Enterprises will eventually move to open-source post-training but more slowly than people think, because fine-tuning requires gathering data, building custom evals and benchmarks; open source becomes strictly better once a use case is solidified and in production at scale.
Jesse predicts enterprises will adopt open-source fine-tuning but slower than expected due to the difficulty of data, custom evals, and governance. He sees a pattern: new/experimental use cases start on frontier models (which is why open-source inference share is falling), but once a use case is solidified in production, there's strong economic incentive to switch. ✦ AI generated
Jesse · a16z Podcast · 2026-07-31 · original ↗
starts at this moment · 7:19
“do you think that enterprises will get there as well on post training open source models... what's the timeline”
I think they'll get there but it'll probably take longer than people think because... even with our team fine-tuning these models is non-trivial... You have to get the data. And then more importantly, you have to like have good eval... it's strictly better to use open source models because when your use case is solidified and you're in production at scale and you're pretty sure this is the the sort of shape of the agent then there's no reason not to use open source because you you get these latency benefits and at the same time you get the cost benefits
verbatim transcript · starts at 7:19
7:19but it'll probably take longer than people think because you know even with our team fine-tuning these models is non-trivial it's not just like oh you And it's like all right, we made the decision to use open source. Like let's just use open source. Like you have to get the data. And then more importantly, you have to like have good eval. Um and if you think about our evals, right, our
7:36evals are very specific to us. You can't just like use some public eval set and like that that just does the job. It's like we're testing it on our task and so we have to generate our own benchmarks and evals. Uh but I think the point is that um at a certain point it's strictly better to use open source models because when your use case is solidified and
7:58you're in production at scale and you're pretty sure this is the the sort of shape of the agent then there's no reason not to use open source because you you get these latency benefits and at the same time you get the cost benefits. I mean again we didn't do these for cost benefits but that's like a nice side effect right >> and once you're there it's like why why
8:14use Frontier for that? But for everything that's new and sort of experimental or you're as Asha was saying what or like kind of these products where you really need the intelligence you're still going to use frontier models and it's just so much easier to use frontier models you know they're worrying about the infra you just like like they just um you just use the APIs and I think that's why in
8:34enterprises right now even though there's there's a lot of hype for open source the sort of share of of open source inference is actually going down right now because people are spinning up all these new use cases and if you're spinning up new use caseas is of course you're going to use the frontier models until until they're working. >> Yeah. >> But of those use cases, you know, some
8:50might die off, but like some might like the enterprises are like, "Okay, great. We want to keep shipping this and like roll it out." >> Once it's at that point, they're heavily incentivized to use open source. It's way cheaper and faster. >> At that point, they'll maybe they can do it in house or maybe they'll need help uh from people to to help them fine-tune it. But that that will eventually
9:08happen. I just think it'll be kind of slow. Even right now in our experience >> like enterprises have a lot of desire to move but they can only do so many use cases at once. You know they there is inertia there and they have to you have to go through all the you know model risk governance and all the security things and so um I think it'll take time
9:25but it will get there. >> Yeah. the the the other reason I think it makes a lot of sense for uh enterprises to kind of build their you know cool bundle of labs is that the shape of these models is changing constantly right we don't just build our set of open source models and then you know it's done we can move on to our next thing and maybe we'll revisit this