Big news just in… stock responding much higher in the AH, currently up 37% (edit, now up 57%!) in AH.
Amsterdam, September 8, 2025 — Nebius Group N.V. (NASDAQ: NBIS) (“Nebius” or the “Company”), a leading AI infrastructure company, today announced an agreement to deliver AI infrastructure to Microsoft.
Under this multi-year agreement, Nebius will deliver dedicated capacity to Microsoft from its new data center in Vineland, New Jersey starting later this year.
Arkady Volozh, founder and CEO of Nebius, said:
“Nebius’s core AI cloud business, serving customers from AI startups to enterprises, is performing exceptionally well. We have also said that, in addition to our core business, we expect to secure significant long-term committed contracts with leading AI labs and big tech companies. I’m happy to announce the first of these contracts, and I believe there are more to come. The economics of the deal are attractive in their own right, but, significantly, the deal will also help us to accelerate the growth of our AI cloud business even further in 2026 and beyond.”
Nebius expects to finance the capital expenditure associated with the contract through a combination of cash flow coming from the deal and the issuance of debt secured against the contract in the near term, at terms enhanced by the credit quality of the counterparty. The Company is also evaluating a number of additional financing options to enable significantly faster growth than originally planned and will update the market on its financing strategy in due course.
Nebius has been forecasting explosive growth in revenue this year and the market did not price in that growth. Now Nebius is showing the market where the revenue ramp is coming from the market is rewarding Nebius.The doubting Thomas market has been shown.
This is huge. Nebius, I believe was guiding to around $1 billion of ARR ending in 2025, to inking $17.4 billion from Microsoft over 6 years, with a MSFT option to increase to $19.4 billion.
From the 6-K:
The GPU Services will be deployed in several tranches during 2025 and 2026. Subject to the satisfaction of the deployment and availability of the GPU Services, the total contract value is about $17.4 billion through 2031. Microsoft may also acquire additional services and/or capacity under the Agreement, which would increase the total contract value to about $19.4 billion. Cash flow coming from the Agreement will be utilized to finance part of the capital expenditure associated with the Agreement.
A few months back I wrestled with NBIS vs. RBRK and RDDT.
While all three remain wonderful companies, I am so happy I chose NBIS. Now it’s my largest holding at 19.4%. A win like this every once in a makes up for an occasional loser.
I own my losses, but also try to enjoy my victory laps to their fullest.
I’ve been a long time skeptic of this company. However, I cannot recall ever seeing a company that’s not a micro-cap sign a contract like this bigger than their market cap! I’m in for a small starter position in the pre-market.
A new question I’ve been asking to each company in my portfolio is “how fast are they scaling up”? I’m really looking to see companies which are capable to expand at a staggering pace.
I’ve held NBIS for a few quarters now and I’ve also added more on the news.
@wpr101 You also have sizable positions in IREN, ALAB, and CRDO. How large of a stake are you targeting in NBIS? Are you at all concerned about being overexposed to the AI infrastructure side?
@SailingDev I actually closed my positions on Nebius and Credo recently. My exposure to the sector comes from IREN, ALAB, HIVE, and AAOI now. The positions in IREN and ALAB are largely unchanged around 18%, but I have bumped up HIVE to about 8%, and AAOI to about 5%.
The debt raise of 3B by Nebius was a reminder of why I passed on the company initially in favor of IREN. From what I can gather Nebius has 100 MW of “active online data center capacity” and 220 MW of “connected power” as of September 2025. While IREN currently has 810 MW of connected power, with +1GW coming online in April 2026. HIVE has 430 MW of connected power and is a sub 1B market cap. My view here is that IREN and HIVE have done the harder work of electrifying the grid, while Nebius has focused heavily on sales and very heavily on marketing.
Basically the debt raise was a reminder that Nebius still has a ton of work ahead of them to electrify the grid, or pay contractors to do so. I view the Microsoft contract as a massive land for them, but they’ve still got to deliver against that timeline. I’m just not sure the former Yandex team which was known for a copy-cat Google product, has the right experience.
Additionally I have concerns about the management team from Nebius. Their founder was previously sanctioned by the EU, and where the company originates from is known for being fast and loose with accounting. On a tactical level I’m wondering if their workers will be able to get visas to Finland and the US. I thought Nebius would have compliance and regulatory issues selling into the US and EU. So far I’ve been proven completely wrong by their ability to sell and the stock price appreciation.
