Cohere at $7B: Funding and AMD Deal Explained  ## Metadata - Author: [[Hard Fork AI]] - Full Title: Cohere at $7B: Funding and AMD Deal Explained - Category: #podcasts - URL: https://share.snipd.com/episode/943b09f7-17da-4c33-8e47-09b312cd9aa9 ## Document Notes ## Highlights - **Cohere's Big Enterprise Raise** - Cohere raised an extra $100M on top of an oversubscribed $500M round, valuing it near $7B. - The company targets enterprise AI and has drawn strong investor demand despite being smaller than top rivals. Transcript: Jaeden Schafer So just on Wednesday, Cohere, which is really an enterprise AI model maker. I think a lot of people, they try to compare them to ChatGPT, but it's not really ChatGPT. They're specifically building AI models for the enterprise. They have raised an additional $100 million, which is really, really impressive. And this brought their valuation up to $7 billion. I'll also say that their round was oversubscribed. They announced it in August, and it was an oversubscribed $500 million round. So it seems like they had a lot of appetite. They raised at a $6.8 billion valuation, oversubscribed at $500 million. And then it seems like they're like, all right, there's still a lot of demand. We'll go and take in another $100 million. So altogether, it was around $600 million that they raised and a little bit, you know, somewhere between $6.8 billion and $7 billion was the valuation for all of that. ([Time 0:01:06](https://share.snipd.com/snip/3af1989f-090a-4723-af34-96fbbdedbf89)) - **AMD Partnership As Strategic Leverage** - Cohere signed a partnership with AMD to run its command-family models on AMD Instinct GPUs. - That hardware alliance mirrors how NVIDIA invests in startups and helps secure strategic funding and market differentiation. Transcript: Jaeden Schafer Cohere also just announced a really interesting partnership. So of course, OpenAI just got a $100 billion investment from NVIDIA. And so, you know, OpenAI is kind of like a big competitor, even though they're not directly competing, because Cohere is more focused on the enterprise. But OpenAI obviously serves enterprise as well. And so, Cohere has actually just signed a deal with AMD, one of NVIDIA's big competitors, and it's also one of their investors. So right now, what the offer is that Cohear has, how they kind of set themselves apart, is that they have what they're calling a command family AI model. So they have command vision, translate reasoning models, and they can all run on AMD's Instinct GPU on an NVIDIA competitor, basically. And so it seems like they're, it's kind of interesting this time where you have these AI models that are not just trying to set themselves apart by, you know, who they're serving, but Who they are using to power their models. And it seems by, you know, making an alliance with AMD, and I'm sure you can still run Cohere on an NVIDIA GPU, but by making it specifically so you can run this on AMD's Instinct GPU, AMD Is happy and they're going to invest money into your company. The same way that it feels like NVIDIA basically gives money to every major AI startup because they know they want these companies to be running their AI models on NVIDIA's GPUs. So Cohere is obviously not abandoning support for NVIDIA's GPUs. They're just adding support for AMD, which not every AI model does that. ([Time 0:02:01](https://share.snipd.com/snip/4b1ef2ab-a784-4798-a69f-f37193625f0d)) - **Founder's Transformers Roots** - Aidan Gomez co-founded Cohere in 2019 and co-authored the Transformers paper "Attention Is All You Need." - His academic pedigree helped position Cohere early in the model race alongside OpenAI and Anthropic. Transcript: Jaeden Schafer I remember talking about them at the beginning. They were kind of up there with OpenAI, Anthropic, you had Cohear, and a lot of people talked about how they sort of fell off. Aiden Gomez is their co-founder. He co-founded it in 2019. And he's also one of the authors on the famous Transformers paper that really is what started this whole big AI race where everyone is, you know, attention is all you need is the name of The paper, but is outlining how to make these LLM models. And he was one of the authors on this specific paper. ([Time 0:03:37](https://share.snipd.com/snip/99675feb-c350-4db1-868e-6a941105bc3a)) - **Why Cohere Feels Smaller Than It Is** - Cohere's growth is substantial but overshadowed by mega-valuations like OpenAI and Anthropic. - Comparing absolute dollars can mislead; Cohere serves a narrower enterprise niche with steady returns. Transcript: Jaeden Schafer What's interesting to me though is obviously like zero to $7 billion valuation in six years. That is absolutely insane in my opinion. And if you had said that like 10 