OpenAI demand sinks on secondary market as Anthropic runs hot

4/1/2026 at 3:09:46 PM

This, to me, reads like the CEOs that are catching up to months ago.

I was subbed to claude for CC since July/aug (can’t remember when I stopped paying the API pricing).

I canceled 3 weeks ago. Since then everything with the usage limits being slashed before being announced, their cache bug that eats limits (and won’t reset), and just rolling paper cuts with every single release I feel pretty good.

I personally think Anthropic is lost and are chucking products out left and right. They might vibe their moment away

Edit: their last month of releases for CC are heavily favored towards building openclaw into their ecosystem

by Robdel12

4/1/2026 at 3:16:07 PM

Most people I talk to who pay for their own subscriptions have switched providers at least once, or they keep 2 or 3 plans active from different companies so they can try them all out.

It’s interesting how the long business purchasing cycles and have made it difficult for companies to keep up with these rapid changes. Companies like to spend months getting a vendor approved and then not change it for years. Depending on when your company got on the AI coding bandwagon you may be locked into Copilot or ChatGPT while the rest of the coding world knows that Opus 4.6 is king, at least for this month. The situation may change again next month, but today that’s how it goes.

by Aurornis

4/1/2026 at 3:56:10 PM

Indeed .. my company got on Cursor when Cursor's fame started to fade. We've just got out of Cursor now to go on Claude, and I feel like we are again "buying the top"

by WinstonSmith84

4/1/2026 at 3:35:50 PM

I dropped Claude Code months ago. I'm just an amateur who plays around with bringing screens in from Figma and turning them into websites so I can get practice with AI tooling, but their limits were absurdly low on the $20 plan. Codex, for now, is much better in terms of daily limits.

by GenerWork

4/1/2026 at 3:52:16 PM

The amount of tokens you get on Codex for $20/mo compared to Claude Code is indeed insane.

by strongpigeon

4/1/2026 at 4:08:48 PM

And probably unsustainable. OpenAI desperately needs to catch up so they’ll throw yet more cash at it, while Anthropic are market leaders in this particular space.

by afavour

4/1/2026 at 4:16:23 PM

It probably is unsustainable, but until they change it I'm sticking with Codex.

by GenerWork

4/1/2026 at 8:24:49 PM

I suspect GPT-5 models are sparser and/or smaller than Opus which is why they can afford to give away so much usage.

by BoorishBears

4/1/2026 at 7:14:54 PM

[flagged]

by mountlatmus

4/1/2026 at 7:06:11 PM

I sound like a broken recorded in these threads but the $20 dollar Google AI Pro is unbelievable good value. You get more Claude tokens in antigravity than you get with the pro subscription from anthropic AND you can share it with your family for free and they also get the same amount of tokens. That's 5x the tokens for just $20.

by BoredPositron

4/1/2026 at 3:04:15 PM

I don’t know why anyone would want to invest in OpenAI on the open market. They are wildly over priced, very much in the red, don’t have the momentum at this time. Their pitch is “trust us even though we are losing money every quarter we are building the user moat”. That is not their strength at this point. Llms have shown they are $ hungry and only enterprises have a proper use case for them. As evident from anthropic, even the $20 per month is not enough to sustain the token usage as they put limits. So OpenAI is far away in that race and their enterprise adoption is just chat, which is barely useful. So not sure what their pitch or near term target should be. Oh don’t forget they are also pushing on hardware and robots which are also huge cash sinks

by yalogin

4/1/2026 at 3:07:47 PM

Yes, most of their top talent has left, except for Jakub. The top researchers I know have no interest in the company.

by harmonic18374

4/1/2026 at 4:47:01 PM

Yeah, and Jakub didn't seem to have much background in AI research. I'm sure he's a great coder and did a PhD on fast algorithms but it's a different area to pushing forward AI really.

by tim333

4/1/2026 at 3:18:26 PM

Open market investing in private markets is kind of a dumpster fire everywhere. The offering prices are way too high, the contorted investment vehicles can skim a lot of your investment, and the platforms may not even be able to get the shares you bought because the company exercises right of first refusal.

It’s mostly a FOMO play for people who think they need to have some exposure to these companies they see all over the news.

by Aurornis

4/1/2026 at 4:11:53 PM

> I don’t know why anyone would want to invest in OpenAI on the open market

I can only assume hype. That’s why Sam Altman has the job he has. You don’t see the CEO of Anthropic going on the Tonight Show. He’s there to bring OpenAI to the forefront of people’s minds, and uninformed investors will follow.

All catches up to you eventually though.

by afavour

4/1/2026 at 6:25:22 PM

Hasn't seemed to catch up with Tesla, which is still highly valued despite making a pretty mediocre car compared to the competition. Even if one makes the argument that Tesla vehicles are of good quality, it's still a high valuation that seems to show no sign of dropping.

by doom2

4/1/2026 at 4:25:58 PM

I generally think the frontier model labs are doomed as businesses. There's just no economic case for them, especially when a few months later open source models catch up.

by vrganj

4/1/2026 at 4:57:44 PM

Yup. It's not trivial now to setup a system where you can get a frontier model to help do the research, draft a spec, humans read and comment on the spec, and then you get an open model to do the grunt work.