I’ve seen the equity stakes from Nebius in companies like Clickhouse and others be mentioned as a big positive. My take here is it’s a lot to track all these different equity stakes and allocations they have. It sounds like Nebius has made some smart investments. At the same, I’m not thrilled about investing in a conglomerate of businesses.
The management team at Credo has lost my trust. Their CEO was recently at the Goldman Sach’s conference and their presentation was in stark contrast to Astera’s recent conferences. There was zero talk of their new PRISM software platform during the entire Q&A. Additionally, the CEO was asked about pricing increases with product, and he replied “You want me to talk about increasing pricing in a public forum?”. He then said it was a joke, but gave a long answer that did not answer the question well.
The Credo CEO made it sound like they are just selling wires practically, there was hardly any vision specified. They said they won’t play in the Switch area where Astera has a big market with it’s Scorpio products. Additionally, I noticed Astera seems much less worried about Credo as a competitor. Here’s what Astera said recently in reference to their markets,
So it’s a slightly different way simply because we don’t believe in carrying the inventory and all the headache that comes with cable type of products just given that’s not our expertise. Our expertise is to develop silicon and software and all that. So we’re trying to stay true to what our strengths are and not trying to be a manufacturing house for cables.
Astera seems to give some props to Credo for doing well, but at the end of the day I see Credo as staying more in that “manufacturing house for cables” space. I believe Amazon threw Credo a lifeline as their 86% revenue customer two quarters ago to keep them in the race. Amazon has since decreased spending with Credo the past two quarters. The reasoning for this from Amazon’s side is likely they do not want Astera to have pricing power and it makes more sense to have two players that Amazon can try and pit against each other.
This is a good point. It’s true that establishing connected power may take significant efforts. But it’s also worth thinking that why Microsoft would sign long term contract with Nebius who only has 220 MW connected power rather than IREN who has 810 MW. I don’t think advertising can be the primary reason of singing such a critical deal. There must be something else that appears unique advantages of Nebius to Microsoft. It could be the high performance of both the vertically integrated solution and the bare metal options provided by Nebius. It could be the track record to serve enterprise customers (e.g. Cloudflare, Shopify). It could be the proven track record from the engineering team.
IREN has advantage and expertise in dealing with power, which is arguably a difficult part of GPU cloud. But Nebius has a full stack of engineering team that optimize the stack end to end. They design and build their own data center, their own servers and their own cooling solutions. They have built software stack to allow startups to initiate inference or training tasks with minimal efforts.
I think IREN has a good risk-reward at the moment, but I can not turn down Nebius as an investment when it has $3 billion yearly contracted revenue on average for the next 6 years. That itself is almost 6X Nebius’s 2025 revenue.
Long NBIS with full position; small position in IREN
I always really appreciate your comments and strong contributiuon to this board and I get where you are coming from, but I respectfully disagree with a number of your points.
I have been a big Nebius bull on this board since I first bought it back in January of this year. In fact, I decided very early on this year that Nebius is the real deal - and so I made it my biggest stock. It is still my biggest stock by a huge margin today. YTD I am now up 112% on the entire portfolio - much of which is due to Nebius.
First, the $3B plus debt raise isn’t a red flag — it’s a growth accelerator. Nebius isn’t raising money to hope for customers later; it already has hyperscaler demand in hand, including the Microsoft contract. That kind of contracted revenue is the dream scenario, and debt financing makes sense when you have guaranteed ROI ahead of you.
On the MW comparison: IREN and HIVE do have more connected capacity today, but utilization is what actually drives returns. Nebius is scaling demand-first. I’d much rather see 220 MW tied to real workloads than lots of MW sitting idle.
As for “marketing vs. execution,” Microsoft and others don’t hand out contracts of this size that easily. If Nebius didn’t have real execution, it wouldn’t even be in the room with hyperscalers, let alone signing deals.
It’s fair to say that the founder’s background has baggage, but Nebius is winning contracts in the US and EU, which means it has already passed regulatory and compliance hurdles. I don’t think it’s right to call Yandex a “copy of Google”. That’s not really relevant. What matters is that Nebius built world-class engineering talent under difficult conditions, and now that talent is building AI infrastructure that hyperscalers trust. Everything I have read about the Nebius team ranks it as amongst the best in the world.
