Nebius has picked up another major Wall Street endorsement, and this one goes straight to the heart of the bull case surrounding the AI infrastructure company. Truist Securities initiated coverage of Nebius stock with a Buy rating and a $355 price target, implying roughly 48% upside from NBIS’s September 9 close near $240. The firm’s thesis is not simply that demand for GPUs remains strong. Truist argues that Nebius’s engineering depth and increasingly favorable cloud-contract economics could allow the company to generate more revenue from the capacity it is building than investors currently assume.
That distinction matters because Nebius is already one of the market’s most aggressively valued AI infrastructure plays. The company has signed massive agreements with Meta, secured a $2 billion strategic investment from Nvidia, and just became Palantir’s preferred sovereign AI infrastructure partner. Its second-quarter revenue surged 454% year over year to $582.3 million, while the core Nebius AI Cloud segment produced $285.7 million of adjusted EBITDA. Those are extraordinary growth numbers, but the stock’s valuation means investors are already assuming years of rapid expansion. The new Truist call effectively argues that the market may still be underestimating one critical variable: how much money Nebius can earn from each unit of scarce AI compute capacity.
The result is an unusually interesting setup. Nebius does not simply need to build more data centers and install more Nvidia systems. It needs to prove that those systems can command attractive prices, stay highly utilized and generate enough return to justify the tens of billions of dollars being committed to expansion. Recent pricing data suggest that economics may be moving in Nebius’s favor.
Truist’s $355 Call Is Really About Pricing Power
Truist analyst Arvind Ramnani’s new Buy rating centers in part on what the firm views as improving cloud economics. Recent Nebius capacity auctions have indicated that premium AI compute can command substantially higher pricing than longer-term contracted capacity, suggesting that supply remains tight enough for customers to pay up when they need hardware quickly. One recent analysis cited short-duration Blackwell capacity at roughly $40 million to $50 million per megawatt, compared with approximately $20 million to $25 million per megawatt for mid-term commitments.
That gap is potentially extremely important for NBIS shareholders. AI cloud companies usually trade off certainty against upside. Long-term customer contracts provide predictable utilization and help justify the enormous upfront cost of building data centers, but they can lock providers into lower prices if compute scarcity becomes more severe later. Shorter contracts or auction-based pricing can capture that scarcity premium, although they also introduce more revenue volatility.
Related articles
Nebius appears to be trying to balance both approaches. Large anchor customers such as Meta provide long-duration commitments, while other capacity can potentially be sold into a market where customers are willing to pay substantially more for immediate access. If Truist is correct that recent pricing reflects sustainable economics rather than a temporary shortage, Nebius could outperform revenue expectations without necessarily building materially more infrastructure than currently planned.
That is the source of the upside embedded in the $355 target.
A 454% Revenue Surge Shows the Business Is Already Scaling
Nebius does not need analysts to imagine future growth from scratch. The latest quarterly numbers already show a company expanding at extraordinary speed. Second-quarter 2026 consolidated revenue reached $582.3 million, compared with $105.1 million a year earlier, an increase of 454%. Revenue for the first six months of the year reached $981.3 million, up 529% from the comparable period.
The economics inside the core AI cloud operation were even more impressive. Nebius AI Cloud produced $285.7 million of adjusted EBITDA in Q2, compared with just $9.5 million a year earlier. Across the entire group, adjusted EBITDA reached $236.2 million because losses at Avride and TripleTen partially offset the cloud business’s profitability.
That margin profile helps explain why Wall Street is becoming more enthusiastic. AI infrastructure is extremely capital intensive, but high utilization can generate strong operating leverage because the expensive GPUs and data-center systems have already been purchased. Once the infrastructure is installed, every incremental unit of paid usage can contribute meaningfully to profitability.
The question has always been whether enough customers would exist to fill the capacity. Nebius’s recent agreements are increasingly answering that concern.
Meta Has Already Put Billions Behind the Nebius Story
The largest customer validation came in March when Nebius signed a five-year agreement with Meta covering $12 billion of dedicated AI infrastructure capacity across multiple locations. The infrastructure will use Nvidia Vera Rubin systems, with deliveries beginning in early 2027.
The broader economics could be even larger. Meta also committed to purchase additional available compute from certain upcoming Nebius clusters, potentially worth up to another $15 billion over five years. Nebius intends to sell that capacity to third-party customers first, with Meta effectively acting as a backstop for whatever remains. That means the total potential economic value connected to the relationship can reach roughly $27 billion, although only the $12 billion dedicated component represents the clearly specified base commitment.
That structure is especially attractive when considered alongside Truist’s pricing thesis. Nebius can use Meta to help de-risk capacity utilization while retaining the opportunity to sell portions of that infrastructure to other customers at potentially higher prices. If market pricing remains strong, the company could capture upside. If demand softens, Meta’s commitment provides a degree of protection.
For a company spending aggressively to add capacity, that is unusually valuable contract architecture.
Nvidia’s $2 Billion Investment Adds Another Layer of Validation
Nebius also has something few independent AI cloud companies can claim: a direct strategic investment from Nvidia. In March, Nvidia agreed to invest $2 billion in Nebius as part of a broader engineering partnership designed to support more than 5 gigawatts of Nvidia-powered systems by the end of 2030. The companies are collaborating across AI-factory design, inference, fleet management and early deployment of next-generation hardware including Vera Rubin.
For NBIS investors, that relationship has several potential advantages. Early access to new Nvidia architectures can help Nebius offer scarce capacity before smaller rivals, while engineering collaboration may improve utilization and system performance. In a market where GPU availability and deployment speed can determine customer wins, those advantages can translate directly into pricing power.