If the spec is very detailed, you've solved most of the problems you might encounter with open models.

You can then get a frontier model to then do a review against the spec.

Doing this cuts down frontier usage by a lot, as all the real work is local, tool calls are instant . It just feels nicer.

I think this is why you're seeing frontier models like Claude suddenly ban people using opencode/pi etc with a subscription (API users still good).

by intothemild

4/1/2026 at 4:30:12 PM

I they tread very carefully, I do think their free+ads strategy has a huge potential payoff.

by strongpigeon

4/1/2026 at 4:38:27 PM

How could this possibly work? Google is profitable because they can insert ~4 ads into a search query. An LLM query costs about 2 orders of magnitude more resources to run than a Google search, so.. I'm not seeing it unless OAI can figure out how to shoehorn 400 ads per prompt into the interface somehow.

by parliament32

4/1/2026 at 4:52:49 PM

Longer sessions. You also have more potential for brand advertising compared to Google Search (which is mostly conversion ads). That being said, it's risky and can ruin your product if done wrong, but I think there are ways to do it right.

by strongpigeon

4/1/2026 at 5:19:47 PM

But longer sessions means more queries, which means more costs, which makes the problem even worse, right?

by parliament32

4/1/2026 at 5:33:38 PM

Only if the margins are negative. Yes, the cost to serve is most likely higher than Google's SRP, but I think the ads can be even better targeted and potentially have a higher CPM than Goog's.

What I'm saying is that I believe their ARPU could be higher than Google's, while what I think you're saying is that their cost will also be higher. I agree with that, but where we differ is that I think that while the margin will be lower, there is still potential to make a ton of money there.

by strongpigeon

4/2/2026 at 3:16:04 AM

They have a dataset of your deepest thoughts and questions AND a way to ask you stuff. Its literally the most valuable dataset on the planet

by whattheheckheck

4/1/2026 at 4:59:17 PM

Google now ads an LLM result to about half the search results. I think they've figured out how to do that without too many resources.

by tim333

4/1/2026 at 5:21:03 PM

Google doesn't have to fight context windows. They can cache and store an AI response to a Google query without having to worry about much other than locale etc. You can't do that a dozen messages into an LLM conversation.

by parliament32

4/1/2026 at 3:33:41 PM

> The large gap between OpenAI’s $852 billion valuation and Anthropic’s $380 billion has investors rushing to grab equity in the latter before it rises, according to Augment co-founder Adam Crawley.

That's the only thing you need to get from this article. They're doing mostly the same thing, aiming the same market. But Anthropic's shares are at 50% off discount.

by raincole

4/1/2026 at 7:39:54 PM

Due-diligence once again proving to be mostly cargo-culting.

by tren_hard

4/1/2026 at 2:56:53 PM

https://archive.is/PMC5t

by Liftyee

4/1/2026 at 3:12:01 PM

I think the main point is that these two companies have been thought of as equals. Yet OpenAI has a 850B valuation while Anthropic has a 380B valuation. So that means either OAI is overvalued, or Anthropic is undervalued. If you believe they are roughly equal, then it makes sense to have demand for Anthropic.

by donkeyboy

4/1/2026 at 3:17:09 PM

Your logic is flawed, comparing them doesn't mean anything about their absolute valuation, they could be both overvalued or both undervalued too

by TheRoque

4/1/2026 at 4:11:31 PM

But if you are going to put some money in one of those companies, you would pick the cheaper one that has more of a moat.

by marcosdumay

4/1/2026 at 4:24:20 PM

Maybe both are overvalued.

by surgical_fire

4/1/2026 at 3:42:27 PM

Both are overvalued.

It's just that people are rotating the musical chairs from OpenAI to Anthropic because the former is cheaper than the other.

They (investors) will do this until there is no more upside left to extract.

by rvz

4/1/2026 at 3:22:49 PM

I just switch between Gemini, Codex, Claude, Z.AI, ... whichever offers the best value.

by throwaway2027

4/1/2026 at 7:15:29 PM

How does that work? Is there a VSCode extension that works with all of them? I’ve only used the Claude Code extension for VSCode and would prefer something like that.

by csiegert

4/1/2026 at 3:51:09 PM

Man, for some reason the ass kissing of Claude is just insane. Claude this, Claude that. I do Elixir for a living and Claude is just absolutely garbage in comparison to Gemini Pro. But hey, everyone hates Google - despite them being the only model provider that is competent and the one that explicitly states for paid providers that they do not use any of our data for training right under the chatbox.

