D2C3 🚴 (@D2C3FiveA) header
D2C3 🚴 (@D2C3FiveA) profile picture

👨‍💻🔧💻

Joined October 2022
Photo by D2C3 🚴 (@D2C3FiveA)Photo by D2C3 🚴 (@D2C3FiveA)Photo by D2C3 🚴 (@D2C3FiveA)Photo by D2C3 🚴 (@D2C3FiveA)Photo by D2C3 🚴 (@D2C3FiveA)Photo by D2C3 🚴 (@D2C3FiveA)Photo by D2C3 🚴 (@D2C3FiveA)Photo by D2C3 🚴 (@D2C3FiveA)Photo by D2C3 🚴 (@D2C3FiveA)Photo by D2C3 🚴 (@D2C3FiveA)
$100K BTC baby! 🤑 Soaking in the vibes at Apurva Kempinski Bali—nothing beats the view 🔥
2
9
595
D2C3 🚴 retweeted
$IREN: Addressing the Semi-Analysis FUD on Reliability What's the Issue? I heard of these issue in August and the only customers that $IREN has had long enough to report "multi-day power outages, network upgrades, storage failures, air quality controls leading to link flaps, XIDs everywhere" are FireworksAI and TogetherAI at Prince George. Mackenzie is undergoing GPU commissioning as we speak (1) so I don't know why @SemiAnalysis_ mentions Mackenzie having customer problems already when there are no customers there yet. What @dylan522p says is true but convinently omitting critical industry context makes this a cleverly disguised Semi-Hit-Piece. I had heard that TogetherAI filed this complaint to IREN and Dylan was able to obtain the information from TogetherAI's CEO. I was also told that Dylan is personal friends with TogetherAI's CEO and are from the same region in India and speak the same Indian dialect which is how Dylan is able to get the info. I will show that these are industry wide problems and the affected customers renewed despite these issues. First Industry Fact First, the most important fact that @SemiAnalysis_ conveniently leaves out is that on August 27, 2026, it was confirmed that “Together AI and Fireworks AI have both renewed and expanded” (2). Despite all these problems, why did TogetherAI and FireworksAI renew? Let's look at the rest of the industry facts. Second Industry Fact Second, I will address the most alarming sounding issue: multi-day power outages. Down-time happens to everyone in the industry. @SemiAnalysis_ Platinum Cloud, $NBIS, had a outrage spanning August 19-August 20, 2026 (3). It's not narrated by SemiAnalysis in hit piece voice but the results are similar: "a storm-related event affected the building of the data center facility hosting the us-central1 region, taking the facility’s building management system offline and shutting down the chilled-water cooling loop. With active cooling lost, the data halls overheated within roughly two hours, and servers, network switches and rack power systems began shutting down on thermal protection. This caused a region-wide disruption of us-central1 services: GPU and CPU compute, external network connectivity for virtual machines, object and block storage, Managed Kubernetes, Token Factory, managed platform services, and the regional API and console endpoints" (3). "Storm-related event" makes NBIS outrage explained. SemiAnalysis's biggest bias is that they conveniently leave out the reason for IREN's outraged: BC Hydro's grid had a brown out. That's why TogetherAI renewed, not because they were okay with unexplained, random multi-day power outrages. At the time, IREN did not have enough backup generators in BC since generators produce significantly more pollution than a hydro powered grid and are hard to obtain licenses for. By March, IREN was able to obtain license to put generators for all of PG and multi-day power outrages will no longer be a problem. IREN has designed Childress and SW1 with full N+1 backup power since MSFT which requires a higher uptime which is why SemiAnalysis says those sites "look much better". Third Industry Fact Third, let's address the rest accusations: "network upgrades, storage failures, air quality controls leading to link flaps, XIDs everywhere". Using fancy technical terms without explaining them is a hallmark of a hit piece, so I'll give definitions. Link flaps means a network connection is going in and out. XID is an error code logged and reported by Nvidia GPUs. Link flaps are common: SemiAnalysis Platinum Cloud $CRWV (4), Silver Cloud Lambda (5), $MSFT (6) all had 2hr15m (4), 5hr14m (5), 4h57m (6) network out rages due to link flaps this year. Why? Because when optical networking links and transceivers collect dust particles the connection can be interrupted. SemiAnalysis conveniently uses "air quality controls" to make it sound even more scary. "XIDs everywhere" sounds scary too but they have a much larger frequency on GPUs than other compute types. $META had a paper showing that they had 419 XIDs in 54 days on their infrastructure. No other AI Cloud is this transparent about reporting problems but $META blog is one of the most honest tech