Google to Invest Up to $40 Billion in Anthropic Through Cash and Computing
Google just confirmed one of the biggest financial moves in recent AI history. The Silicon Valley giant will invest up to $40 billion in Anthropic, combining cash and computing power in a deal that redefines the weight of this strategic partnership between two of the most influential companies in the global AI ecosystem.
The announcement, reported by Bloomberg, caught a lot of people off guard — not because there was a relationship between the two companies, that was already public knowledge, but because of the sheer scale of the financial commitment that was put on the table.
The structure is straightforward: $10 billion comes in now, with Anthropic valued at $350 billion, and another $30 billion arrives later, contingent on meeting performance milestones previously agreed upon by both parties. This means Google is not just betting on Anthropic’s present — it is also tying the future of this relationship to concrete results, which is a pretty calculated move to ensure that money translates into real progress rather than just promises.
That is no small thing.
This investment comes at a time when Anthropic is under pressure from every direction, whether from widespread complaints about Claude usage limits or from the race to secure enough infrastructure to scale its AI models without hitting bottlenecks. The company has been growing at a breakneck pace, but rapid growth in AI comes with a massive cost — and that cost has a name: computing.
The Mythos Model and the Strategic Timing of the Investment
The timing of this announcement is no coincidence. The investment promise came right after Anthropic launched its most recent model, called Mythos, to a limited group of partners. According to the company itself, Mythos is the most powerful model it has ever created and has significant applications in the cybersecurity space.
Precisely because of the potential for misuse, Anthropic decided to restrict broader access while working with selected organizations to evaluate and mitigate risks. Even so, the model reportedly already ended up in unauthorized hands, according to Bloomberg — a sign that when something is powerful enough, maintaining control over it becomes a challenge in itself.
Another relevant detail is that Mythos is likely quite expensive to run at scale, which directly connects the need for massive infrastructure with the billion-dollar deal signed with Google. Without the necessary computational support, even the most advanced model in the world would be stuck in a lab, unable to reach the people who actually need it.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
The restricted launch is a classic strategy for validating capabilities before a wider rollout, but it also signals that Anthropic is playing in a different league, laying the groundwork for something much bigger than what is available to the general public today. 🚀
Why Computing Became the Center of Everything
When we talk about modern AI models, we are talking about systems that consume an absurd amount of computational resources to function. Training a large-scale language model requires thousands of specialized chips running for weeks or months on end, plus energy and cooling infrastructure that very few companies in the world can sustain.
That is exactly why the computing component within the Google and Anthropic deal is just as relevant as the money itself — in practice, they are the same thing dressed up in different forms.
Google has a massive advantage in this scenario because it operates one of the largest computing infrastructure networks on the planet, including its own AI chips, the TPUs (Tensor Processing Units). These chips are developed in-house and optimized specifically for machine learning workloads, and they are considered one of the best alternatives to Nvidia processors, which remain in high demand on the market.
Anthropic already relies heavily on Google Cloud for chips and infrastructure, including access to these TPUs. By expanding that capacity so significantly, Google is not just opening the vault — it is opening the data center. And that makes an enormous difference for a company that needs to scale its AI models quickly and efficiently without depending on external suppliers in increasingly competitive and scarce market conditions.
The Broadcom Partnership and Gigawatt-Scale Computing
The relationship between Anthropic and Google on infrastructure did not start with this announcement. Earlier this month, Anthropic had already revealed a partnership with Google and chip manufacturer Broadcom, which designs custom AI chips for Google, to access multiple gigawatts of TPU-based computing capacity starting in 2027. A Broadcom regulatory filing put that number at 3.5 gigawatts.
Google’s new investment expands that arrangement considerably. Google Cloud will now provide an additional 5 gigawatts of capacity over the next five years, with room to scale even further. To put that in perspective, we are talking about an amount of computing power that could supply a mid-sized city — all dedicated to running AI models.
This move also creates an interesting strategic side effect: it positions Google as an indispensable infrastructure partner for Anthropic, generating a technical dependency that goes well beyond a traditional financial investment relationship. When one company’s systems run on top of another’s infrastructure, the ties that form are deep and hard to undo, which guarantees Google a central role in Anthropic’s future growth, regardless of how the AI market evolves in the coming years.
The Infrastructure Race Defining the AI Era
The race for leadership in artificial intelligence is increasingly being defined by access to the computing power needed to train and deploy these systems. And Anthropic is not alone in this pursuit.
