IREN Vs. Nebius: Why The Latter Is A Better AI Play

Summary In this article, I'm comparing Nebius Group (NBIS) to IREN Limited (IREN) in order to find a better AI infrastructure play for growth investors. IREN's transition from Bitcoin mining to AI infrastructure has fueled growth, but reliance on external financing erodes its previous profitability ￰0￱ isn't just a cloud ￰1￱ run great software and have a massive ￰2￱ can't compete on pricing with NBIS, although I like their alternative energy ￰3￱ trades at a similar 2-year forward price-to-sales multiple as IREN, yet boasts stronger growth potential and a broader tech ecosystem. I decided to upgrade NBIS to "Buy", expecting upcoming earnings and ARR growth to catalyze further upside for ￰4￱ is a "Hold" Introduction I haven't covered IREN Limited ( IREN ) before, but I had been calling for buying Nebius Group ( NBIS ) stock when it dipped close to (and at some point even below) the $20/share mark earlier this ￰5￱ I turned neutral on NBIS, not selling my stake completely, but trimming it back in June 2025 (I know, it was a bit premature, but I had to sell partially because of my extreme overallocation to a single stock back in the day).

Anyway, with the AI neocloud technologies race gaining traction in recent months, I'm seeing some comparisons made on social media platforms, and particularly on Seeking ￰6￱ of my readers even reached out to me and asked to take a look at IREN and say whether it's worth changing their long positions in NBIS to this ex-Bitcoin miner. I've already had an article comparing NBIS to its larger peer, CoreWeave ( CRWV ), where I argued that I'd pick the former stock to get the best neocloud positioning in the longer ￰7￱ I think that the same conclusion should be applied regarding the direct comparison of Nebius and IREN because Nebius Group's long-term prospects look much brighter thanks to its seemingly better-developed tech within its core infrastructure business, as well as the additional tailwinds its non-core segments can add to the consolidated ￰8￱ Do I Think So?

When both stocks started to correct in recent weeks after their massive YTD run-ups, it grabbed investors' ￰9￱ started to think which stock to buy on the 15-20% ￰10￱ by YCharts I understand the growth profile of IREN and why it soared by over 500% on a YTD ￰11￱ originally started out as a Bitcoin mining company in 2018, and since they were quite profitable in that activity, they managed to collect sufficient capital for repositioning towards AI infrastructure, built out as it started to look more promising and less crypto cycle-effective. Theoretically, it's indeed a major advantage over peers like NBIS or CRWV, who had to scale through equity capital raises and thus ￰12￱ now, IREN is still running both BTC mining and AI Cloud/HPC services, and it's making them one of the only bottom-line-profitable AI infrastructure plays out ￰13￱ their bitcoin mining production is slipping rapidly as they're pausing further expansion to focus on AI growth and prioritize the other segment.

IREN's latest IR presentation It's clear to me that with the fall in bitcoin production, IREN has no other choice but to start financing its further expansion with external sources. We're already seeing this happening: Earlier this month, the firm priced $875 million convertible debt offering due 2031, and since the proceeds will likely go to Nvidia's ( NVDA ) Blackwell GPU deployments , the offering was largely oversubscribed, even taking into account the terms of an initial conversion price of ~$85.63 per ordinary share. It's not an immediate dilution thanks to that massive premium to the current price, and there's no interest they're obligated to pay, which is ￰14￱ I think it's a sign that IREN is losing its advantage over its peers in terms of its ability to invest and stay profitable at the same time.

I see IREN's unique selling proposition evaporating. I'm not questioning IREN's other feature, which is the fact of running entirely on renewable energy, and thus offering some cost advantages through some of the lowest power costs in the ￰15￱ adds a lot of value to potential customers seeking low-latency and high-throughput GPU compute, so I think their hardware-driven moat is actually quite significant right ￰16￱ in the long run, I think it's all about software, and what else beyond GPU racks all the neoclouds will be able to brag ￰17￱ Group's infrastructure was purpose-built for AI workloads from the ground up; it's not an ex-miner who decided to reposition.

They've always been here as a cloud provider, and hence, I believe their software tech is superior. NBIS's "NeoCloud", a specialized cloud computing platform designed for high-performance GPU and AI applications, helps them optimize every layer of their AI stack, from hardware to software, and it provides higher compute ￰18￱ keep pushing the boundaries on multiple software-related fronts, signing partnerships with AI-native platform companies like WEKA , and thus adding more features to their full-stack AI cloud ￰19￱ the level of tech and support Nebius's customers are getting, they're likely to choose it over IREN, in my opinion. Moreover, IREN's pricing per hour for NVIDIA H100 is identical to that of Nebius (for the same H100), so I don't see a reason to switch from NBIS to IREN for existing customers.

