OpenAI Lowers Funding Goal to $20B Amid Compute Rivalry with Anthropic

As OpenAI scales back its latest fundraising target to $20 billion amid broader market cooling, its race for compute against Anthropic is splitting into two distinct capital paths.

According to reports from The Wall Street Journal and other financial outlets, OpenAI has lowered its fundraising target from an initial $30 billion to $20 billion, adjusting its target valuation from $1.4 trillion to $1.25 trillion. The company has reportedly secured around $11 billion in initial commitments from backers including UAE-based tech fund MGX, Sequoia Capital, Andreessen Horowitz, and Founders Fund. These figures are based on press reporting, and OpenAI has not issued an official announcement.

The adjustment comes as rival Anthropic accelerates preparations for a mid-November Nasdaq listing after filing its Form S-1(the registration paperwork a company files with the U.S. Securities and Exchange Commission to go public) with the SEC. Both labs face the pressure of securing hundreds of billions of dollars in data center capacity to train and serve frontier AI(the most advanced AI models). The latest financing efforts reflect a battle to shore up long-term balance sheets before the next wave of model scaling.

Confirmed details and unverified reports

Reporting indicates that OpenAI adjusted its fundraising terms in response to investor caution regarding the timeline for monetizing massive AI infrastructure expenditures. To bridge the remaining gap toward the $20 billion mark, OpenAI reportedly plans to approach family offices and high-net-worth individuals at the same $1.25 trillion valuation. Because the round has not officially closed, final commitments and terms remain subject to change.

In contrast, the figures surrounding Anthropic stem directly from regulatory disclosures. Anthropic is pursuing an initial public offering (IPO, a company’s first sale of shares to the public) with an SEC filing that outlines long-term contractual obligations with hyperscalers(the large cloud companies) and hardware vendors, providing a rare window into the financial requirements of operating frontier models.

What the compute commitments reveal

The infrastructure obligations disclosed by both organizations resemble capital expenditure programs typically seen in national telecommunications or energy utilities.

OpenAI has outlined a compute budget of roughly $750 billion through 2030. With annual infrastructure and inference(using a finished model to produce answers) costs running into tens of billions of dollars, maintaining active training clusters while serving millions of API requests requires continuous capital injections.

Anthropic’s S-1 filing outlines more than $587 billion in binding long-term commitments for compute and data centers through 2045. That includes an obligation of at least $44.9 billion with Microsoft Azure through 2033. In the first half of 2026 alone, Anthropic spent $8.9 billion on compute and data center capacity, with roughly 73% dedicated to research and development rather than serving production traffic.

Contrasting infrastructure deployment models

The structural approach to acquiring compute differs sharply between the two organizations.

OpenAI relies on a centralized hyperscaler model, using Microsoft Azure as its primary backbone supplemented by large dedicated clusters on Oracle Cloud. Its long-term roadmap focuses on massive single-site deployments, such as multi-gigawatt data center campuses planned in Georgia, optimizing network topology for distributed training of large models.

Anthropic has pursued a multi-cloud, multi-silicon distribution strategy. Alongside its primary footprint across Amazon Web Services (AWS) and Google Cloud, it holds commitments with Microsoft Azure and leases tens of billions of dollars in Google TPU capacity. Anthropic also spreads workloads across specialized neoclouds(cloud providers specializing in renting out GPUs for AI workloads rather than broad IT services) and alternative hardware providers to avoid single-vendor bottlenecks.

Their financing paths reflect this structural divergence. OpenAI continues to rely on mega-rounds in private markets, leaning on sovereign funds and venture partners to preserve its private capitalization structure. Anthropic is using a public listing to tap institutional capital markets, aiming to fund its long-term infrastructure liabilities through publicly traded equity.

What this means for developers

These financing adjustments do not signal immediate price reductions or sudden service disruptions for production workloads. However, capital availability directly influences long-term development roadmaps and operational stability.

When a research lab commits more than 70% of its compute footprint to research and training, funding delays can slow the rollout cadence of frontier checkpoints or postpone inference capacity expansions. Conversely, securing sufficient capital helps ensure reliable API latency and higher request throughput limits during usage surges.

For engineering teams with architectures built around a single provider, the diverging compute and infrastructure strategies highlight the practical value of maintaining multi-vendor routing layers to mitigate provider-level supply constraints.

What to watch next

Key milestones include the final closing terms of OpenAI’s $20 billion round and whether its valuation holds at $1.25 trillion. In parallel, market pricing for Anthropic’s Nasdaq debut in mid-November will offer the first public market test of investor appetite for long-term frontier AI capital requirements.