blogs out there bcause META culture is "Move Fast and Break Things". Network upgrades or other control/management plane upgrades do cause problem from time to time. AWS had a DynamoDB DNS defect (8) and GCP had a policy update trigger crashes globally (9). Storage outrage is the most outrageous accusation as those are the most common across both GPU and CPU Clouds. In a completely separate incident from the storm-incident, NBIS had a storage outrage in 2025: "we had an outage of the S3 service in the eu-north1 region triggered by a steady increase of migration traffic accompanied by unexpected behavior of the program" (10). Storage downtime are not typically reported these days but in the past generation, they use to more commonly reported by GCP (11), Azure (12), AWS (13). Storage is check pointed and there is often none to very little data loss. Managed Services Don't know why SemiAnalysis has such strong shade on "Managed Services". It's a twist of the facts. IREN has no managed services but Mirantis does. Mirantis provides managed services with 99.99% uptime SLA (15) to customers like Adobe, Cornell University, PayPal, MetLife Insurance, Royal Bank of Canada (14). With the Mirantis acquisition closed, I don't know why SemiAnalysis says that IREN advertisement of managed services is misleading, NBIS is advertising the results of their acquisitions on full blast. Conclusion @SemiAnalysis_ gives a master piece lesson on how to write FUD pieces without straight up lying but cleverly crafting context. I unravel their knots for you.
IREN continues to come up in our conversations with neocloud customers, and usually for the wrong reasons. In Prince George and Mackenzie in Northern BC, Canada, we’ve heard the stories: multi-day power outages, network upgrades, storage failures, air quality controls leading to link flaps, XIDs everywhere. The #1 worst site in the industry, according to users. And we are some of their users, as we’ve gotten to test IREN GPUs from multiple providers that resell their capacity. A lot of our users we are hearing from who aren’t having a good time aren’t renting from IREN resellers, but from IREN themselves. (1/3)🧵
116
137
34
1,002
207,297
$IREN: SemiAnalysis crossed a line… If you recall my post from yesterday, I wasn’t going to keep dragging this topic out. $IREN clearly intends to handle the situation pragmatically and collaboratively instead of turning it into a public fight with @SemiAnalysis_, and I didn’t want to stand in the way of that by pouring gasoline on the fire. However, after hearing what none other than Kent Draper, $IREN's Chief Commercial Officer, had to say yesterday, I don’t think I can stay quiet on it anymore... There are several parts of Semi’s portrayal of $IREN that deserve additional context, because once you hear IREN's side of the story, some of the claims in Semi’s ClusterMAX report start to look very different. Kent went on @McnallieM yesterday for a 1 hour interview and, toward the very end, directly addressed the new ClusterMAX ranking and the accusations surrounding $IREN's Prince George site. I’ve attached his full response below because I think everyone should hear it for themselves. But read this post first, then watch the clip... It hits much harder with the full context and my added commentary. The first issue starts with the ranking methodology itself. ClusterMAX explicitly says it is ranking managed Slurm and Kubernetes clusters. Yet as Kent confirmed, $IREN does not even have a live managed-services environment today. Investors following the company already knew this. Managed services are a capability IREN is only now building out following the @MirantisIT acquisition, which brought in the orchestration layer and enterprise cloud expertise needed to move beyond predominantly bare-metal compute. That immediately raises a pretty obvious question: If ClusterMAX is specifically a ranking of managed clusters, and $IREN does not currently have a managed cluster product available for Semi to test, why is IREN being ranked as "Underperforming" in the first place? The fair classification was always "Unavailable". That may technically sit below Underperforming on the ranking, but the implication is completely different. Unavailable simply tells the reader that there was no qualifying product available to test. Underperforming tells the reader that Semi tested $IREN's managed offering and found it poor. According to Kent, they couldn’t have done that because the product isn’t live yet. This becomes even stranger when Semi goes as far as recommending that $IREN “stop pretending to offer managed clusters and inference endpoints” because of the supposed shortcomings of those