And Anthropic isn't even the saint that everyone pitches them out to be:

https://www.cnbc.com/2026/02/12/anthropic-gives-20-million-t...

by neya

4/1/2026 at 4:10:10 PM

Gemini is great, but when I last tried it it wasn't as good as Claude Code with agentic workflows. When I tried Antigravity is was very unreliable, like the tooling had yet to catch up to the model such that it wasn't able to fully leverage the model's intelligence and capabilities. I think it comes down to how you're using the models, so I'll ask: how are you interfacing with the LLM?

by TheAceOfHearts

4/1/2026 at 3:08:05 PM

So a buy order for $122000M, and a sell order for $600M is presented as bad news for OpenAI?

by vlaaad

4/1/2026 at 4:19:55 PM

“First of all, we’re doing well more revenue than that. Second of all, Brad, if you want to sell your shares, I’ll find you a buyer,” Altman said, prompting laughs from Nadella. “I just — enough. I think there are a lot of people who would love to buy OpenAI shares.”

by therobots927

4/1/2026 at 4:27:12 PM

[dead]

by jamiebeach

4/1/2026 at 2:44:16 PM

OpenAI increasingly looking like the Kozmo.com of the AI bubble.

by ForHackernews

4/1/2026 at 2:55:08 PM

Further proof that the "first-mover advantage" is dramatically overstated

by bensyverson

4/1/2026 at 3:09:47 PM

It could alternatively be proof that Sam Altman's incompetence is dramatically understated.

by philipov

4/1/2026 at 3:19:58 PM

Have never seen an interview or piece of writing where he comes off sounding like he knows what he's doing.

by blehn

4/1/2026 at 2:49:06 PM

That's what happens when you're so untrustworthy as Sam Altman. This is coming to roost now. Ever since Paul Graham even said he'll get whatever he wants by any means necessary, basically. And then his board turns on him.

He's a constant liar.

by johnwheeler

4/1/2026 at 3:00:02 PM

I wish being untrustworthy and being unsuccessful were linked in this way, but sadly they are not.

Altman's untrustworthiness goes all the way back to Loopt and Paul Graham's comments about him were meant as praise, not criticism.

OpenAI's problem is that in addition to Altman being untrustworthy he has also just been flailing in terms of focus.

by bayarearefugee

4/1/2026 at 3:13:07 PM

> ...just been flailing...

Brains fried post late-00s to 10s endless ZIRP waterfall helping them fail up. Not used to reality where everything they touch does not turn to gold.

by yabutlivnWoods

4/1/2026 at 2:53:51 PM

Something I don't quite understand is why software developers are flocking to CC - product of a company whose CEO literally said they will be out of job because of it.

by baal80spam

4/1/2026 at 3:07:30 PM

I am a software developer that believes LLMs will ruin the market for us in terms of destroying a lot (not all, but more than enough to be devestating) jobs and depressing wages, but I still have no real option but to use them to remain relevant for now.

That said, I dont have a horse in the race and have used all the available options and find it very easy to switch among them, so I'm not a specific booster of Claude Code or any other option.

by bayarearefugee

4/1/2026 at 3:05:25 PM

Because whether I use it or not doesn’t affect the fact that it’s going to take my job. It’s still the easiest way to get things done today.

by sd9

4/1/2026 at 2:56:47 PM

ostrich, meet sand. I mean, have you tried CC? It's fun to build stuff with it. I think developers that don't use it (or something like it) will be out of a job, yes (unless they're very niche).

by lacunary

4/1/2026 at 3:13:40 PM

Like Jensen Huang said the job of an engineer is to solve problems not writing code. The code is a means to an end. This is certainly true for me as an engineer and why I'm not worried about AI.

by verst

4/1/2026 at 2:59:30 PM

That's like asking why people who shovel shit all day started using backhoes to do their shit shoveling

by IncreasePosts

4/1/2026 at 2:55:34 PM

If he's right they'll be out of job regardless whether it's because of Claude or another AI.

by chpatrick

4/1/2026 at 3:52:56 PM

Because fear sells easily to those who do not know the future.

Dario constantly fearmongers to them and 98% of software developers all fall for it, and he needs to sell you access to his product.

The future is local LLMs and Dario knows this is their main threat. Just as Cursor, they don't want to pay for Claude anymore and are going with local models.

by rvz

4/1/2026 at 3:08:40 PM

It’s just because ChatGPT is worse than Claude as of today. If that flipped, which is probably will at some point, if only temporarily, OpenAI will be back on top. Unfortunately it has nothing to do with Altman’s morals.

by sd9

4/1/2026 at 2:56:55 PM

People don't know who Dario is yet.

Sam looks like a used car sales man. Dario looks like he'd hand out coolaid at Jonestown.

by noosphr

4/1/2026 at 3:40:16 PM

What are you actually claiming? Your take on their looks is irrelevant, and I don't know if you're hinting at anything specific when you say "People don't know who Dario is yet".

by retsibsi

4/1/2026 at 3:07:12 PM

As long as it means Sam doesn't get another dollar from me. What's the flavor?

by kelseyfrog

4/1/2026 at 3:09:04 PM

Death

by achierius