The equity stakes (AVride, Clickhouse, etc.) are a big positive not a negative. You’re basically getting hidden upside here.
That’s why I’m long Nebius. It’s not a miner story, it’s a hyperscaler-aligned AI infra play with hidden upside through equity stakes. The debt raise is exactly what I want to see — it shows they’re scaling aggressively to meet real demand.
OK, I get it. I had taken a small position in NBIS based on nothing more than the big pop after earnings. But, like you, I was later reminded of why I passed on it in the first place - primarily the management, the CEO’s background made me nervous. And yeah, there was the debt load, but for whatever reasons it I wasn’t unduly concerned about it.
And CRDO, thanks for making me aware of what appears to be a lack of confidence by the CEO in their own products. I sold some outright and I also sold covered calls on the rest.
But I must admit, I’m puzzled by your investments in HIVE and AAOI. I didn’t dig in and do much of an analysis, but just looking at their revenues over the last 4 quarters doesn’t make me want to throw money at either of them. Care to elaborate? I know, it’s the middle of the month and you’ll probably have more to say in your monthly review - but maybe just brief description of what makes you interested enough to place a substantial bet on these two companies. Or I can wait.
Though I don’t know if I will be able to look at your videos for the next two months. I’m on my way to China this afternoon and the Chinese block all things Google. That includes YouTube. I installed a VPN, but my experience in the past has been that VPN performance in China is spotty at best.
There are a couple areas where I can see Nebius having a competitive advantage,
They have been talking a lot about their software stack and optimizations they have regarding their platform. If they are able to produce a superior interface or squeeze more efficiency out of the underlying hardware that could be a game changer. On the other hand IREN is saying customers are demanding bare metal and don’t need that hand holding.
Nebius (and CoreWeave) are building in more populated areas like New Jersey. I believe this offers a competitive advantage to Wall Street firms which need the low latency, and the closer the machines are to their location the better for tasks like high frequency trading. IREN is building in Texas and Canada for the renewables and low cost optimizations, but there are likely some clients who are sensitive to losing a some milliseconds of timing because of the physical location. HIVE on the other hand is in Canada, Sweden, and Paraguay quite far away from the main customer bases for HPC.
I’m in agreement this is an absolutely staggering contract amount, and why I’m on the fence with them despite those other concerns. It is wild to see a company sign a contract for more value than their market cap at their scale.
On the MW comparison: IREN and HIVE do have more connected capacity today, but utilization is what actually drives returns. Nebius is scaling demand-first. I’d much rather see 220 MW tied to real workloads than lots of MW sitting idle.
All of those MWs for IREN and HIVE are going to be used for Bitcoin or HPC one way or another. It’s a fair critique that IREN and HIVE aren’t as experienced in the sales side and do not have that marquee contract that Nebius and WULF landed. I view the electricity side of the equation more difficult than the procurement and setup of machines from Nvidia. All that debt that Nebius is taking on is going to build up data centers and buy machines, but IREN already has half of that equation solved on the electricity side.
It’s fair to say that the founder’s background has baggage, but Nebius is winning contracts in the US and EU, which means it has already passed regulatory and compliance hurdles. I don’t think it’s right to call Yandex a “copy of Google”. That’s not really relevant. What matters is that Nebius built world-class engineering talent under difficult conditions, and now that talent is building AI infrastructure that hyperscalers trust. Everything I have read about the Nebius team ranks it as amongst the best in the world.
I am questioning where this narrative comes from that Nebius has the world’s best engineers? From my viewpoint it seems like marketing that they seeded the internet with these stories about their superior engineering skills. I’ve never heard anybody say Yandex was some great innovator. Or what innovations did Yandex have that were technological breakthroughs?
When I posted about a company called Kaspi years back there was a lot of push back that this was a “Russian company” even though they are based out of Kazakhstan.
Nebius has done a remarkable job rebranding themselves to a European company, but some of these lines from their earnings call rub me the wrong way,
“A bit longer answer is that we’re very well connected in Europe. We came from Europe. We have and we’ll have even more data centers in Europe.”
Sure I can give a quick highlight on these two,
HIVE, they are also in the space of a Bitcoin miner that has some small HPC revenue. I mentioned they have 430 MW which is a lot, but their latest announcement is they expect to have 25 exahash for mining by Thanksgiving, and are currently at 18 exahash. That’s about 3% of the global mining capacity alone, which doesn’t really correspond to their small market cap right now. They already are hosting some Hoppers and A100s with their software platform Buzz, so it’s not like they are totally new to HPC. They got a really great deal with electricity and miners down in Paraguay from another miner collapsing. I don’t think the market has factored this in at all. They are dual listed on the Canadian exchange, and I see them as vastly under-appreciated by the market, probably in part because institutional money is oblivious to them.