The investment also acts as a strong external endorsement. Nvidia does business with virtually every major cloud provider, but committing $2 billion of its own capital suggests it sees strategic value in Nebius becoming a large independent AI infrastructure platform.
That does not eliminate risk, but it makes the company harder to dismiss as simply another speculative GPU renter.
Palantir Could Open a New Enterprise Revenue Channel
Nebius added another strategic partner this week when Palantir named it its preferred sovereign AI infrastructure partner. Under the agreement, Nebius compute and inference endpoints will be integrated into Palantir’s enterprise environment, allowing eligible customers to access AI infrastructure while maintaining control over data, models and compute.
No contract value or minimum purchase obligation was disclosed, so investors should not attach billions of dollars of immediate revenue to the announcement. Its significance is more strategic. Palantir is expanding extremely quickly among enterprises and governments that want AI systems connected to proprietary operational data, and many of those customers also care intensely about sovereignty, security and infrastructure control.
If Palantir starts routing meaningful customer workloads through Nebius, NBIS could diversify beyond giant hyperscale customers such as Meta and build a broader enterprise revenue base. That would potentially improve both pricing and customer concentration over time.
Truist’s new coverage arrives at exactly the moment that this distribution opportunity is opening.
The Bull Case Still Comes With an Enormous Capex Bill
The biggest challenge to the bullish thesis is simple: Nebius has to spend enormous amounts of money before it can earn the revenue Truist is modeling.
The company has been raising billions to fund data centers and GPU purchases, including an approximately $5.75 billion convertible-note offering completed in August. The capital is intended to finance rapid infrastructure expansion, but the financing also highlights how cash-hungry the business remains.
Independent estimates put Nebius’s fiscal 2026 capital expenditure ambitions in the $20 billion to $25 billion range. That type of spending can create extraordinary growth if capacity is quickly filled at premium pricing, but it creates serious risk if AI demand slows, customers delay projects or the company misjudges where to build.
The fact that Nebius has strong customer contracts helps, but it does not eliminate the mismatch between cash spent today and revenue received over years. This is one reason the cloud-pricing discussion matters so much. Higher revenue per megawatt means faster payback and better return on invested capital. Lower pricing would make the same infrastructure considerably less attractive.
Customer Concentration Is Another Risk Investors Should Not Ignore
Nebius’s explosive growth has come partly from a relatively concentrated customer base. Recent analysis indicated that three customers accounted for around 59% of Q2 revenue, highlighting how strongly near-term results can depend on a limited number of large counterparties.
Large contracts are not inherently problematic. Hyperscale AI deployments naturally involve enormous customers because relatively few companies can consume hundreds of megawatts of accelerated compute. But concentration increases risk if a customer delays capacity, renegotiates deployment schedules or decides to move more workloads onto internally developed silicon.
That makes the Palantir relationship strategically interesting even without a disclosed dollar value. A larger mix of enterprise customers could eventually reduce dependence on a handful of hyperscalers while potentially improving pricing because smaller customers often lack the bargaining power of Meta-sized buyers.
Nebius therefore needs both types of customers: huge anchor contracts to justify capacity construction and smaller high-value workloads that maximize economics once the infrastructure exists.
The $355 Target Also Shows How Much Optimism Is Already in NBIS
Nebius closed September 9 around $240.35, giving the company a market capitalization of roughly $65 billion. Truist’s $355 target implies upside of about 48%, but it would also push Nebius toward a valuation approaching $100 billion depending on future share count.
That is an enormous valuation for a company expected to generate only a few billion dollars of revenue this year, and it means conventional earnings multiples provide limited help. Investors are instead valuing Nebius based on future infrastructure capacity, long-term customer commitments and the assumption that AI compute remains scarce enough to support attractive pricing for years.
Other analysts are also bullish but vary considerably in how much upside they see. Citi recently raised its target to $324, while Baird moved to $340. DA Davidson, by contrast, maintained a Neutral rating with a $175 target, underscoring how widely opinions differ around the amount of future success already priced into the shares.
That disagreement is exactly what makes NBIS so volatile. Small changes in assumptions around pricing, capacity or capital spending can produce enormous differences in theoretical valuation.
Nebius Stock: Truist’s $355 Target Depends on One Number More Than Any Other
The new Buy rating reinforces a bull case that has become substantially stronger over the past six months. Nebius has posted 454% quarterly revenue growth, produced significant adjusted EBITDA from its core AI cloud business, secured a potentially massive Meta relationship, attracted a $2 billion investment from Nvidia and opened a new enterprise channel through Palantir.
But Truist’s most interesting argument is not simply that Nebius will deploy more GPUs. Investors already know the company is expanding aggressively. The more powerful question is how much revenue each megawatt of that capacity can generate.
If recent Blackwell pricing proves durable and Nebius can combine long-term anchor customers with premium short-duration workloads, revenue could rise faster than capacity alone would suggest. That would improve returns on the company’s enormous infrastructure investments and make the path toward Truist’s $355 target easier to defend.
If cloud pricing normalizes sharply once more capacity enters the market, the opposite happens. Nebius could still grow quickly while producing lower returns than investors currently expect, putting pressure on a valuation that leaves little room for disappointment.
That is why the next stage of the NBIS story is no longer simply about signing bigger contracts.
Nebius has already proved customers want the capacity. Now it needs to prove just how valuable every megawatt can become.
Disclaimer
This article is for informational purposes only and does not constitute financial or investment advice. Readers should conduct their own research and consider consulting a qualified financial advisor before making investment decisions. This article was researched and drafted with the support of AI, then reviewed, fact-checked and edited by the editorial team before publication.
