Anthropic CFO Krishna Rao is leading early IPO meetings with investors and has not discussed valuation, sources say
By Imani Sutton ·
The entire conversation surrounding Anthropic’s potential IPO—the one that CFO Krishna Rao is steering toward—is running on the same kind of speculative vapor that once inflated…
This IPO is Running on Speculative Vapor, Not Line Items
The entire conversation surrounding Anthropic’s potential IPO—the one that CFO Krishna Rao is steering toward—is running on the same kind of speculative vapor that once inflated NASDAQ indices to absurd heights. We are supposed to be impressed by the sheer numbers: sources familiar with the matter suggest investors expect a valuation of $2 trillion or more, while fortune.com confirms Anthropic recently closed its Series H funding at a staggering $965 billion post-money valuation. These figures, which have been splashed across outlets like CNBC, feel less like financial metrics and more like pure narrative engineering designed to distract from the ledger.
The meetings are reportedly high-level, focusing on Claude’s models and its position in the enterprise market, but conspicuously avoiding specific financials or a valuation discussion, according to CNBC. This isn't due diligence; it is performance art for venture capital hype cycles. The problem isn't that they haven't set the price—the company itself stated: "The number of shares to be offered, and the price have not yet been set." The problem is that the valuation being floated doesn’t account for the actual physical load required to sustain it.
Compute Procurement Is Not a Line Item on an SEC Filing
I keep coming back to the same fundamental point: every promise of a clean, smart, frictionless future runs on somebody's grid, and nobody in this room seems to be looking at the utility bill. Rao himself stated that "Compute that we procure is the lifeblood of our business." He noted they are customers of Amazon’s Tranium chip, Google's TPUs, and Nvidia's GPUs. This isn't a software problem; it is an immense, physical power consumption problem.
The sheer scale required to maintain this 'frontier AI' capability means Anthropic is not simply selling algorithms; they are buying megawatts of electricity and acres of specialized cooling infrastructure. When you talk about $2 trillion valuations, you have to account for the cost of keeping those servers running—the actual load factor that determines whether a metro can survive two inches of snow or if a data center will trip its breaker during a heat advisory. The fact that dnyuz.com reports Rao confidentially submitted a draft registration statement on Form S-1 only highlights that they are merely filing paperwork for an economic fantasy, not documenting sustainable profit.
Selling Potential Energy Instead of Kinetic Energy
This isn't new; it’s just wearing better branding than dial-up internet. We are witnessing a classic repeat performance, one that finds its clearest precedent in the dot-com bubble. Both periods share a mechanism: revolutionary technology creates an investment environment where perceived future market dominance and disruptive potential are valued far higher than current financial metrics. In the late 90s, investors were willing to fund anything with a ".com" suffix because they believed the internet was fundamentally changing everything—a belief that blinded them to basic profitability or sustainable infrastructure costs.
The hype around Anthropic is doing exactly the same thing. It’s selling potential energy instead of kinetic energy. The market has decided that "disruptive potential" is worth more than a stable, readable utility bill. The $2 trillion valuation is a spectacular, beautiful hallucination built on vaporware capital. It ignores the fact that power grids have finite capacity, and that cooling servers in Atlanta during a summer heatwave costs money—real, physical dollars paid to Georgia Power or whoever owns the conduit.
The market has forgotten how to read line items.