services. How exactly do you reach a conclusion on the quality of a managed service that has not launched? If anything, the limited evidence we currently have around $IREN's forthcoming managed-services capabilities points in the opposite direction. Kent specifically said IREN has already signed managed-services clusters with NVIDIA for its own internal R&D workloads. That sits alongside NVIDIA’s broader five-year, $3.4B cloud agreement with $IREN across 60MW of capacity. NVIDIA obviously understands GPU infrastructure better than virtually anyone on the planet, and as Kent pointed out, they conducted extensive diligence around $IREN's ability to deliver before signing a contract of that size for their own workloads. So before $IREN's managed-services platform has even properly launched, one of the first companies willing to underwrite it is NVIDIA itself. Semi somehow found room to tell IREN to stop marketing a service that isn’t live yet, but apparently didn’t find that worth discussing... Then we get to Prince George, which is where most of the really aggressive language in Semi’s write-up comes from. Kent did not deny that Prince George has experienced issues, nor did he try to pretend the site was built from day one like IREN's newer AI infrastructure. He explained that Prince George was originally a Bitcoin mining facility that $IREN has been retrofitting for AI compute, and the company was installing additional power redundancy as part of that process. The important part is WHY the GPUs came online before all of that redundancy was finished. According to Kent, customers wanted access to the compute as quickly as possible. $IREN therefore brought the GPUs online while the redundancy buildout was still progressing, and those customers were fully aware of the setup from the beginning. That is a very different picture from IREN secretly cutting corners, handing customers an inferior product and then getting caught after the fact. It was a known tradeoff... Customers wanted the GPUs immediately, $IREN had the capacity available, and they were willing to accept reduced redundancy during the retrofit in exchange for getting access to scarce compute sooner. And today, the very power redundancy Semi criticized as “missing” is already being commissioned. There is also a pretty funny detail here that I think deserves more attention. The only customer relationship Semi actually names in connection with the Prince George problems is TogetherAI (@togethercompute). Semi tested Together capacity running out of Prince George and says it knows of multiple unhappy Together customers using that infrastructure. Semi also claims to have heard complaints from customers renting directly from $IREN, but none of those customers are identified. Meanwhile, Kent confirmed in the exact same interview that Together recently renewed or expanded its relationship with IREN. In other words, the one company sitting between Semi and the Prince George cluster they use as their clearest example of $IREN's supposed unreliability decided to do more business with IREN. That doesn’t mean an outage never happened or that every downstream Together customer had a flawless experience. It does, however, make the broader portrayal of Prince George as some completely dysfunctional operation rather difficult to square with the commercial behavior of the customer Semi itself links most directly to the site. If the experience was truly as irredeemable as the language in ClusterMAX suggests, why is Together signing up for more? The fiber accusation gets even more interesting... Semi specifically mocked Prince George for supposedly having a single Internet Service Provider (ISP) with no redundancy. Yet, Kent directly contradicted that. $IREN does have redundant fiber at Prince George. What happened over the summer was that wildfires in British Columbia damaged parts of the surrounding fiber network, temporarily reducing the level of redundancy available to the site. Those are two very different things... A site being architected without redundant fiber is an infrastructure decision, while a site having redundant fiber and then temporarily losing portions of that redundancy because wildfires physically damaged external lines is an operational event. Yet if you only read Semi’s version, you would walk away believing $IREN simply built the site with one ISP and called it a day. Prince George also tells you very little about how IREN’s next generation of infrastructure is being built. Kent confirmed that Horizon, the future Childress deployments and Sweetwater will incorporate full power, mechanical and cooling redundancy from day one, including full concurrent maintainability. Semi