Google Plans Up to $40B Investment in Anthropic for Cash and Compute
Google will put up to $40 billion into Anthropic to cover cash needs and computing resources, according to Bloomberg. The Alphabet unit promises $10 billion right away. This values Anthropic at $350 billion. Another $30 billion follows if the AI company meets specific performance goals, Anthropic states.
Mythos Model Sparks New Compute Demands
Anthropic just shared its newest model, Mythos, with a small set of partners this month. The firm calls Mythos its strongest model yet. It offers key uses in cybersecurity. Concerns over possible misuse led Anthropic to limit access. The company works with chosen groups to check risks and fix them. Still, the model has reached unauthorized users. Running it widely would cost a lot.
Anthropic, founded in 2021 by former OpenAI executives including CEO Dario Amodei, focuses on building reliable AI systems. Its Claude models have gained popularity for safety features. Mythos builds on that line, targeting secure applications where risks demand tight controls.
Compute Becomes Central in AI Competition
Control over computing power shapes the AI field more each day. Training and running advanced models requires massive resources. OpenAI pushes hard to lock in supply. It signs deals worth hundreds of billions with cloud firms, chip makers, and power sources. This month saw OpenAI expand ties with Cerebras, a chip producer.
OpenAI, started in 2015 by Sam Altman and others, leads with models like GPT series. Its moves highlight how top AI players chase hardware to stay ahead. Energy and chips limit growth for all firms in this space.
Anthropic Secures Multiple Deals Amid User Limits
Anthropic deals with user gripes over limits on Claude access lately. It fights back with infrastructure pacts. This month, it agreed with CoreWeave, a cloud provider, for data center space. Amazon added $5 billion this week. That fits a larger plan where Anthropic could spend up to $100 billion for about 5 gigawatts of compute over time.
Amazon Web Services, or AWS, competes in cloud services. This investment deepens Anthropic's options beyond one supplier.
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
Google Balances Competition and Supply Role
Google builds rival AI models, yet supplies key tools to Anthropic. The AI firm counts on Google Cloud for chips and setup. It taps Google's tensor processing units, or TPUs. These handle AI tasks well. Many see TPUs as top rivals to Nvidia's busy chips.
Google Cloud Platform launched publicly in 2011. It trails AWS and Microsoft Azure but grows in AI strengths. TPUs, introduced around 2016, suit machine learning workloads.
Building on Past Ties with Google and Broadcom
Anthropic's link with Google goes back further. Earlier this month, it teamed with Google and Broadcom. Broadcom makes custom AI chips for Google. The group aims to deliver multiple gigawatts of TPU compute from 2027. A Broadcom filing lists 3.5 gigawatts.
Broadcom, a semiconductor giant, partners across tech. This setup gives Anthropic steady access to specialized power.
Expanded Capacity from Latest Deal
The fresh Google funds grow that setup. Google Cloud adds 5 gigawatts over five years. It can expand more later.
High Valuation and IPO Talks
Anthropic hit $350 billion valuation in February. Investors now push for $800 billion or higher, Bloomberg notes. The company eyes an initial public offering as early as October.
Alphabet, Google's parent restructured in 2015, invests across AI to match rivals. Such deals secure influence in fast-moving tech.
Related on Neura Market