AAOI, the last couple quarters were somewhat flat on revenue but their guide is for a big jump. This lines up exactly with what Marvell and Applied Optoelectronics are saying about their businesses doing well. A ton of items I like from their conference call, but their is a good chance I’m reading too much into their growth story too early. I definitely want to see them come in about 120M+ of revenue to have some confidence going forward. That’s right in line with their guidance range, which alone would be over +80% yoy and about +20% qoq growth.
I dug a little bit into the 810 MW connected power that IREN has. As you said, it will be used either on bitcoin mining or AI. The management stated that they would stop expanding mining after it reaches 50~52 EH/s level. Grok told me that, in order to maintain this level of mining, the required power is already 891.25 MW to 926.9 MW. So I guess IREN will need to deploy their planned new power to enable the AI workloads.
From this standpoint, does IREN actually have that much advantage in landing power, compared to Nebius?
IREN plans to have an additional 1,400 MW capacity energized by April 2026, just a few months from now. For reference, the MSFT/ NBIS deal is for just 300 MW.
HOWEVER, the Nebius deal included not just the datacenter, but also GPU servers, which make up the majority of the costs.
Just how big those costs are, I tried to find out using Perplexity, as I’m not an expert on this topic. If anyone has different numbers, please correct me - because as it stands, I don’t see how this deal is very positive for Nebius.
tl;dr: of the $17.4 B paid by Microsoft, most of it ($10-15B) will be eaten up by the GPU costs alone.
GPU Models for the Deal
Nebius uses Nvidia’s top-tier AI GPUs, mainly H100s initially and B300 GPUs as newer generation hardware becomes available in late 2025 and beyond.
These GPUs are deployed in large clusters within Nebius’s dedicated data centers, such as the 300 MW Vineland facility delivering capacity in multiple phases during 2025-2026.
Number of GPUs for 300 MW
For H100 GPUs (~700 watts each), about 428,000 GPUs could run on 300 MW assuming full power availability.
For B300 GPUs (~1,400 watts each), about 214,000 GPUs could run on the same 300 MW power budget.
Estimated Total Cost
Nvidia B300 GPU systems cost roughly $400,000+ per system (containing multiple GPUs, e.g., 8 GPUs per DGX B300 system).
To estimate cost at the GPU level, assuming roughly $50,000 per B300 GPU (approximate since exact standalone GPU pricing isn’t public), then for 214,000 GPUs:
214,000×50,000=$10.7 billion
For H100 GPUs, price per unit may be slightly less (~$30,000 to $40,000 per unit historically), so for 428,000 units:
428,000×35,000=$15 billion approx.
Considering additional infrastructure costs (servers, networking, cooling), the $17.4 billion deal value aligns well with these GPU purchase and deployment costs, combined with operational expenses for 300 MW of dedicated AI infrastructure.
To summarize, it looks like NVDA is the biggest winner of this deal.
Hi William - just wondered whether you have explored TeraWulf (WULF) in depth and what you make of their BT mining / HPC DC hybrid business model build out taking advantage of their low cost peak power supply and the arbitrage opportunity to flip between bitcoin mining and DC with their processing compute?
Regarding „one time charges“: A GPU under full utilization has a life expectancy of 3-5 years. Maybe more with liquid cooling - but even then they will be technically obsolete due to newer models coming out. GPUs are quickly depreciating assets.
@wpr101 Thanks for the detailed breakdown, really useful! I also bailed on CRDO after a good return. Too many risks compared to other plays.
On Nebius, I see it a bit differently. Google’s stack is world-class, and building a credible alternative requires serious engineering chops. From my own experience working with Russian engineers, the technical talent is top-tier! The MSFT contract is strong external validation. Satya isn’t a bean counter, and if he picked Nebius, that tells me they can execute. I got in early ($20-$30) and added after the MSFT news, so now it’s a medium-conviction (7-8%) position for me.
