The article talks about averages, but what I want know is the median. The usual situation, and I have zero reason to believe OpenAI is different, is that stock options are top heavy leaning heavily toward executives.
I want to know rank and file salaries as opposed to stock options
Business & back office employee salaries are standard but not impressive. Similarly, stock grants are better than most places but not wildly high unless you're in specific engineering & research functions. This is the same at Anthropic, too (I recently interviewed for director level business roles at both).
I think the simpler answer (and thus more likely) is that statistics education in the US is very bad (they should teach stats in high school, not calculus), and reporters just don't understand it. And the ones who do, believe (correctly) that their audience does not understand it.
They do, but you sell forward contracts instead. This is perfectly legal, and the approach I've seen. There are a few companies, and even funds that will engage in this, in an effort to attain future upside.
They've changed the laws recently which makes it far easier - I believe you'd still need to be accredited but for most of HN, that's a low bar. For OpenAI specifically, they've allowed employees to participate in the funding rounds and they did a separate tender offer with Softbank to provide liquidity to early employees as well;
Browsing OpenAI's careers page, I'm seeing at most $275k for most positions, so I'm assuming the median is much lower than an average being pulled up by a few rockstar positions.
Yes, you might need approval, but if there's regular secondary sale does it matter?
> OpenAI has finalized a secondary share sale totaling $6.6 billion, allowing current and former employees to sell stock at a record $500 billion valuation, according to a person familiar with the transaction.
Again, that doesn’t make the assets particularly liquid. They did this and that’s fine, but if they hadn’t done this you’d be stuck with a piece of paper even after the lockup period ends.
Until traditional RSUs that once they are vested you can sell them, with few exceptions
Indeed, it's not guaranteed. But now it seems to be expected for those large private companies to unlock employee equity at regular intervals (I'm not sure it was the case 5-10y ago).
Aren't their secondary markets for this? My wife gets offers constantly for her options even though they're not public yet and the offers are higher then what she was awarded them at. Maybe it's scams? we never took any of them up on it.
You must be looking at non tech positions, most of their research/applied role go up to ~550k, and they do offer more than advertised for strong candidates. + hiring cash bonus + equity (which is a lot).
I just can't see any clear winners in the AI race. At least not as far as the models/products go.
Maybe a wild round of mergers & acquisitions, combined with regulatory capture and some monopoly will be what settles everything. Probably with a crash in the middle of it all.
The market has not settled yet. We're in the late 90s internet era when it comes to "AI". As with the dotcom bubble, real value will only become evident after the crash.
I'm not surprised they have to pay higher, since no amount of money could convince me to work towards human irrelevance, or BigAdTech.
When I see this technology improve and free the lives of those whose salary is akin to slavery, then I might reconsider.
Context: I've been reading about the Mondragon Corporation, and it seems a much better model than this maximum extraction economy we are building. I'll submit a story for it, although I discovered it through a HN book recommendation (Kim Stanley Robinson).
Really, I found 5.2 to be a rather weak model. It constantly gives me code that doesn't work and gets simple APIs wrong. Maybe it's just weak on the domain I'm working in.
I heard that the environment there is 996 with high turnover. So you might be paid double in comparison to a FAANG job but you work double as well. (This was about dev positions not researchers)
Anyone know if that’s true? I only heard it second hand.
Employees also have no moat, and that is not unique to OAI. You either work as much as the other people on your team or you get replaced by one of the many people eager to work harder than you for that money.
This assumes a constant stream of an available workforce. Meanwhile, in the US where OpenAI is based, scrutiny and pressure from the current administration is making it harder to hire at their largest locations.
OpenAI is exactly what happens when a company finds itself in such a far, far away blue ocean strategy that there are no more traditional "economic anchors" (to call it that) to reason with.
It usually ends in blood and tears, for both employees and investors.
BUT: the SOTA has been greatly advanced, which matters a great deal more than the destiny of a particular corporation or the social status of sam-i-am.
I like this viewpoint - it basically casts VC-backed AI startups as privately-subsidized applied R&D projects, which largely seems to be the case for foundational model companies.
I don't think this way at all. RSU (for public companies) is real, tangible money. Sure, there is some amount of risk involved with it, but it is WAY more tangible than options than can just be diluted if the person with the cap table is feeling ornery that day. The key that makes RSU's real money is their immediate liquidity on award, so you can literally just treat them as cash compensation in this regard if you decide not to hold them.
When you work at BigCorp for an extended period of time, your salary often ends up being majority by RSU as the vest rolls start to stack up
Yeah, agreed. It was options I always thought of as a "nice bonus", but RSUs I always thought of like cash.
The tax situation around options also makes for many more variables in the risk calculation (i.e. do I exercise as soon as I vest to reduce potential AMT, but then I may need to find sizable cash to execute), while RSUs are just much simpler
This is not true. Most fast-growing private tech companies with valuations a fraction of OpenAIs switch from ISOs to RSUs because exercise costs start to become prohibitive.
Of course they aren’t liquid yet, but they’re still RSUs, and it’s much more common than OpenAIs profit shares structure
Anyone that works for startups knows that it’s not really “compensation” until it’s cash in your bank account. Until then it’s just a theoretical number on paper, which tends to end up being worth a lot less than originally advertised/hoped.
I’ve lost track of the number of times that someone’s startup got acquired for (insert what sounds like a big number) and everyone is like “wow the employees must all be rich” only to find out later that after preferred cap tables and other terms the employees got very little.
