←News detail

ZhidxUpdated Oct 31 01:55

Cohere Launches Series E Funding, Targeting Up to C$7 Billion, with the Canadian Government Planning to Take a Stake

AI summary

Canadian large-model unicorn Cohere announced the completion of the first close of its Series E funding. The round targets raising up to C$7 billion (approximately RMB 33.4 billion), led by German retail giant Schwarz Group, with existing investors including Nvidia and Saudi Aramco's Prosperity7 Ventures continuing to participate, while the Canadian government and others joined as first-time investors. Cohere said that if it reaches its fundraising target, its valuation will reach C$35.7 billion (approximately RMB 170.2 billion), setting a record for the largest single private financing round in the history of Canadian startups. The company plans to carry out two AI company acquisitions within a year, launch the enterprise cybersecurity product Cohere Cyber, and train a new generation of frontier large models; the Canadian government plans to invest up to C$1.6 billion in stages for an equity stake, on the condition that Cohere add at least 1,000 Canadian technology jobs by 2030.

Body · AI translation

Smart Things (Zhidongxi) October 30 news: Last night, Canadian large-model unicorn Cohere announced the first closing of its Series E financing, with this round targeting up to C$7 billion (about RMB 33.4 billion). Cohere said on its official X account that if it reaches its fundraising target, this financing will set the record for the largest single private financing round in the history of Canadian startups.

According to Cohere, it plans to advance two AI company acquisitions within a year, launch Cohere Cyber, a product for enterprise cybersecurity, and train a new generation of frontier large models.

This round was led by German retail giant Schwarz Group, with existing shareholders including Nvidia and Prosperity7 Ventures under Saudi Aramco continuing to participate, while the Canadian government and others invested for the first time. In September last year, the company was valued at about C$9.8 billion (about RMB 46.6 billion). If it reaches its fundraising target, Cohere's valuation will reach C$35.7 billion (about RMB 170.2 billion), growing to 3.6 times its original value within a year.

Cohere was founded in 2019 and mainly provides enterprises and governments with privately deployable large models, AI agents and search tools.

Its three co-founders are all University of Toronto alumni. Among them, co-founder and CEO Aidan Gomez is one of the eight authors of the foundational Transformer paper Attention Is All You Need. Another co-founder, Nick Frosst, studied under "godfather of deep learning" and 2024 Nobel Prize in Physics laureate Geoffrey Hinton.

Previously, Cohere had completed multiple large financing rounds: a Series C of about C$380 million in 2023, a Series D of about C$700 million in 2024, and another round of about C$840 million in 2025, with participants including Nvidia, Oracle and Salesforce Ventures.

According to Cohere, by the end of the third quarter this year, its ARR had grown 165% year on year. Its number of global customers grew 20-fold over the past year to 300, including Royal Bank of Canada, Fujitsu and PwC.

1. Self-developed models integrated into enterprise customer service, problem resolution time shortened to 2-5 days

Cohere mainly develops models for enterprise document analysis, reasoning and AI agent tasks. In May this year, the company open-sourced the Command A+ large language model; its quantized version requires only two Nvidia H100 GPUs to run and supports enterprise self-deployment.

To let AI participate in enterprises' daily work, Cohere also launched the agent platform North. The platform can connect to enterprises' internal knowledge bases and office software, allowing employees to create AI agents through natural language to complete tasks such as information queries and customer service. On October 5 this year, Cohere launched North 2, adding features such as cross-session memory and shared knowledge bases, enabling AI to remember previous work information and continuously execute multiple tasks.

According to a customer case published by Cohere in February this year, US AI cloud provider CoreWeave took only 90 days to integrate the previous version of North into its customer service system, using AI to collect customer information, find technical documents and help transfer issues to engineers. Combined with internal process optimization, CoreWeave's average problem resolution time was shortened from 4-8 days to 2-5 days.

In response to enterprises' data security needs when using AI, Cohere provides private deployment solutions and has partnered with Nvidia and Microsoft to launch confidential computing services. In October this year, the company launched the public beta of a confidential computing service based on Nvidia H100 GPUs and Microsoft Azure, using hardware encryption technology to protect sensitive data processed by AI models in the cloud.

2. Plans to acquire two AI companies; Canadian government to invest up to RMB 7.6 billion

Cohere plans to advance two AI company acquisitions in the coming year. On October 15, the company signed a letter of intent to acquire US AI cybersecurity startup AISLE, planning to use its technology to launch Cohere Cyber to help enterprises automatically discover, verify and fix software vulnerabilities.

According to AISLE's official disclosure, over the past year the company has discovered more than 500 software vulnerabilities assigned CVE numbers, including 18 of the 24 vulnerabilities recently fixed in the open-source networking tool curl. Cohere plans to combine AISLE's technology with its own large models so that enterprises can complete vulnerability screening on local servers or isolated networks, avoiding the external transmission of sensitive code.

▲AISLE AI cybersecurity platform workflow, covering software vulnerability analysis, risk screening, remediation and verification (Source: AISLE)

The other transaction targets the European market. In September this year, Cohere signed a merger agreement with German AI company Aleph Alpha, planning to integrate both sides' model R&D capabilities and European government and enterprise customer resources. Schwarz Group, the lead investor in this round, also plans to deploy Cohere models on its STACKIT cloud platform to provide AI services to European customers. At present, neither transaction has been completed.

In addition to acquisitions, Cohere will also receive funding support from the Canadian government. According to an agreement announced by the Canadian Prime Minister's Office on October 29, the government plans to invest up to C$1.6 billion (about RMB 7.6 billion) in Cohere in stages for an equity stake. As a condition of the investment, Cohere has committed to adding at least 1,000 Canadian technology jobs by 2030 and investing at least C$1.7 billion (about RMB 8.1 billion) to build local computing facilities, while retaining its Canadian headquarters and place of registration.

This financing will support Cohere in continuing to develop large models and expanding into cybersecurity and the European enterprise market through acquisitions. Together with local computing infrastructure construction and cloud platform cooperation, Cohere is extending further from model R&D to the deployment and delivery of enterprise AI products.

Competition in the enterprise AI market is no longer limited to the models themselves. This month, another Nvidia-backed AI office company, Pigment, completed a US$250 million financing round and disclosed that it aims to compete for enterprise customers of vendors such as Microsoft, Oracle and IBM.

Although the two companies' product roadmaps differ, both are trying to bring AI into enterprises' existing business systems to take on real work. For Cohere, what it must compete for next is not only model invocation volume, but also enterprises' software procurement budgets in areas such as office work, operations and maintenance, and cybersecurity.

Sources: Reuters, Cohere

Source

Zhidx