OpenAI, its main rival, has been moving aggressively to secure that capacity through a network of multi-billion-dollar deals with cloud providers, chip suppliers, and energy companies. This month, OpenAI expanded an agreement with chip manufacturer Cerebras, in a transaction involving more than $20 billion in chips and even an equity stake.
On Anthropic’s side, the company has also been racing to close infrastructure deals on multiple fronts:
- CoreWeave: Earlier this month, Anthropic closed a deal with cloud computing provider CoreWeave for data center capacity.
- Amazon: This week, the company secured an additional $5 billion investment from Amazon, part of a broader deal in which Anthropic is expected to spend up to $100 billion on roughly 5 gigawatts of computing capacity over time.
- Google and Broadcom: The partnership for access to gigawatts of TPU-based capacity, now expanded by the new investment.
This web of deals shows that Anthropic is building computing infrastructure at a truly global scale, distributed among multiple partners to reduce risk and ensure it does not run out of capacity at a critical growth moment. 🤖
What This Move Means for the AI Market
The relationship between Google and Anthropic has a peculiar dynamic worth paying attention to. Despite being direct competitors in the AI model market — Google with Gemini and Anthropic with Claude — Google also serves as an essential infrastructure provider for Anthropic. This kind of relationship, which blends competition with dependency, is increasingly common in the tech industry and creates complex strategic situations for both sides.
For Google, investing in Anthropic even while competing against it makes sense for several reasons. First, it ensures substantial revenue for the Google Cloud division, which gains one of the largest AI infrastructure clients in the world. Second, it keeps Anthropic within the Google ecosystem instead of pushing it into the arms of other cloud providers. And third, it gives Google a privileged view into the technological advances of one of the most innovative companies in the space.
The size of the current commitment changes the nature of this relationship significantly. We are talking about an investment that places Anthropic among the most well-capitalized companies in the tech sector, with a valuation that rivals giants in other segments and reflects market expectations about the value AI models will generate over the coming decades.
Anthropic on the Path to an IPO
Another point worth watching is the company’s valuation trajectory. Anthropic was valued at $350 billion in February, but according to Bloomberg, investors are already willing to put money into the company at a valuation of $800 billion or more. That valuation leap shows the market’s appetite for cutting-edge AI companies and the confidence that Anthropic will keep delivering results in the years ahead.
On top of that, the company is also reportedly considering an initial public offering (IPO) as early as October of this year. If that materializes, it would be one of the most anticipated IPOs in the tech sector and could set new valuation benchmarks for artificial intelligence startups around the world.
The Role of Safety in This Equation
It is worth remembering that Anthropic was founded with a very clear premise: developing AI safely and aligned with human interests. That mission has always been the company’s differentiator in its messaging, but the resources needed to compete with players like OpenAI, Meta, and Google itself are immense.
With this new infusion, Anthropic gains the breathing room to keep investing in both safety research and the development of increasingly capable AI models, without having to choose one at the expense of the other — which was a real tension before this deal.
The situation with Mythos illustrates this challenge well. The model has significant cybersecurity applications, but precisely because it is so powerful, it carries risks that need to be carefully evaluated before an open release. The fact that it was already accessed by unauthorized users shows that balancing innovation and safety remains one of the biggest practical challenges in the AI industry.
What Becomes Clear After This Announcement
When a company the size of Google puts $40 billion into a single bet, it is saying a lot without needing to use words. It is saying it believes AI models will be central to the future of digital business, that Anthropic has something worth protecting and expanding, and that control over the computing power needed to run these systems is just as valuable as any product or service that can be built on top of them. It is a statement of intent disguised as a financial transaction.
For anyone following the sector, this investment also reinforces a trend that has been taking shape for some time: the consolidation of the AI market around a small number of major players with privileged access to computational, financial, and talent resources. Smaller companies that cannot secure this kind of infrastructure will face growing difficulties competing at the large-model level, which could further concentrate power in the hands of a few and make the entry of relevant new competitors increasingly difficult.
The picture taking shape goes well beyond a simple capital injection. It reveals how the battle for AI leadership is increasingly concentrated around who controls the infrastructure, the chips, and the energy to run these systems in the real world.
What is at stake is not just money or technology. It is about defining who will shape the development of artificial intelligence in the coming years, with everything that implies for products, services, regulation, and the impact these systems will have on people’s lives. And Google, by betting so heavily on Anthropic, is making it clear that it wants to be at the center of this story — not just watching from the sidelines while others write the script. 💡