itself even acknowledges that Childress and Sweetwater look considerably better than Prince George. Which makes the attempt to use a retrofitted Bitcoin mining facility in Northern Canada as some sweeping indictment of where $IREN's cloud platform is heading even more questionable. Reporting genuine outages, reliability problems or customer complaints is completely fair game. I want publications like Semi digging into that stuff. But there is a difference between reporting those problems and taking issues at one transitional retrofit site, assigning an Underperforming rating to a managed service the company doesn’t even offer yet, telling $IREN to stop pretending it offers that service, calling Prince George the “#1 worst site in the industry according to some anonymous users”, and then suggesting NVIDIA’s involvement might help IREN “cut less corners this time around.” That is where, in my opinion, the report crossed from analysis into grandstanding... Especially when a considerable amount of the missing context could have been cleared up with a direct conversation with $IREN. And that brings me back to how IREN itself handled this. I have never shied away from criticizing $IREN's investor relations and communication when I thought they deserved it. I dedicated an entire section of our latest earnings report to the issue and have raised it publicly numerous times over the past year. But I’ve also been equally vocal in recent weeks about the improvement I’ve seen. There has been a meaningful uptick in investor communication, more context coming directly from management, and a much greater effort to shape the company’s own narrative before somebody else shapes it for them. This situation is another example where I genuinely could not have asked for a better response. $IREN didn’t put out some dramatic press release attacking Semi. They also didn’t start throwing accusations around on social media, and they didn’t turn a research report into a corporate feud that would only give the entire story more oxygen. Instead, their Chief Commercial Officer went onto an investor podcast, answered the question directly, explained the facts behind the situation and left it there. No grandstanding, no personal attacks, just context. At the same time, as I mentioned yesterday, IREN appears intent on working more closely with publications like Semi as its managed-services platform actually rolls out, which is exactly the right approach from a corporate perspective. There is zero reason to burn bridges with an influential industry publication when you can instead give them access to the real product once it exists and let the results speak for themselves. As a shareholder, I couldn’t be happier with how $IREN handled this. And massive credit to @McnallieM for hosting another excellent interview and asking Kent the question directly. This entire discussion would have remained one-sided without it. Now everyone has both sides of the story. Watch Kent’s full response below. Cheers! ✌️
86
93
21
626
69,147
$IREN At the request of a few homies on X, I’ve updated my IREN revenue roadmap using the latest company disclosures and my own deployment assumptions. 2026: $770M in AI revenue, exiting the year at a conservative $4B in AI ARR. 2027: $6.3B in AI revenue, exiting the year at $16.55B in AI ARR. The key distinction: revenue is earned throughout the year as capacity comes online. Exit ARR is the annualized run rate at year end. Here are the assumptions behind my updated IREN revenue roadmap: 2026 • Horizon 2–4 are delivered at the end of October, November and December, respectively. • Mackenzie’s 80MW is delivered in four phases beginning in late October. • Childress’s 50MW is delivered in two phases beginning in late November. 2027 - liquid-cooled • Horizon 5–6 and Sweetwater 1 use VR200, come online in CY Q4 and generate an assumed $27.5M in annual revenue per IT MW. • Canal Flats uses GB300, comes online in CY Q4 and generates an assumed $25M per IT MW. 2027 - air-cooled • NVIDIA’s deployment arrives in three tranches beginning in February. • Childress’s 190MW arrives in four tranches beginning in June. I’ve placed many deliveries toward the end of their expected windows. So far, IREN has not demonstrated a consistent ability to deliver tranches on schedule, much less ahead of schedule. I have not included Mirantis revenue. I’ll add it once the company discloses a figure or we have enough data to make a grounded estimate. Just 80 days ago, IREN closed a quarter with $71M in AI revenue. In 106 days, we enter a quarter in which I expect IREN to generate more than $1B in revenue. Not ARR. Actual quarterly revenue. The ramp hasn't even started!
