To The Moon

$NEBIUS(NBIS)$  , an Amsterdam-based AI infrastructure company, is positioning itself as a leader in GPU-powered cloud services, data labeling (via Toloka), autonomous vehicles (Avride), and edtech (TripleTen). Its Q1 2025 financials showed robust revenue growth of 385% year-over-year, reaching $55.3 million, driven by its core AI infrastructure business. Despite a net loss of $113.6 million, the company reaffirmed its guidance for an annualized run-rate revenue (ARR) of $750 million to $1 billion by year-end 2025, boosting investor confidence and stock performance.

Toloka’s Strategic Investment: In May 2025, Toloka, Nebius’s AI data solutions arm, secured a $72 million investment led by Bezos Expeditions, with participation from Shopify CTO Mikhail Parakhin. This move relinquished Nebius’s majority voting control, granting Toloka greater independence to scale its operations, particularly in the U.S. market. Toloka specializes in high-quality data labeling, agent safety evaluation, and red teaming, serving clients like Amazon, Microsoft, Anthropic, and Shopify. The investment underscores the critical role of human-in-the-loop data services in ensuring AI models are accurate, safe, and scalable. Nebius retains a significant economic stake but will no longer consolidate Toloka’s financials, focusing instead on its core AI infrastructure.

Capital Expenditure and Expansion: Nebius is aggressively expanding its AI infrastructure, with a planned $2.5 billion CapEx for 2025 to build data centers in New Jersey (up to 300 MW, first phase online by summer 2025), Kansas City (40 MW, first phase Q1 2025, second phase Q2 2025), and Iceland (geothermal-powered, operational Q1 2025). It’s also tripling capacity at its Finland data center. These facilities are tailored for Nvidia’s Blackwell GPU architecture, with Nebius deploying 22,000 Blackwell GPUs in 2025, supported by a $700 million financing round from Nvidia, Accel, and others. This partnership ensures early access to cutting-edge hardware, strengthening Nebius’s competitive edge in AI compute.

Meta’s Potential Investment in Scale AI

Meta Platforms is reportedly negotiating a landmark investment exceeding $10 billion in Scale AI, a leading data-labeling startup valued at $14 billion in 2024 and potentially $25 billion in a 2025 tender offer. Scale AI provides data labeling services critical for training machine-learning models, serving clients like Microsoft and OpenAI. This potential deal, if finalized, would be one of the largest private funding events, reflecting the soaring demand for high-quality AI training data amid the generative AI boom. Meta’s move aims to bolster its AI infrastructure to compete with Microsoft, Amazon, and Alphabet, though terms remain fluid.

Broader Context and Critical Analysis

The surge in investments in data labeling firms like Toloka and Scale AI highlights the foundational role of high-quality, human-annotated data in generative AI development. As AI models grow more complex, the need for accurate, safe, and ethically aligned datasets is paramount, creating a lucrative niche market estimated at $10 billion for Toloka alone. However, the high valuations—Nebius at 20–25x projected 2025 revenue and Scale AI potentially at $25 billion—raise concerns about overvaluation in a competitive, capital-intensive sector. Nebius’s $283 million cash burn in 2024 and -59% EBITDA margin signal execution risks, despite its $2.9 billion cash reserve.

Meta’s potential $10 billion investment in Scale AI, while strategic, reflects the escalating AI arms race among tech giants. However, the lack of finalized terms introduces uncertainty, and Scale AI’s reliance on a few key clients like Microsoft and OpenAI could pose risks if market dynamics shift. Additionally, the broader AI infrastructure market faces challenges like energy constraints and regulatory scrutiny, which could impact Nebius’s data center expansions and Scale AI’s growth.

Conclusion

Nebius is leveraging Toloka’s $72 million investment and its $2.5 billion CapEx plan to solidify its role in AI infrastructure and data services, backed by Nvidia and Bezos. Meanwhile, Meta’s potential $10 billion investment in Scale AI underscores the critical demand for data labeling in the generative AI boom. While both companies are well-positioned, their high valuations and operational risks warrant caution. Investors should monitor Nebius’s Q2/Q3 2025 earnings call  for further clarity on Toloka’s spin-off and expansion progress, and track developments in Meta’s Scale AI deal for signs of strategic shifts in the AI landscape

@TigerObserver  @Daily_Discussion  @TigerPM  @TigerStars  @Tiger_comments