Anthropic ended July with an annualized revenue run rate of $65 billion, up roughly 600% from the end of 2025, underscoring how quickly generative-AI demand is translating into large-scale commercial buildouts. While the company remains private, its growth trajectory is already feeding into two major partners’ capacity plans: Amazon Web Services and Space Exploration Technologies.
Key takeaways
- Run-rate surge: Anthropic’s annualized revenue hit $65 billion in July, about 600% higher than at the end of 2025.
- Primary catalysts: The company has committed to major compute and cloud relationships with Amazon and SpaceX, including pricing and capacity arrangements through 2029.
- Partner implications: Amazon and SpaceX are positioned as critical infrastructure providers if Anthropic’s growth persists.
- Profitability questions: The report highlights uncertainty over model-building costs, which may pressure profitability even as revenue scales.
What drove Anthropic’s growth visibility
“Run rate” figures reflect taking current revenue and annualizing it, meaning the number is a snapshot rather than a finalized financial statement. Still, Anthropic’s reported trajectory is significant. The article says its preliminary Q2 revenue was around $11.5 billion, which it frames as roughly 14 times what Anthropic generated over the same period last year.
Anthropic is preparing for a public market debut, according to the article. It filed confidential IPO paperwork with the Securities and Exchange Commission in June 2026, but has not disclosed a listing date.
How Amazon and AWS are positioned to benefit
Amazon is described as Anthropic’s largest corporate backer, with about $13 billion invested so far and an option to commit up to $33 billion. The article also notes Amazon holds a minority stake and has no board seat.
Beyond equity, the commercial commitment is the focal point for investors. According to the article, Anthropic has committed $100 billion to Amazon cloud services over the next 10 years, and has secured up to 5 gigawatts of capacity running on Amazon’s in-house AI chips. The piece adds that if Anthropic continues expanding at its current pace, it may need to increase that commitment.
For Amazon, the implication is straightforward: rapid scaling from Anthropic increases the likelihood of sustained demand for high-performance compute and power-intensive infrastructure—provided Anthropic can keep financing model development at scale.
SpaceX compute deals extend through 2029
The article states that Anthropic pays Space Exploration Technologies roughly $1.25 billion a month for compute capacity through 2029, citing SpaceX’s IPO filing. It characterizes the arrangement as commercial in nature, with no equity involved.
While the two companies have competed in the past, the piece says the relationship has evolved: Anthropic is described as SpaceXAI’s biggest customer. According to the article, Anthropic leases compute running on more than 300,000 Nvidia chips from SpaceX, sustaining the capacity payments described above.
The article also notes Alphabet’s separate lease with SpaceX at $920 million a month, with that agreement ramping up starting in October.
It further claims that Anthropic accounts for roughly half of SpaceXAI’s total sales, meaning SpaceX’s near-term revenue visibility is tightly linked to Anthropic’s continued compute needs.
Bigger picture: IPO timing and the profitability debate
Even as revenue run-rate accelerates, the article flags a central risk: the economics of frontier model building. It argues that while the market has growing clarity on revenue momentum, costs may scale differently—especially because developing frontier models is extremely expensive.
According to the article, Anthropic has been reported as “profitable operationally,” but that framing may exclude the most costly component from the perspective of model developers: building the models themselves. The piece suggests that costs appear to scale with revenue so far, but emphasizes that this remains an open question.
From an investor standpoint, the article implies a two-sided dynamic for both partners. More Anthropic demand can mean more compute utilization and steadier infrastructure contracts for Amazon and SpaceX. At the same time, if Anthropic’s spending requirements outpace its ability to monetize models efficiently, it could eventually strain financial outcomes for customers locked into compute-heavy relationships.
On the market’s timeline, an IPO could provide a clearer view of these economics once full financial disclosures arrive. The article cautions, however, that more detailed cost visibility could also change investor perception if profitability—particularly the model-building layer—is weaker than expected.
Investors will likely look next for updates from Anthropic ahead of its planned listing, including the IPO timeline and, when available, full financial disclosures that separate operational profitability from model-development costs. For partners, the key watch items include whether Anthropic expands compute commitments beyond current levels and whether additional customer demand offsets the concentration risk highlighted by the article.