43
72
17
477
81,795
D2C3 🚴 retweeted
The AI riptide is already underway
a surprise I think is coming the quick payback periods and extraordinarily high IRR for AI infrastructure, even at monumental scale, will push up cost of capital sufficient to push many traditional businesses into the abyss, even those without obvious direct AI counter-exposure
1,077
1,100
162
11,037
8,453,779
D2C3 🚴 retweeted
Horizon 1: delivered. IREN has delivered Horizon 1 to Microsoft and achieved NVIDIA Exemplar Cloud status on NVIDIA GB300 NVL72. Read more: iren.gcs-web.com/static-file…
371
889
308
5,552
1,460,481
$IREN For anyone paying attention to any of the accounts actually providing real insights and analysis of the IREN story, this ARR raise was obvious. To state the obvious, GPU unit economics are improving - yes, a portion of it is coming from the increased capex costs of GPUs, but a lot of it is from improved hourly rates. Coupled with the hourly rates improving, getting 45% prepayment across the associated GPU Capex is absolutely massive. Getting both the hourly rate improvements and 45% prepayments is validation of the "no need to rush" approach IREN's been taking. No need to lock in rates early for the sake of adding to a Backlog number that will be many months or years from generating revenue. Beyond the obvious there are a few other things we can glean from these disclosed numbers. Incremental ARR: >980M ARR (>4B * 85% - MSFT ARR - PG ARR) 2026 Contracted ARR is now at 3.4B Total Contracted ARR is now at 4.1B Implied hourly rate: ~3.60/GPU-hr (1.58B / 50,000/ 8760) Contract length: ~3 years (2.8B Total Contract Value / 980M ARR = ~3) In all likelihood these unit economics are for the majority if not all of Mackenzie GPUs, and none of the Childress 50MW-Air GPUs. It makes Mackenzie per MW unit economics: Implied $M/Gross-MW: $12.25/Gross MW (980M/80MW) Implied $M/IT-MW: $14.1/IT-MW (980/(80/1.15)) There is a possibility as the Childress 50MW-Air GPUs get contracted that this 4.0B ARR floor is raised again sometime this year. It's not a coincidence the company said "Contracted pricing continues to strengthen". If the remaining GPUs from these 50K are as in demand as IREN says they are and pricing continues to strengthen, I could see the remaining block clearing above these figures and raising the 4B floor. I won't assume it yet, but 4.0B is probably not the real ceiling for 2026 end of year ARR. 🚀🚀🚀🚀🚀🚀🚀🚀 More later.
IREN has signed $2.8bn in new multi-year AI Cloud services contracts with leading AI developers and raised its year-end 2026 AI Cloud ARR target from $3.7bn to over $4.0bn. “Our vertically integrated AI Cloud platform is scaling at pace. In the past 12 months we have expanded from approximately 3MW of self-built AI Cloud capacity to 480MW being delivered this year, with 1.2GW targeted for 2027, broadening our customer base across hyperscalers, enterprises and AI developers.” “We are proud to support leading companies building frontier applications across design, physical AI and robotics, generative media, AI search and model development.” - @danroberts0101 Press release: iren.gcs-web.com/static-file…
24
40
10
403
80,401
Hello London! ✈️🇬🇧 Juicy steak + ice cold Fuller's London Pride = perfect start !
2
278
D2C3 🚴 retweeted
The NYSE welcomes @twentyone to celebrate its listing! $XXI | @jackmallers
119
304
46
2,557
624,160
😂🐕
It’s unhealthy how many times I’ve watched this 😂
233
D2C3 🚴 retweeted
Thank you to @DavidSacks, @chamath, @friedberg, @Jason, and the entire @theallinpod @allinsummit team for an incredible event! It was a spectacular three days connecting with business and technology leaders, learning about the exciting AI innovations being built, and how $IREN Cloud™ can help power the next big thing!
