Anthropic's Pre-IPO Playbook: Tightening Control, Widening Moat, and a $6 Billion Bet on Decart

Anthropic nears record IPO with $65B annualized revenue, dual-class governance, and new chip efforts to cut Nvidia dependence.

The road to what could be the largest public market debut in history is rarely a straight line, and Anthropic is no exception. As the AI heavyweight barrels toward a potential listing that some project could rival SpaceX's record-setting $1.77 trillion valuation, the company is simultaneously tightening its internal governance, expanding its hardware ambitions, and preparing to absorb a major acquisition target.

Behind the scenes, the machinery of an IPO is already grinding into gear. A revolving credit facility that was initially sized at roughly $10 billion is now expected to overshoot that target as banks jockey for mandates ahead of the listing. Meanwhile, the company is putting the finishing touches on a dual-class share structure that would hand CEO Dario Amodei and his co-founders supervoting rights, alongside a separate share class for non-shareholding trustees who will appoint the majority of the board. That governance layer has recently thinned: following the departure of Mariano-Florentino Cuéllar, who stepped up to Chief Global Affairs Officer, the Long-Term Benefit Trust now counts just three members — Neil Buddy Shah, Richard Fontaine, and Ben Bernanke.

The financial trajectory underpinning all this activity is nothing short of extraordinary. Annualized revenue hit $65 billion by the end of July, up from $47 billion in May and a mere $9 billion at the close of 2025. Bloomberg reports that investors are now penciling in full-year 2026 revenue between $100 billion and $120 billion. That torrid pace helped fuel a Series H round in May that valued the company at $965 billion.

Yet for all the growth, Anthropic is signaling a more cautious internal posture. Its August risk report — dated the 15th — nudged the assessment of misalignment risk in high-stakes scenarios from "very low" to "low," citing recent cybersecurity incidents. The report also references an internal model dubbed "Model 2," which outperforms the current flagship Mythos but is not slated for external release. Axios has reported the model shows marked improvement on internal tasks, though no launch is planned. Automated AI research risk remains rated low overall, though the company flags early signs of acceleration in research automation.

That measured tone extends to the enterprise data front. A new policy, reported by Reuters on Thursday, retains the existing 30-day storage window while handing corporate clients more say over how their data interacts with Anthropic's most advanced models. A new security system is slated for later this year.

On the hardware side, Anthropic is moving to loosen its dependence on Nvidia. The company has brought on Amir Salek, formerly head of Google's custom chip program and TPU business, to lead semiconductor development. That effort is complemented by a $250 million chip order from UK-based Fractile, with an option to expand, plus capacity agreements with Riot Platforms and Volta Infra Holdings.

The most conspicuous piece of the pre-IPO chessboard, however, is the reported acquisition of Decart, an Israeli startup backed by Nvidia. Bloomberg pegs the potential deal at roughly $6 billion — which would mark Anthropic's largest known acquisition to date, though its completion remains uncertain.

Competitive pressure is mounting even as the company holds its ground. Payment processor Ramp's data shows Anthropic retained 44 percent of tracked US business users in July, versus 40 percent for OpenAI — though OpenAI is growing faster in the third quarter. The Financial Times has noted customers increasingly gravitating toward cheaper alternatives like GPT-5.5 or open models rather than Anthropic's premium Fable 5 offering, raising questions about whether technological advances will translate into higher customer spend or force prices downward.

There have been regulatory bumps as well. Crypto exchange OKX briefly suspended Anthropic's corporate account over a compliance violation and subsequently barred employees in Hong Kong and China from accessing Claude.

For investors eyeing a future listing, the picture is one of deliberate preparation: explosive revenue growth, structural safeguards for founder control, a beefed-up security regime, and strategic bets on both silicon and startups. Whether the market will reward that combination at the scale of a $1.77 trillion debut remains the open question — but Anthropic is clearly building the scaffolding for something historic.