Really intrigued by your reallocation into HIVE and AAOI. I’m looking forward to your next video. HIVE in particular looks like another IREN-style bet: crypto miners with massive power footprints that could pivot into HPC. But doesn’t that still require heavy capex to retrofit for HPC (especially for training workloads considering the location of their expansion in Paraguay), e.g.: the same drawback you flagged for Nebius? I doubt it’s as easy and cheap as IREN makes it sound (it’s in their best interest to frame it as trivial). Especially for HIVE which has only 24 employees?
Why did MSFT choose NBIS as its data center partner? Doesn’t it stand to reason that NBIS is one of the most capable data center counterparties? However, as Wrl has pointed out, NBIS has a formidable marketing machine, and they have painted a very clear picture of the depth of NBIS’ technical stack.
Now let us consider for a moment a few things we know about IREN. The Roberts brothers who founded IREN started out as investment bankers for Macquarie Group. If you ask Perplexity: Does Macquarie Group have a particular area of expertise? The answer is: infrastructure asset management and finance. The Roberts brothers cut their teeth doing infrastructure deals. From the inception of IREN their sights were set on developing the infrastructure for high performance computing. They are not wedded to bitcoin mining, as they pointed out in the most recent quarterly earnings call when Dan Roberts clearly stated: “We aren’t religious” [about how the infrastructure is used.] Their purpose is to use their infrastructure in whatever way generates the highest return for shareholders of IREN, and that is why they are now pursuing the AI Data Center opportunity. In 2021 IREN hired Denis Skrinnikoff as Chief Technology Officer. Skrinnikoff brought extensive Tier 3 Canadian data center and cloud engineering experience from his leadership roles at RackForce Networks and TeraGo. Consider that the buildings that IREN built in British Columbia to house it’s bitcoin mining rigs are now being redeployed to house Nvidia GPUs. In its most recent presentation, there is a picture of bitcoin rigs and Nvidia Hoppers in adjacent isles. IREN’s data centers have been designed to allow up to 200kw rack density. Without going into further detail, suffice it so say that IREN intended to pivot into High performance compute from its inception. It’s not a mere bitcoin miner, it’s an outfit that specializes in developing infrastructure for high performance computing. And it’s led by a team that knows how to use leverage in making deals. Could it be that MSFT chose NBIS because it was a relatively weaker counterparty, one in which it could use its leverage to drive a better bargain? That seems plausible to me.
Now let’s consider a recent development. On August 27th 2025 IREN announced that it would lease and expand its AI Cloud infrastructure to approximately 10900 Nvidia GPUs. The announcement detailed securing NVIDIA Preferred Partner status, purchasing 2,400 next-generation NVIDIA GPUs, and financing via a 24-month lease structure covering the liquid-cooled GB300 GPUs. Leasing Nvidia GPUs, rather than owning them, is new to data center development. Why is this important? In a word, optionality. If you listened to IREN’s most recent conference call, you heard the word optionality several times. At the end of the lease term, IREN can: (1) buy the GPUs, or (2) return them. Owing assets comes with attendant risks, and in technology one of the biggest risks is obsolescence. DELL is the owner of the GPUs and as such carries the risk of obsolescence, but IREN retains optionality. The implicit interest rate in this lease is “single digits.” When factoring in data center infrastructure costs along with the GPUs, the payback period extends to three to four years. This rapid ROI is driven by high hardware margins and near-full GPU utilization rates in IREN’s AI cloud business. So, this arrangement will allow IREN to fully recoup not only it’s cost of leasing the Nvidia GPUs, but the entire cost of its data center infrastructure in just 3 to 4 years. While this is an innovation in GPU financing, it is an adaptation of the standard playbook that the Roberts brothers learned doing infrastructure deals at Macquarie. It has been assumed that a mountain of debt and dilution would be prohibitive for IREN and force it to do co-location deals for much of its GWs, which is akin to operating an REIT, which carries a much lower return on investment with a much longer payback. Now we see there is another way forward that carries much more attractive returns and is much more capital efficient. NBIS, on the other hand, seems to have chosen the well-trodden path of debt and dilution and the risk of obsolescence.
As a final point, consider that IREN (according to Perplexity) has $1,949,506 in revenue per employee. NBIS has $181,838 in revenue per employee. NBIS appears bloated to me, and that together with the $1.2B in debit it is carrying on its balance sheet doesn’t seem like a great combination. MSFT and hyperscalers in general have office buildings filled with software engineers, why should they need even more? I understand why other clients might need that kind of assistance, but certainly not a hyperscaler. Yes, I know, IREN and NBIS, while in a similar line of business, operate in a different manner. But as investors, our future returns are married to return on investment, and we should be agnostic to how that ROI is earned.