years ago, people would have been like, there's no way. Like that's incredible. That's an amazing company. But I think Cohere definitely gets really overshadowed by just how fast OpenAI is growing and also Anthropic. OpenAI was just recently had a valuation of $500 billion. So obviously this is, you know, really, really massive. They were in talks to sell about $6 billion in their stock, and it was all happening in a $500 billion valuation. So when you have these $500 billion valuations and you're comparing this to $7 billion, people are like, Cohere isn't growing as fast. And even if you compare that to Anthropic, they hit $183 billion valuation in their Series F, and they were actually raising $13 billion there, which is honestly, it's pretty staggering To think, you know, you see Cohere and they just raised 600 million, which, wow, that's a lot of money. But you're comparing this to, you know, in the same month, Anthropic is pulling in 13 billion. Like it's it's just crazy. These numbers are absolutely staggering. So I think that's why people down, like talk down on Cohere, but obviously it's growing in an incredible multiple. ([Time 0:04:13](https://share.snipd.com/snip/1d68366d-dba5-4b1e-b401-15316f51e14d)) - **Enterprise Focus And Data Sovereignty** - Cohere positions itself for enterprise customers needing AI sovereignty and local control of data. - That focus attracts investors like the Business Development Bank of Canada and Nexus Capital Management. Transcript: Jaeden Schafer What I will say, Cohere basically is always focused on the enterprise market. It markets itself as an, basically to enterprises where AI sovereignty is really urgent. So, you know, local control of your data and models, right? You're, you're running a lot of these things yourself rather than putting them into the hands of a foreign entity. And so because of that, they've gotten a lot of interest from a bunch of people. The Business Development Bank of Canada, and by the way, Cohere is a Canadian company, but the Business Development Bank of Canada and Nexus Capital Management were new investors In this $100 million round that they just pulled in. ([Time 0:05:31](https://share.snipd.com/snip/652f9b62-b495-4301-acf5-61ffe8d8f092)) - **Make A Visibility Push** - Cohere could expand by promoting its models to broader consumer and developer markets. - Push visibility and tailor models for specific verticals to capture more usage and adoption. Transcript: Jaeden Schafer They're, you know, getting a lot of value out of it. I, you know, I am curious if, if Cohere could try to address a broader market and if they could get a lot of demand for it. I know for myself on my own platform, AI box.ai, we basically for 20 bucks a month, you get access to all of the top AI models And Cohere is, we have the latest Cohere models on our platform. So like you can embed them into any products. I think Cohere could probably do a lot better job of putting a push and being more visible. It's just not a brand you see out there a lot. But at the end of the day, it's a powerful company. I'll say like, basically, I can see usage metrics of our AI model. Cochure is not incredibly popular on our platform, not a lot of people use it. And so I know they have a lot of specific like use cases where you might be running this locally on your own hardware for different high, you know, high concern industries like legal or Healthcare. But at the end of the day, I think that I think they can still make a really big push to kind of have the general consumer use this. ([Time 0:07:28](https://share.snipd.com/snip/41cf89d8-bb7f-4bba-8752-1ad867b75a78)) - **Win With Vertical Specialization** - Specialized model features (like Anthropic's tone or Cloud Code) create strong product advantages. - Cohere can win by tailoring models and tools to clear vertical needs rather than generic competition. Transcript: Jaeden Schafer And also Anthropic's Cloud is better for coders. They have Cloud Code, where they kind of built the Cloud Code and kind of this tool that developers can plug into their whole data, their whole code base, and it can write code for them. We use Cloud Code exclusively at OpenAI, or at AI Box, because it's just so much better at understanding everything going on, and OpenAI doesn't really have a strong competitor to it. So, and I know there's Codex, but I'll spare you the Cloud Code Codex debate for the moment. In any case, I think Code here could focus on a couple of these areas, and I think they could see a lot of growth. I'll be curious to see what happens with them, but at the end of the day, do not count Cohear out. They're still alive and kicking. ([Time 0:08:56](https://share.snipd.com/snip/9680a855-76b8-46d6-a0ea-d1cd99d9ffa7))