Anthropic plans two-way voice mode and memory for Claude

If you have been waiting to talk to Claude like a real person, Anthropic is finally getting ready.

Speaking to The Wall Street Journal at Davos, Anthropic CEO Dario Amodei revealed that the company plans to roll out a "two-way" voice mode for Claude, alongside a new memory feature to remember past conversations and user details. Amodei also noted that smarter AI models are slated to arrive in the coming months.

Catching up on voice and compute

The bigger confession, though, was about infrastructure. Amodei admitted that skyrocketing demand over the past year—particularly the last three months—has completely overwhelmed Anthropic's ability to supply enough compute.

Anthropic is burning massive capital to keep pace with OpenAI. Despite raising $13.7 billion to date, the startup reportedly lost billions of dollars last year. It is now said to be in talks to raise another ~$2 billion at a $60 billion valuation.

Why it matters

Claude has earned deep loyalty from developers and writers, but its lack of memory and hands-free voice mode made it feel like a text-only tool next to ChatGPT. Adding these features is essential if Anthropic wants Claude to become an everyday companion rather than just a smart workspace.

However, we think Anthropic's biggest hurdle right now is raw capacity and burn rate. Voice interaction is computationally heavy. Launching real-time audio while existing servers are already overwhelmed, all while burning billions, shows just how expensive staying in the frontier AI race has become.