Anthropic, the artificial intelligence startup behind the Claude chatbot, has revealed in its IPO filing that it has committed to spending more than $800 billion on computing infrastructure. The eye-popping figure underscores just how capital-intensive the race for AI dominance has become—and how much of that spending is locked in for decades to come.
The filing, which has already lifted futures and tech stocks, shows that Anthropic's signed deals stretch all the way to 2045. Among the biggest commitments: $161 billion with chip designer Broadcom, $150 billion for additional Google chips, and $120.5 billion with SpaceXAI, a provider of AI-focused cloud infrastructure.
Explosive growth, but at a price
Anthropic's revenue has grown at a staggering pace. The company brought in $4.6 billion last year, and in the first nine months of 2026 alone, revenue topped $33 billion. That kind of growth is rare even in the fast-moving tech sector.
But the spending commitments are equally dramatic. The company is essentially betting that the demand for AI will keep growing for the next two decades, justifying the massive upfront investment in data centers, chips, and other infrastructure.
There are signs that the bet is paying off—at least in terms of efficiency. Anthropic says its computing costs have fallen from over $2.50 for every dollar of revenue in early 2025 to less than 50 cents by mid-2026. That means the company is getting much more value out of each dollar spent on computing, a trend that could help narrow its losses over time.
Why the spending is so large
Building and running large AI models requires enormous computing power. Companies like Anthropic need thousands of specialized chips, vast data centers, and reliable energy supplies. The costs are so high that even the biggest tech firms are forming partnerships and making long-term deals to secure capacity.
Anthropic's approach appears to be to lock in that capacity early, even if it means committing to spending billions of dollars years in advance. The deals with Broadcom and Google are designed to ensure a steady supply of the most advanced chips, while the SpaceXAI agreement likely covers cloud computing services.
This is not unique to Anthropic. Rivals like OpenAI and Google have also made massive infrastructure commitments. But the scale of Anthropic's commitments—relative to its current revenue—is notable.
What it means for investors
For everyday investors, the key takeaway is that AI is a capital-intensive business. Companies that want to lead in AI must spend heavily on infrastructure, and that spending can weigh on profits for years.
Anthropic's revenue growth is impressive, but its spending commitments mean that the company will need to keep growing at a rapid clip to justify the outlays. If demand for AI services slows, or if competitors introduce cheaper alternatives, Anthropic could find itself with more capacity than it needs.
On the other hand, if AI adoption continues to accelerate, these long-term deals could give Anthropic a competitive advantage by ensuring it has the computing power to meet demand.
The IPO itself is one of the most anticipated in recent memory, and its filing has already lifted European tech stocks and rallied emerging-market tech. Investors will be watching closely to see how the company prices its shares and whether it can maintain its growth trajectory.
The bigger picture
Anthropic's filing is a reminder that the AI boom is not just about software—it's also about hardware and infrastructure. The companies that supply the chips, data centers, and energy will benefit from this spending spree, even if the AI startups themselves take years to turn a profit.
For investors, it's worth paying attention to the entire AI supply chain, from chipmakers like Broadcom to cloud providers. The broader market backdrop also matters: if interest rates stay high, the cost of financing these massive projects could rise, putting pressure on companies like Anthropic.
Ultimately, Anthropic's $800 billion bet is a bet on the future of AI. Whether it pays off will depend on how quickly the technology is adopted and how efficiently the company can turn its spending into revenue. For now, the numbers are staggering—and so are the risks.