IREN's pricing page Nebius Group's pricing page IREN's access to inexpensive, renewable power and its strong hardware partnerships with Nvidia are great selling points, I agree, but I don't see them monetizing that efficiency better than Nebius because the latter can use its own software-driven advantages to keep pricing reasonably affordable and ￰20￱ NBIS also has the same mega caps in partners, so IREN doesn't stand out solely on its ties with Nvidia or anyone ￰21￱ is also well-capitalized after its recent financing ￰22￱ addition to the $1.6 billion in short-term liquidity on its balance sheet , the firm can secure debt on exceptionally favorable terms using the recently signed contract with Microsoft as "near-risk-free collateral".

That's what CoreWeave used to have as an expansion vehicle in the past (and also in the present time), and it's been working for them quite well. Yes, Nebius is going to be a "CapEx absorber" for MSFT and big techs, and the management guided for about "$500 million in quarterly CAPEX", but as long as the demand side is strong amid all the AI buildouts, it's going to be justified, so I'm not scared by the current cash runway at ￰23￱ addition to its well-developed orchestration layer (inherited from its Yandex origins), which can lower deployment friction for AI models, Nebius holds stakes in Avride (~80% of the firm), TripleTen, Toloka, and ClickHouse, as well as in smaller ￰24￱ of these stakes has its own value proposition, and it lets them expose operations and gain ground simultaneously across different tech ￰25￱ effective TAM that Nebius can use for its valuation justification is bigger than that of IREN, but the latter isn't really that cheap compared to ￰26￱ me ￰27￱ Isn't Priced Cheaper Than Nebius Nebius is ~70% highly valued than IREN, although it's going to show only $580 million in sales for FY2025, while IREN's top line is already slightly above $500 million on a TTM ￰28￱ am I calling NBIS a better pick than IREN, given this fact?

It's a matter of forward growth. Yes, IREN is valued at a positive P/E ratio of 60x on a forward basis , while NBIS isn't projected to have its first positive P/E ratio any time soon: Seeking Alpha, NBIS But it's important to keep in mind here that IREN's bitcoin-related sales are not going to drive past years' stability to the bottom ￰29￱ other words, the firm is going to come through the same heavy-intensive capital stage and its earnings are going to be hammered in the medium ￰30￱ forward-looking EPS estimates and the implied P/Es don't look that much better, and they're even more cyclically dependent (based on what today's consensus shows ).

Seeking Alpha, IREN What we can use as an alternative to the P/Es, which tend to be useless for valuing growth stock anyway, are sales-related multiples, for example, ￰31￱ the operational expansion in both cases (for NBIS and IREN) is just getting started and we're about to see the first meaningful results only next year, I think it's better to assess their 2-year forward ￰32￱ taken their current market caps and having divided them by the corresponding sales estimates, I got the following result: Data by YCharts As you can see, IREN is currently trading at 7.53x FY2027 sales, while NBIS trades at only ￰33￱ difference of ~6.66% isn't that enormous, I agree, but I think the market is either overestimating IREN's growth potential, or underestimating Nebius' one, not putting sufficient premium on top of the latter's forward ￰34￱ Thoughts I recall that I wrote in August 2025 that I wouldn't buy more of NBIS stock in addition to my existing position, and the stock managed to skyrocket by over 60% since that call of mine.

Naturally, I was wrong, but my reasoning at the time lacked some essential corporate news (for example, the NBIS-Microsoft ( MSFT ) $20 billion agreement ). In August, Wall Street analysts were projecting $3.11 billion in FY2027 sales, and now the estimate for the same year stands at $4.34 billion, with likely more positive revisions to come throughout ￰35￱ the 20% off high correction that took place in recent weeks, I think it's time for my upgrade to "Buy" today. I'm waiting for next week's earnings release: NBIS is projected to show 52 cents in net loss per share, but it's not as important as its quarterly sales ($155.73 million consensus estimate), and ARR growth updates.

I'm expecting their ARR to increase at a higher-than-expected pace thanks to the recently signed deals amid overall modest coverage from the analysts, and it can catalyze the stock to the ￰36￱ reports 1 month later than Nebius (in late November), and it's going to show a massive QoQ reduction in EPS (from 66 cents to 13 cents in just 3 months) due to the reasons that I discussed today. It's not going to be an "ugly earnings print" solely based on this expectation, because, again, it's all about their future capacity built out. I'm curious to see what their corporate strategic pivot brings next, and how they manage to position themselves without developing the long-supporting Bitcoin mining ￰37￱ it's valued for right now, I see better options (like Nebius).

Based on all the above, I decided to rate IREN as a "Hold" and upgrade NBIS from "Hold" to "Buy". Thank you for reading!