AI Profitability: A Comparative Analysis of Anthropic and OpenAI’s Paths to Success

Craig: if you think the ai war is about who has the smartest chatbot, you're misking the bigger picture. it's actually about who's gonna go broke first. today, we are looking at massive shift in 2025 and moving on to 2026 that reveals why the boring ai company might end up owning the future and what that means for your wallet. we are witnessing a divergence between the two ai superpowers or all the ai superpowers. openai is chasing the massive consumer growth but burning cash at a rate, uh, that could lead to a $74,000,000,000 loss by 2028. meanwhile, anthropic is quietly corning the enterprise market and is on track to break even this uh, to breakeven in 2028. so where openai is on track to have a $74,000,000,000 loss in the year of 2028, that's the same year anthropic is projecting to break even. so this headline news because it signals the end of experimental phase of ai and start of the sustainable phase where reliability and business models matter more than hype. i'm always saying nobody's making money in ai except for hardware companies. this is an example. now you were just talking about, um, sam altman declaring a code red, um, the last couple of days. what's going on there?

Jon: yeah. so, apparently, um, he's declared a code red effort, um, in order to improve the quality of chat gpt. uh, so, apparently, other products are gonna be delayed as a result and which makes sense because openai is facing immense pressure from competitors who have narrowed the gap in the ai race. me, personally, i've always thought that claude has much much better quality output and it's much, uh, much better to work with than openai. but google gemini last month, they've surpassed openai's models in an industry bet benchmark tests. and, also, um, openai is facing pressure from anthropic, which is becoming, uh, popular among business customers, which flows kinda nice into this because, um, the difference between anthropic and openai and the reason why anthropic is so much closer is because anthropic focuses on selling ai tools to businesses. so think corporate customers paying premium prices, while openai is betting on, uh, they're betting big on consumer products and massive infrastructure investments hoping to dominate the entire ai landscape. what are your what do you think about all this?

Craig: i i i think it's interesting to watch how they carve out the market. i think it's interesting to watch where they each excel and get better at. and i think, um, you know, when it was six months ago, a year ago when google came out with gemini and we're having all those problems and they were, you know, really bad, people just dismissed them immediately, and maybe we did too. i don't remember if we did or not. uh, we did dismiss that version of it. that's for sure. but you cannot us underestimate, you know, the power of google's data.

Jon: mhmm.

Craig: they have all pretty much crawl every website in the world every freaking day or close enough to it. um, it has all the data, what you do, what you send, all your email, all, like, just everything, all the social media. um, so to to think that it count it out when they've got all the money and all the data, they just have to work at it long enough, um, i think was a mistake. and i think it was arrogant for a lot of these other companies like openai. you know, they've been through a lot of stuff, um, that has not, uh, had them focusing on the product. yeah. um, and, you know, for example, i've i've seen some road maps where they're putting in ads in it, you know, and they had the, uh, board controversy where they fired sam altman last year or at the beginning of this year, and, um, then they brought them back. and and then immediately within a month, uh, half the board or most of the board were gone, the ones that voted against them. yeah. and so, anyway, i mean, weird stuff is going on. um, and it's interesting to see where they're going here. and they're burning money like crazy. they need more money all the time. i think openai might be too big to fail at this time from their relationships with governments and other things, but that is only a matter of time before that changes. but it's interesting to see their different models, what they're going after, what their spend rate, and what happens when, um, you take your eye off the ball.

Jon: yeah. so there's a couple of things here. my question to you is, so a lot of their investments right now is going into infrastructure. so do you think that their investment in the infrastructure, even though they're gonna take longer to be profitable, they will, in the long run, be worth not worth more, but be more profitable because of that investment in infrastructure?

Craig: it depends. yeah. i i mean, the infrastructure is definitely one thing, but it's product. right? and i've mentioned this before, and it's see it it's a meme now or and it's in the, uh, culture, the end certification of products. they get good. they get users, and then slowly over time, like google. like, you search on google. it's just not at all what it used to be. um, it's because they started focused on money. so as openai shifts and start focused on money, um, they changed. and i think that's what happened. they were folk as a nonprofit, they were focused on the product, and now they folk they're focused on profitability. and i those aren't the same things, and it doesn't one doesn't eliminate the other. but, you know, when you have this cash burn that is crazy and they're always have to raise billions and billions and billions of dollars, we'll have to see it. openai commands a staggering 300,000,000,000 valuation following its softbank funding round. um, this makes it one of the most valuable private companies globally. anthropic's valuation stands at 61,500,000,000.0 in march. however, new funding rounds are approaching 170,000,000,000 suggesting investors see tremendous gains. right?

Jon: and, also, we can cover this next time, but anthropic is reportedly preparing for one of the largest ipos ever. uh, it just that just came up on my thing, so we can talk about that next time.

Craig: so there's gonna be, uh, reliability versus features, privacy trade offs, rise of agents. you know, one of the main things i see why we i run flowgrammar events. we get ai builders in toronto. we talk to them every month. it's great. the number one complaint, openai is slow. it's way slower than every other model out there, and people are leaving it for that alone. um, so they need to

Jon: yeah. i'll say this. um, sonnet 4.5 is actually pretty slow. uh, i know i always praise anthropic, but i have to say, sonnet 4.5 is pretty slow. it it gives openai vibes, but the output is phenomenal.

Craig: and that's a in openai's defense, the one thing i do really love is structured output, and not many of the other ai models support structured output. yeah. and most people don't give a crap, but if you're running apis and you do a call and you want a specific, uh, structured json output out, openai is the freaking king of that. okay? because some of them are starting to support it now. i think gemini started supporting it, uh, recently. i'm not sure if, uh, claude and sonnet support it yet. oh, yeah.

Jon: claude has been doing it. yeah. i do it all the time with claude. yeah.

Craig: because when i get it back, i don't get a, uh, structured output back from claude. so i think we may not be talking about the same thing, but, uh, i could be wrong.

Jon: like a structured json prompt output?

Craig: um, all the fields in the exact way and the exact format that i asked for go. not text in the middle of json, improperly formatted json, not everything. you know?

Jon: it very specific output. yeah.

Craig: that's what i mean. so the thing with openai you don't understand is everything comes out proper with their structured output, so you can just grab the field and use it immediately, not run it through a code node or make strip out all the other stuff or what have you. and that's a game changer when you're doing ai and automation, that that structured output. now the other ones are getting better. and and, uh, but if they don't support structured output, it it's not something i usually use too much in my ai workflows, but i'll use it for other things.

Jon: yeah.

Craig: so, um, openai is building the ultimate consumer toy, but anthropic is building the ultimate business machine. and in a recession, the machine that saves the company money usually wins. so in the future or just is this the future or just a cash burning bubble? let me know what you think in the comments below, and we'll keep watching where the money's going.