Nvidia Cuts OpenAI Deal as Anthropic Revenue Surges

Nvidia has scaled back its planned financial backing for a massive data center project with OpenAI, reducing its initial guarantee from $250 billion to just under $120 billion. The decision, driven by investor concerns over Nvidia's risk exposure, covers the first construction phase of a project designed to deliver approximately five gigawatts of capacity. OpenAI is independently negotiating a lease for the complete 10-gigawatt development, which is being built by SoftBank subsidiary SB Energy. Additionally, Nvidia remains in discussions regarding separate financing for OpenAI's chip acquisitions, a deal valued at up to $350 billion.

While Nvidia's cautious move might suggest growing industry hesitation, Anthropic's financial performance presents a starkly different picture. The AI startup saw its revenue jump from $4.73 billion in the first quarter to more than $11.5 billion in the second quarter, representing a massive 14-fold increase year-over-year. Anthropic projects its annual revenue will reach between $190 billion and $200 billion by 2028, a massive leap from the $45 billion run rate it reported in May. The company, which achieved tenfold annual revenue growth for three consecutive years leading up to early 2026, is reportedly preparing for an initial public offering in late September or early October at a valuation close to $1 trillion.

For AI practitioners and developers, these contrasting developments signal a complex market transition. On one hand, hardware and infrastructure providers are facing stricter financial scrutiny, which could eventually bottleneck the rapid expansion of massive compute clusters. On the other hand, the soaring demand for proprietary models suggests that enterprise adoption remains incredibly strong. However, practitioners should monitor potential shifts in usage; financial services firm Ramp recently noted a slight flattening in business demand for Anthropic's tokens, indicating that while overall market revenue is surging, individual enterprise consumption patterns may be starting to stabilize.