23
101
17
813
154,355
$IREN's Next Step into Multi-Cloud (Many of my post long but all are hand written. This one is my coup de grâce.) Everyone is speculating on whether IREN’s next customer is Google, Meta or some other hyper-scale company. This is not the way to look at it. The first step to understanding what will happen is to understand Multi-Cloud. To explain Multi-Cloud, I’ll use an example none other than the Nebius-Shopify relationship. Shopify specifies exactly what their relationship is with NEbius: “we [Shopify] are partnering with GCP, as our main infra provider, and also engaging with neo-cloud providers such as Nebius to utilize large training clusters” (1). In other words, GCP is Shopify’s main cloud but Shopify is using Nebius for GPU intensive functions due to Nebius' lower GPU cost. Before any Nebius analyst jump up and say it’s not for lower cost, further down the Shopify blog post, Shopify says “Our tooling abstracts away from the underlying cloud provider, allowing us to move between vendors to get access to the latest GPUs as they're made available” (1). Shopify is a Tier 1 Tech Company as an E-commerce company but not a Tier 1 AI Research Lab, when Shopify says large training clusters, they mean training clusters are larger than inference but not the massive scale training cluster xAI, OpenAI, etc are building for their next big run. Now why doesn’t Shopify run everything on Nebius? Shopify needs GCP for user analytics, fraud detection, Spanner DB, customer data protection, Identity Services, Dataproc, etc. Let’s put it this way, hyperscalers are hyperscalers for a reason: the breath and depth of their software platform/services. Now why doesn’t Shopify use GCP for everything? GCP is “out” of GPU compute even though Google has already spent $75B in capex, or as Microsoft CEO Satya Nadella puts it, “I’m good for my $80B” (2). Note that GCP is not really "out" of GPU compute, but it’s got so much demand that GCP GPU time commands ridiculous prices. In other words, nothing is ever “out”, it just depends on the price. Hyperscalers are conscious in their capex in order to maintain their asset-light model and focus on high margins from IP/Software. Why doesn’t Google just build more datacenter faster and take that free business away Nebius? First, Google is max allocated the capex that it’s investors are comfortable with. Second, Google already maxed out its DC expansion capacity and operations teams. While GPU datacenters are not nearly as complex as TSMC fabs, just like TSMC fabs are not copy and paste, GPU datacenter that have 99.8% uptime are not copy and paste. When you guarantee 99.8% uptime, you are guaranteeing the long tail of hardware bugs. You get hardware bugs that show up on one board and not anther. Some bugs only show up above certain hardware temperatures. Some networking bugs only show up above certain bandwidth and/or certain packet types. Uneven levels of hardware wear across heterogenous and homogenous chips. Inverse logic for poorly documented driver code from the antique ages. Tracing voltage fluctuations all the way down the schematic. God Forbid EMI. Google has already assigned all its teams to build and operate its own datacenters and part of the bottleneck is the training of new teams. Colocation gives Google one of the most important component of power but doesn’t solve all the bottlenecks. The building and more importantly operating of GPU datacenters for IaaS once you have power is not copy and paste! @brianfry01 knows! Thirdly and most importantly, Google doesn’t want to get stuck with all the current gen GPUs - multi-cloud distributes the risk! So what’s the value of IaaS? Why is Microsoft signing a deal with Nebius? By extending Azure compute on top of Nebius IaaS, Microsoft gets X PaaS revenue and pays Nebius Y IaaS revenue where X > Y. X-Y is a cashflow machine for Microsoft without the capex burden, GPU obsolecence risk, and Microsoft like Google cannot build GPU datacenters fast enough anyways. The more important question is how can IREN participate in Multi-Cloud? $NBIS's Inference as a Service stack (3) is: DataOne (Power + DC Construction + DC Physical Operations) with Nebius (DC Design + IaaS + PaaS) = Full Inference Stack’s Value Chain. IREN’s official partner is TogetherAI (4) and TogetherAI’s Inference as a Service stack (5) will be: IREN (Power + DC Construction + DC Physical Operations + DC Design + IaaS) and TogetherAI (PaaS). Now why does inference need PaaS? Didn’t this Jim bro