Cohere opens a record Series E, and Ottawa takes a stake

Covers Thursday, Oct 29 through early Friday, Oct 30 (Eastern). Thursday's biggest check came with conditions attached. Cohere opened a Series E aimed at up to C$7 billion, and Canada's government will buy equity in exchange for promises on jobs, compute and headquarters. Sequoia priced a ServiceNow challenger at $3B, and IVP made a consumer AI agent a unicorn. Fourteen rounds in the window cleared our checks, plus one Oct 28 evening round the last Daily Brief missed (marked "catch-up"). Every round below is confirmed by the company, its lead investor or both. Figures companies report about themselves are labeled company-reported.

Cohere, the Toronto-based maker of AI models and agent software for regulated industries and governments, announced the initial close of a Series E that targets up to C$7 billion. If the round fills, it would value the company at C$35.7 billion (Cohere release via PR Newswire · Cohere blog · CBC / Reuters). Cohere has not said how much is committed so far. The companies of Schwarz Group lead the round. Bell Ventures, Bullhound Capital and the Government of Canada are new investors, and Inovia Capital, NVIDIA and Prosperity7 Ventures are returning. Schwarz first said it would lead, with a €500M (about $600M) structured financing commitment, in April, when Cohere and Aleph Alpha announced plans to combine (Cohere, Apr 24). Prime Minister Mark Carney's office says Ottawa will invest up to C$1.6 billion in equity, paid in stages with regular progress reports (Prime Minister of Canada). In return, Cohere commits to at least 1,000 new Canadian technical jobs and at least C$1.7 billion of added computing infrastructure in Canada, both by 2030, and to keeping its Canadian incorporation and headquarters. Cohere says annual recurring revenue grew 165% year over year as of the end of the third quarter and its customer count rose 20x in a year to 300 (company-reported). It calls the round the largest private financing ever by a Canadian technology company. Why it matters for investors: Two of the anchors in this cap table have goals beyond financial return. Schwarz wants Cohere's models running on its STACKIT cloud, and Ottawa wants jobs and data centers in Canada. That may make them patient holders. But it also means a late-stage investor is buying alongside a lead whose check is a structured financing commitment with undisclosed terms, and a government whose money comes with operating promises. The price is also provisional: C$35.7 billion is the valuation the round implies only if it is fully raised. Diligence questions: How much was committed at the initial close, and how much of it is Schwarz's structured piece rather than priced equity? Does the pledge to keep Cohere's incorporation and headquarters in Canada narrow the list of future acquirers? What dollar base sits under the 165% ARR growth?

Historical lens. Britain ran a state-backed chip venture with tranches and a negotiated exit four decades ago. In a Commons debate on March 12, 1980, a Department of Industry minister said the original Inmos plan envisaged £50 million of equity from the government-funded National Enterprise Board. The previous government approved £25 million first, with a review at the halfway stage. He added that Inmos was the only British-controlled company set up to make standard integrated circuits in high volume in the UK (Hansard, 1980). On July 12, 1984, Trade and Industry Secretary Norman Tebbit told MPs that Thorn EMI had agreed in principle to pay £95 million for the board's holding, which he put at 76% of the equity. He set that against £65 million of total government investment plus guarantees, and said the government had not sought a golden share (Hansard, 1984). Our interpretation, not established fact: Ottawa's staged payments echo the Inmos tranches. Canada differs in writing its conditions on jobs, compute and headquarters into the investment itself. Inmos suggests the question for co-investors is the exit: who is allowed to buy a state-backed champion, and on what terms.

Serval, a San Francisco company whose AI agents build and run automations for IT, HR, finance and other internal service teams, raised a $150M Series C led by Sequoia at a $3B valuation (Serval release via VC Access Online · Business Wire · Fortune). Existing investors Elad Gil, Evantic, First Round, General Catalyst and Redpoint returned, and Conviction and Headline are new. Total funding is $277M. This is its third round in about 12 months: the Series A was announced in October 2025 and a $75M Series B followed in December 2025. Serval says ARR has doubled every quarter since the Series B and that it signed two Fortune 20 customers in recent months (company-reported). Named customers include Fox Corporation, WPP, Live Nation and Ramp. The company names ServiceNow as its main competitor and says one automaker moved its IT service management, asset management and configuration database off ServiceNow in 45 days (company-reported). It is also starting ProServal, a services arm that will hire more than 100 automation engineers in the next year, roughly half of them recent graduates, to work inside customers. Why it matters for investors: ServiceNow-style systems of record are hard to dislodge because of the switching work involved, not the software alone. Serval is attacking that cost directly with agents that write the automations and, now, with its own embedded engineers. The services arm shows how the contest will be won, and it will also weigh on margins. A company hiring 500+ people, many of them on-site with customers, will not look like pure software for a while. Diligence questions: What is the ARR base behind the claim that ARR doubles every quarter? Are the Fortune 20 deployments full replacements or pilots alongside the incumbent? What gross margin does the business show once ProServal is fully staffed?

Wabi, the personal AI assistant founded by Replika creator Eugenia Kuyda, raised $55M led by IVP at a $1.25B valuation (Wabi release via PR Newswire · TechCrunch). Andreessen Horowitz, Forerunner and the Amazon Alexa Fund joined, along with angels including Tobi Lütke, Marc Benioff, Marissa Mayer, Dick Costolo and Alex Pall. The company does not name a stage. Wabi pairs an AI agent with small custom apps that it builds on request, such as a shared meal plan that becomes a grocery list, and with spaces where couples, families or friend groups use them together. It runs on iOS, Android and Mac, with Windows to follow. In the release, a16z general partner Anish Acharya says the firm has backed Kuyda since the pre-seed. The release does not describe pricing or revenue. Why it matters for investors: Most consumer assistants are built for one person in one chat window, and they compete with the default assistants on phones and in the large chat apps. Wabi's bet is that shared use is the difference: if a household or friend group runs its plans through Wabi, every member who joins makes it harder to leave. At $1.25B with no disclosed business model, investors are paying for a founder's record and an early usage curve. Diligence questions: How many active groups does Wabi have, as opposed to single users, and how do they retain? What does inference cost per active user, and who pays it? What stops a large platform from copying shared spaces inside an app people already use?

Cohere is ranked by its stated target. The amount committed at the initial close has not been disclosed.

Round notes

Not investment advice. The AI Cap Table is published for information only and is not investment, legal, or tax advice. Nothing in this newsletter is an offer or solicitation to buy or sell any security. Funding data comes from public sources we cite and verify, but errors are possible; do your own due diligence and consult a qualified adviser before making any investment decision.