CommSec said it would lead the Australian retail offer if the IPO proceeds, but Anthropic is not seeking an Australian Securities Exchange listing.
Trajectory secures $45 million to automate specialized AI training
The Sequoia-led Series A values the startup at $300 million as it makes its post-training platform publicly available.

Trajectory announced a $45 million Series A investment led by Sequoia, valuing the AI training startup at $300 million. Nvidia, Bessemer Venture Partners and Bain Capital Ventures also participated in the financing.
Alongside the funding announcement, Trajectory is making its platform publicly available. The company seeks to automate the additional training businesses use to adapt general AI models for work in fields including law and finance.
Customers connect an open-source model and their applications to Trajectory, which gathers performance feedback and uses it for further training. Co-founder Ronak Malde said the system enables specialists, including doctors and lawyers, to improve models themselves with little manual work.
Harvey and Rogo use the software for specialized legal and financial tasks, including contract reviews and company valuations using particular methodologies. Nvidia, Decagon and Mercor are also customers.
Three former Stanford classmates established the startup roughly a year ago, and its workforce now numbers 18. Malde and co-founder Michael Elabd previously held roles at Google DeepMind. The third co-founder, Arjun Karanam, worked at Apple on generative foundation models.
Elabd argued that human participation in training makes the company's approach safer than autonomous self-improvement. Karanam said the interface was designed to help customers without technical expertise carry out post-training.