@SailingDev I probably did not articulate my point too clearly here with what I was going for on this topic.
I’ve also seen that first hand when I was working as a software developer where engineers from that region of Russia, Belarus, and Ukraine were top tier engineers. I was working at a 50 person startup that originally had about 15 US based engineers. Over time the company replaced pretty much everyone that was US based with the hired help from those regions. Effectively they would do twice the work at half the price, and when the company needed to make cuts they replaced the more expensive American engineers.
What I’m trying to ask is where does the story that Nebius has superior talent originate from? This is where things seem a bit fishy to me.
We can probably all agree Nvidia has superior engineers, but do we ever hear about them? The only thing I hear about, is how grand Jensen’s vision is and there is very little talk of their engineers. Same thing with Astera, they must have superior engineers or how else would they be producing new innovative products. Yet we rarely hear anything about their engineers, or even CEO.
The closest I can recall seeing something similar was with Supermicro. The founder kept stressing what a great research org they are and how they have the best engineers. This was coming from the founder though. Yet I do not see the CEO of Nebius saying this same thing.
This leads me to believe that Nebius is planting these stories about their superior engineering force, and it’s likely through methods that Western companies would not even consider such as paying social media users to shout from the rooftops about Nebius superior engineers. I see almost every post on Twitter that’s pro Nebius mention the quality of the engineers. But what is the source of this information?
I was searching Saul’s board to see if anybody had investments in Yandex prior to 2024 when they rebranded. There’s not a single mention of the company from anybody or in anybody’s portfolio. So why was nobody invested in Yandex if the company was a such a force in engineering? Again, I would ask can anybody point to a single innovation that Yandex produced over the course its company history? Nearly all of their products were clones of Google products, or copying something like Uber Eats.
@monkeydluffy IREN stopped accumulating hardware for Bitcoin mining when they reached 50 exahash, and of the 810 MW they have online 650 is for Bitcoin mining. That analysis from Grok is basically correct, that a normal mining firm would take 900 MW to host 50 exahash, but again it shows IREN’s efficiency here that they can do that with 650 MW. The remaining 160 MW is being used for HPC or getting ready to host HPC.
The company has said that Bitcoin mining can be swapped for HPC at minimal CapEx in about two months, as it does take some reconfiguration time. This is one area where I prefer IREN over Nebius, is this added optionality. IREN can do Bitcoin or HPC, while Nebius can only do HPC. The economics of HPC are currently better than Bitcoin which is why many companies are making this move, but that may not always be the case depending on the Bitcoin price, and the mining difficulty.
So I do see that 810 MW as a big advantage overall. Currently where things stand, IREN has 4x the electricity ready right now, and half the market cap of NBIS. As someone mentioned another 1.4 GW comes online early next year for IREN as well. I will say it maybe is some risk they bring on electricity faster than they can get Blackwells. I don’t know if keeping the electricity “idle” is a significant cost or not.
I looked at WULF and ended up passing on them, as it seems most of their market cap value is based on announcements. It seems way overpriced on MW + exahash when I do an industry comparison. Just take a look at how WULF and HIVE stack up here, and it may be more clear why I’m so bullish on Hive,
With those numbers I’d expect HIVE to have maybe twice the value of WULF. Yet the market caps are 890M for HIVE and 4.5B for WULF. Keep in mind HIVE has about 200M worth of Bitcoin, while WULF has about 40M.
The only advantage I see for WULF is they already signed a contract with Google for HPC, but should that mean the company is worth 5x a competitor with twice better the metrics on everything else?
It may take more time for HIVE to secure some press releases on contracts, but every company in the space is saying there is way more demand than supply that they can bring on. I don’t think HIVE will have trouble selling it’s compute even if they have not announced anything big yet. Lastly I’ll add many of these Canadian companies which are dual listed trade as tiny valuations compared to their American counterparts. From my perspective this is a huge opportunity as these Canadian companies barely generate any interest, but produce actual results.
I’ll add one more point that I may get back in Nebius at some point. It’s possible a company like HIVE or AAOI could disappoint and Nebius looks like the next best company. I appreciate everyone for keeping the conversation civil when presenting some bear cases.