AI Attacks Will Overwhelm Corporate Defenses | Joelle Pineau artwork

AI Attacks Will Overwhelm Corporate Defenses | Joelle Pineau

MTS

October 30, 2026

Cohere Chief AI Officer Joelle Pineau discusses the company's massive Series E fundraise, expanding enterprise deployments, and building frontier AI models outside the US and China.

0:00

Joelle Pineau

Given the volume, the complexity of attacks these days, if companies are not thinking about that, they are going to have a really, really difficult few months in the current strategy. And certainly, there's a lot of governments who are thinking about that with respect to their own infrastructure and making sure that they have the right access to the technology.

But I would encourage every organization, whether it's a company, whether it's a nonprofit, to think about how to secure their own systems.

0:31

SPEAKER_2

And we are back. We are live with Joelle Pineau, who is the Chief AI Officer at Cohere, which just announced the initial close of a huge Series E targeting up to 7 billion Canadian dollars, 5 billion USD at a 35.7 billion Canadian dollar valuation. So congratulations and Joelle, welcome to MTS.

0:54

Joelle Pineau

Thank you, thank you. Great to be here.

0:57

SPEAKER_2

Absolutely. So tell us about the raise. What will this allow you to do?

1:02

Joelle Pineau

Yeah, I mean, you know, Cohere, for those who may not be as familiar with who we are, you know, model building company headquartered in Toronto, really focused on building frontier technology, but focused on the enterprise side, enterprise governments and so on. So we build our own models, we build harnesses that are suitable for those kinds of deployments. And we do a lot of cases where folks want solutions for AI that are deployed on-premise, air gap with really high guarantees of security and privacy. Obviously, you know, a lot of the raises going towards training the next generation of models.

It's big numbers and at the same time, you know, the competition to build frontier technology is ferocious. And between the data centers, the data itself, the talent and so on, it takes a lot to really maintain our capabilities and make sure that we're building the right technology for the enterprise space, which in some cases is actually quite a bit different from what we're seeing on the consumer side.

2:04

SPEAKER_2

Totally. Well, yesterday we had on someone from Mistral, and I asked them this question. I'd love to ask Cohere the same question, which is Cohere is, you know, an incredible company when it comes to enterprise, fine-tuned enterprise deployments. But it is not a super bleeding-edge frontier lab in the way that OpenAI or Anthropic are. So do you think that it would have been possible to become so when Cohere was founded? And do you think that it would be possible to do so today?

2:39

Joelle Pineau

I mean, that's an interesting question, right? Like Cohere was founded in 2019 If you look at any of these companies in 2019, I think the trajectory of all of them was really quite unpredictable. There's a lot of things that happened along the way that we didn't expect at that time. What it was founded on, though, is that conviction that the technologies was going to be really transformational. Aiden Gomez, our CEO, one of the authors on the Transformer paper, and did some of that foundational work while he was at Google. You know, like a strong conviction from the co-founders, from the initial investors, that this was going to be a major technology shift. They also made some choices of focusing on enterprise versus customer, which maybe gives the work less visibility than others.

And headquartered in Canada, which maybe has some impact in terms of the early funding trajectory. I would say right now, I think we're in a really good position to continue where we are, specifically with respect to the competitiveness of frontier models. We watch all the same indexes as others in terms of monitoring that.

What we really aim for is to be in the ballpark of having frontier model. Whether you fine-tuned such that you're really suited to the latest results or not, it becomes less important. And we saw some shuffle recently, for example, when some of the artificial analysis index sort of shuffled away that we're doing these index and then suddenly like the ranking changed and so on. We try not to over-optimize on that and to produce models that are really high value for money. And yet are within that margin of the frontier. A lot of our customers actually don't want to run frontier models. They're way too expensive. They're not necessary for the workloads that they have. And so we offer like a mix. We're model agnostic in our platform. So we offer a mix of different solutions depending on the needs and try to get the best trade-off they have from like a security point of view, cost-effectiveness, and then capabilities. Not all models are just good if you're doing coding, obviously, like some models are much better than others. But if you're really into chit-chat and conversation, other models might be better. And so we try to pick our lanes where we really bring value and keep growing company.

Read the other 9 minutes of this episode here.

$15 gets you this episode and 99 more. Credits never expire. After checkout, one link brings you back here to read it.

Not what you expected? Full refund within 14 days.

Thousands of transcripts fetched by people building searchable podcast archives

Fetch transcripts with code

The same credits work from a script or an AI agent. The demo key returns a sample episode in full, no card needed:

request
curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.

Using your own key:

request
curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/1000798964530