Anthropic's 2025 Financials Raise Market Alarm

Public response to Anthropic's 2025 financial figures has been largely negative, primarily due to the substantial size of the reported losses.

Meet Kevin, an analyst, frames Anthropic as a 'canary in the coal mine' for the broader AI economy, suggesting its performance could foreshadow trends for the entire sector.

He argues that if major AI firms like Anthropic and OpenAI face funding crises, the current cycle of GPU-heavy investment could collapse, leading to significant market instability from a 'circular investment' model.

The focus on 2025 figures for Anthropic's financials is a standard regulatory procedure for IPO filings, rather than an attempt to obscure recent performance, as audited financial statements for the previous fiscal year are typically due by April 30th.

Companies often utilize PCAOB-audited financials for IPOs or significant fundraising rounds, and while 2026 data is not yet fully audited, supplementary data such as quarterly revenue often helps fill this temporal gap.

Operating expenses hit $12.6 billion against $4.59 billion in revenue, with projected annualized run rates reaching $65 billion.

The company incurred $12.6 billion in operating expenses, leading to an $8.06 billion loss, with an additional reported loss of $42 billion attributed to derivative valuation changes linked to stock price increases.

Furthermore, Anthropic faces substantial future obligations, including $518 billion in cloud compute and infrastructure costs over the next decade, which could equate to approximately 10 gigawatts of power consumption.

The speaker contends this model artificially inflates company valuations and questions the true profitability of AI firms, suggesting a lack of genuine returns for investors beyond the initial loop.

He labels current profitability claims as 'fugazy' due to the specific, often adjusted, methods used for their calculation, suggesting they do not reflect a complete financial picture.

Adjusted profit calculations for Anthropic exclude significant financial components, such as stock-based compensation, which typically forms a substantial part of employee remuneration in tech companies.

Furthermore, these metrics strip out revenue shared with distribution partners, including Amazon, directly impacting the reported top-line figures.

Crucially, the costs associated with training large language models, a core operational expense for an AI company, are also excluded from these adjusted figures, leading to a potentially misleading view of profitability.

OpenAI reports a robust annualized recurring run rate of approximately $50 billion, showcasing significant revenue growth in the AI sector.

In contrast, Anthropic's run rate appears to be flatlining, according to the speaker, making the timing of its IPO critical to maximize valuation before comparative performance with competitors potentially worsens.

When measured against small historical figures, revenue growth can appear as high as 900%, or a tenfold increase, creating an inflated perception of rapid expansion.

However, these percentage growth rates are expected to drop to approximately 4x as the company 'laps' larger revenue quarters, eventually slowing to around 30% or lower, which typically triggers investor concern.

Continued annualized recurring revenue (ARR) growth is the primary requirement for ongoing funding of the AI buildout, as investors maintain confidence despite losses as long as the company demonstrates expansion.

The cessation of AI funding serves as a major 'red flag' for the entire industry, indicating that stagnation in recurring revenue could prompt investors to withdraw support from the sector, thereby undermining its systemic growth.

This governance model is intended to insulate the company from lawsuits regarding profit-focus, enabling Anthropic to prioritize safety over immediate profitability, and includes notable figures like Ben Bernanke and Reed Hastings in leadership and trustee roles.

The actual profitability numbers and sustained growth rates for Anthropic remain largely unknown, making future quarters critical for evaluating the company's trajectory.

The quarters between Q2 2027 and Q4 2027 are highlighted as essential periods to determine if the AI bubble will 'pop,' as the success of the IPO is secondary to the long-term evolution of the company's financials.

The speaker also speculates about a potential euphoric market rally contingent on an eventual Iran deal, which could influence broader market sentiment, though its direct impact on Anthropic is not detailed.