITDA positive in a single transaction. $31 million in annualized revenue and $10 million in EBITDA. Not projected. Operational today.
And then six days after closing they announced something most people missed. The facility prequalified 30MW into Statnett’s ancillary services markets.
The Nordic grid pays data centers to hold flexible load capacity in reserve. $1.9 million in additional annualized EBITDA.
Zero incremental capex. Pure margin on top of existing operations.
Three independent revenue streams from the same single asset. AI tenant lease in active tender with strong hyperscaler inbound interest.
Grid demand response already paying. Waste heat district heating in feasibility. One facility. Three revenue lines. None cannibalizing each other.
Now here is what sits behind it.
Norway has a pending 40MW expansion. If it clears the site goes from 41.5MW to over 80MW. When a hyperscaler signs a long term AI compute lease the revenue per watt jumps 3 to 5x over standard hosting rates.
That single facility at full capacity could generate $150 to $200 million in annual revenue.
Behind Norway: 291MW of secured powered land in Finland. A 25MW sovereign AI data center platform already in place in the UAE. Power-secured land in constrained markets that hyperscalers cannot easily replicate.
Tembo EV spin on deck at a targeted $838 million valuation. Nasdaq already approved the ticker TEMB. When that closes it strips $8 million in annual overhead from a company generating $10 million in EBITDA.
Here is what the market is almost entirely missing.
Management published a formal 10 year strategic growth plan with specific targets.
By 2029: 2GW of controlled power sites. $1 billion in annual revenues. $200 million in operating free cash flow.
By 2033: 10GW of power sites. $3 billion in annual revenues. $1 billion in operating free cash flow.
By 2036: 20GW spanning EU, GCC, ASEAN, and Africa. $5 billion plus in annual revenues.
A $78 million company with management formally targeting $1 billion in revenue by 2029 on assets they are already controlling today.
Board bought 2.65 million shares personally. Terminated the $180M dilution shelf. Former Microsoft Global AI leader on the advisory council. GCC sovereign family offices backing the financing.
The risks are real. Management has a history of big promises. The hyperscaler lease is not signed yet. The float is tiny.
But $78 million market cap. $31 million in current revenue. Three revenue streams from one facility. $838 million spin on deck. Management targeting $1 billion by 2029.
The asymmetry here is one of the more interesting setups I have seen in the small cap space in a long time.
Long $VIVO. Not financial advice. DYOR.
$VIVO
English
The largest datacenter hub in North America and maybe the world is Northern VA at 4GW. $IREN has 5GW of contracted land/power by itself, if that given you an idea of how insane their portfolio is.
McFly@ilzmcfly
$IREN has so much grid-secured power its insane. $HUT $WULF $GLXY $CIFR $KEEL dont even come close Imagine $2.5M-$3M per MW on 5000MW, because its already grid secured they can start colocation even for sites sitting out years away today. $IREN could collect between $12.5B-$15B in Revenue per year. Imagine that on 10-15 Year leases. What’s crazy is that they have even more undisclosed power and sites. Electrons are expensive and not easy to come by. They can pull a $SNDK and contract all the capacity in 2 years. The reality is that $IREN emerging themselves in the AI Cloud world and I believe they can successfully do it. But the issues is that they are slowed down by GPU delays and bottlenecks. It’s also allot more capital intensive. For the company I think AI Cloud is great but for my stock I think colocation is great. At these prices though I think most other colocation players are running out of secured power and Iren still doing AI cloud might give the better return from these levels.
English
As with any hot IPO like this, Ill look to see if Im interested in $CBRS in a month after it drops back to the IPO price.
English