tell me bare metal GPU was the best? When you write inference code for an application, you need to optimize the GPU kernels, CUDA capture graph, and batching for that application. If you are T1 AI Lab, you tune this yourself, but if you are most other Software company working on a Software Application, you just want to focus on the Framework, API calls, DB Schema, fine-tuning your training model, but you don’t care how your inference is optimized as long as it’s optimize. The PaaS portion of “Inference as a Service” is a software service that fine-tunes the application’s inference and runs it on top of bare-metal GPUs. T1 AI Lab will tune it themselves and run on bare metal GPUs. In other words, PaaS is literally a service, not what inference runs on top of. TogetherAI+IREN = Nebius+DataOne and can serve as a AI Cloud in the Multi-Cloud ecosystem. Now here’s the banger. It doesn’t matter if IREN works with TogetherAI or AWS or Microsoft, to IREN they all serve the same purpose: the PaaS layer. Now if AWS/Microsoft is uncertain about IREN’s capabilities, IREN will prove it in its partnership with TogetherAI. In fact TogetherAI needs IREN more than AWS/Microsoft because TogetherAI doesn’t have its power, DC operations, etc, so TogetherAI will give better margins to IREN than AWS/Microsoft. But once AWS/Microsoft sees that IREN can be trusted as Y in its X-Y cashflow machine, it will want IREN IaaS to fight for PaaS market share since PaaS is made by deploying code (aka copy and paste). But just as Nvidia keeps Neoclouds as leverage, it may make sense for IREN to partner with TogetherAI for better margins and better leverage in the relationship. IREN+TogetherAI will be in position to serve as Inference/Small Training Cluster Cloud in Multi-Cloud for companies ranging from startups up to large companies like Shopify. As IREN builds credibility, IREN’s IaaS may be able to partner other strong PaaS/SaaS companies like Databricks. I’ll note that $CRWV PaaS > TogetherAI PaaS which might be the factor to push IREN to consider working with hyperscalers or Databricks, etc. Furthermore, IREN can work with model development focused startups like HumeAI; although HumeAI doesn’t have a PaaS, HumeAI can optimize the GPU kernels, CUDA capture graph, and batching for its own application. This is why Tim Delcourt is in SF: to court the TogetherAI and HumeAI and work with the next generation AI companies.
39
64
13
505
93,363
D2C3 🚴 retweeted
$IREN is pleased to announce an expansion of its AI Cloud business. - Purchase of 1.2k air-cooled NVIDIA B300s & 1.2k liquid-cooled NVIDIA GB300s - Expands AI Cloud to approximately 10.9k NVIDIA GPUs - New liquid-cooling installation at Prince George underway - Additional $96m in GPU financing secured - NVIDIA Preferred Partner status Press Release: irisenergy.gcs-web.com/stati…
42
140
36
1,050
156,827
D2C3 🚴 retweeted
$IREN is pleased to provide our monthly update for July 2025: - Record monthly revenue ($86m) and hardware profit ($66m) - AI Cloud expanded with 2.4k NVIDIA B200/B300 GPUs - Initial B200 shipment completed and fully contracted - 50MW Horizon 1 liquid-cooled AI data center on track for Q4’25 - Mackenzie fiber upgrade complete, further enhancing our ability to support continued growth of AI Cloud beyond Prince George - 1,400MW Sweetwater 1 civil and electrical works continuing Press Release: iren.gcs-web.com/static-file…
59
155
57
1,028
220,369
people asked me what i think about @mavensbot inscribing his own wizards collection i think it’s great, i love mavensbot and the wizard meme, they helped introduce me to bitcoin 12 years ago that’s why i’ve been harassing him since 2022 to make a wizard collection, then built Taproot Wizards with his blessing, then helped onboard him to ordinals to complete the wizard circle there’s no rivalry, try harder losers, wizard domination soon 🧙‍♂️
72
39
7
361
23,419
Good morning paradise Epicview 🏌️‍♂️Golf & BTC vibes! Keep swinging and HODLing
1
3
302
some surprising thoughts on the op_return drama: 🧙‍♂️ on principle, i agree with the technical claims made by core devs who want to remove the op_return relay limits (there’s a BUT coming soon). if/when demand shows up for so-called “spam” onchain, it won’t be possible to stop it with relay policies, and you might as well not try to 🧙‍♂️ BUT. there is currently little to no demand for “spam”. the mempool is empty and the average fees are virtually zero. when there’s no demand for “spam”, the relay policy actually DOES help prevent demand from manifesting! 🧙‍♂️ limiting op_return relay makes it a little more difficult for people to “spam”. they need to contact miners, or use inconvenient schemes like “inscriptions”… in most cases they just give up. i know this because they tell me: “if only we could easily do X we would”, but they can’t so they don’t. 🧙‍♂️ so the “filterooors” are right about one thing: if you don’t make it easy to “spam”, it is less likely that people will want to “spam”. they’ll just do something else instead. 🧙‍♂️ note this only works when demand isn’t there. if demand for “spam” is high, no filter can stop it. when the “filterooors” argued 2 years ago that strict filtering would kill ordinals, they were wrong. people were paying hundreds of millions of dollars to inscribe, miners would gladly help them bypass any new filters and they’d have no effect. 🧙‍♂️ but the “filterooors” so thoroughly convinced everyone that they’re idiots, that devs can’t see that they’re not wrong this time. enabling a new “spam” vector that isn’t currently in use is virtually guaranteed to increase “spam”. 🧙‍♂️ as much as i do respect bitcoin core contributors and their work, i do think they made a serious unforced error here. 🧙‍♂️ the most dumbfounding thing is that no one asked for this. there’s a small, somewhat obscure research group called “citrea” that authored a whitepaper that briefly mentioned how they’d consider to use op_return if they could. they’re not currently using it and they’ll be just fine if they can’t use it (not that it matters). 🧙‍♂️ based on that hypothetical sidebar in an obscure whitepaper, some core contributors decided to open pandora’s box with a debate that no one wanted. everything would’ve been perfectly fine in the world if this wasn’t brought up in the first place 🧙‍♂️ i think core devs mean well, and do a lot of good, and i think they don’t deserve the harassment. but it they made a huge unforced error 🧙‍♂️ at the very least, they could’ve waited to see if this ever becomes a real problem, and deal with it then, otherwise avoiding the debate altogether 🧙‍♂️ i’ve been saying for over a year that bitcoin core doesn’t have the power and influence people think they have, and that eventually this will become obvious due to one blunder or another, at which point the entire ecosystem will smell blood and a power struggle will kick off. i believe this is now in full force. 🧙‍♂️ there are many implications for this: for one thing, i think projects championed by Core like “Great Consensus Cleanup” now have no chance of happening anytime soon. Core has shown its hand, they have almost no political capital and cannot pull off small changes without massive backlash, let alone big one ones 🧙‍♂️ also, i think many groups will form soon attempting to offer alternative clients to Bitcoin Core. this is a good thing, even if many of them will be of poor quality. it’s possible that some mini-disasters will happen along the way, but eventually we will have a balanced competitive ecosystem. 🧙‍♂️ more people will participate in technical discussions. the motivation will be political but that’s better than nothing 🧙‍♂️ competition will drive Core to improve too. there are some great people there so this is a huge win. much can be improved there in building social and political capital, communicating their work, and forming some productive hierarchy 🧙‍♂️ also, thank you all for hosting my JPEGs on your nodes forever
95
67
37
510
136,089
Channel Islands Sunset
1
1
111
JUST IN: 🇺🇸 US Treasury Secretary Scott Bessent said "Bitcoin is becoming a store of value".
204
1,185
172
7,428
573,083
🇺🇸 all-in takes dc part 2! @chamath and @friedberg sit down with @CommerceGov secretary @howardlutnick for another long-form interview 🔥 -- insights from running trump's transition team -- creating @DOGE with @elonmusk -- fixing gdp -- tariff strategy -- how to sell 1m+ trump cards -- government software strategies ++ much, much more (0:00) chamath and friedberg welcome commerce secretary howard lutnick! (1:10) howard describes his 30+ year relationship with president trump and his road from business to politics (14:44) running trump's transition team, doge origin story, what it's like working for trump (38:01) balancing the budget and fixing gdp (52:21) tariff history and strategy, global trade (1:10:34) trump cards, building better government software, ai thoughts (1:22:49) sovereign wealth fund strategy (1:37:16) how his family reacted to his new role
773
1,187
662
5,748
3,895,506