A lot could happen here, but history says “watch this space” on this stock-based comp. Some options on the secondary markets but that only works as long as OpenAI can convince more people to dump money on the burning pile of cash they have going at the moment.
the industry secret is "not really", no, you don't
way before private secondaries got big, there were boutique funds and lawyers that set up the contracts and structures to circumvent this friction
and most companies do not care about that covenant, they care about the optics around their cap table to attract other investors and also satisfy securities regulators
The best people have all seemed to fled Amazon the last 24 months. LinkedIn is flooded with long tenured top tech folks leaving. I doubt that gives those left a great deal of leverage and Amazon was never known for market-leading comp.
I skimmed but missed that, thanks! I have seen so many times where comparisons are breathlessly made without adjustment in the media, so I’m pleasantly surprised it was done here.
The article talks about averages, but what I want know is the median. The usual situation, and I have zero reason to believe OpenAI is different, is that stock options are top heavy leaning heavily toward executives.
I want to know rank and file salaries as opposed to stock options
Business & back office employee salaries are standard but not impressive. Similarly, stock grants are better than most places but not wildly high unless you're in specific engineering & research functions. This is the same at Anthropic, too (I recently interviewed for director level business roles at both).
In general, I wish the media would stop using just the average when the distribution is not normal.
That's exactly why they use averages, though. Propaganda is insidious in that way.
I think the simpler answer (and thus more likely) is that statistics education in the US is very bad (they should teach stats in high school, not calculus), and reporters just don't understand it. And the ones who do, believe (correctly) that their audience does not understand it.
Who is “they”? I believe open AI would share averages for that reason but not that media would choose to cover it that way.
The whole way media treats numbers is more than tiring:
"X increased by Y"
Sure... but:
- What's the relative increase? - Is this increase out of the ordinary? Annually? Globally?
But I don't think this is some sort of conspiracy. Rather: Most journalists are not very smart.
Illlquid “stock options” in a private company is not what I consider compensation.
This is the general rule, but not for ones the size of OpenAI. There’s always a secondary market for prominent enough companies.
Dont all private companies require approval for secondary sales, which I assume are not ever approved?
They do, but you sell forward contracts instead. This is perfectly legal, and the approach I've seen. There are a few companies, and even funds that will engage in this, in an effort to attain future upside.
Its pretty common, Ive personally done this.
Typically. I’d be shocked if OpenAI let employees sell their options like this without requiring approval
They don't but you effectively do it under the table
Do they sell at full value?
Whatever they sell them for is the value.
How do you define value without an IPO?
OpenAI has regular tender offers for their employees, so while this advice is reasonable in general it is less true for this case.
Less true isn’t the same as not true. Simply put we don’t know because it’s not what they are disclosing
This article feels more like paid publicity than it does journalism
I am sure you can make OpenAI stock liquid pretty easily.
Only if you're allowed to which is not always the case.
They've changed the laws recently which makes it far easier - I believe you'd still need to be accredited but for most of HN, that's a low bar. For OpenAI specifically, they've allowed employees to participate in the funding rounds and they did a separate tender offer with Softbank to provide liquidity to early employees as well;
https://fortune.com/2024/12/17/hundreds-openai-employees-10-...
It’s allowed pretty easily.
They had had tender events (where you can sell your private stock super easily)
Dont worry...there is always an acquisition by Meta, on the horizon for any company with nowhere to go.
Browsing OpenAI's careers page, I'm seeing at most $275k for most positions, so I'm assuming the median is much lower than an average being pulled up by a few rockstar positions.
You also need to take into account equity, since it went up 250% in a year it can be a large amount of someone's compensation.
Until it’s really liquid it’s still fiction. The salary and other cash compensation is all that matters until you can actually sell the options
Want does “really” mean?
If there are enough opportunities to offload stock on the secondary market (which seems to be the case of them), then it’s not fiction.
If you can show me that they can sell without OpenAI approval on the secondary market, I would concede in a heartbeat.
If not, I am to assume this isn’t true, and that they are functionally non liquid possible assets at the discretion of OpenAI to sell
Yes, you might need approval, but if there's regular secondary sale does it matter?
> OpenAI has finalized a secondary share sale totaling $6.6 billion, allowing current and former employees to sell stock at a record $500 billion valuation, according to a person familiar with the transaction.
https://www.cnbc.com/2025/10/02/openai-share-sale-500-billio...
Again, that doesn’t make the assets particularly liquid. They did this and that’s fine, but if they hadn’t done this you’d be stuck with a piece of paper even after the lockup period ends.
Until traditional RSUs that once they are vested you can sell them, with few exceptions
Indeed, it's not guaranteed. But now it seems to be expected for those large private companies to unlock employee equity at regular intervals (I'm not sure it was the case 5-10y ago).
Aren't their secondary markets for this? My wife gets offers constantly for her options even though they're not public yet and the offers are higher then what she was awarded them at. Maybe it's scams? we never took any of them up on it.
The article is talking about $1.5 million in stock-based compensation (i.e. equity) per employee. That's in addition to cash salaries.
You must be looking at non tech positions, most of their research/applied role go up to ~550k, and they do offer more than advertised for strong candidates. + hiring cash bonus + equity (which is a lot).
https://openai.com/careers/research-engineer-research-scient...
I just can't see any clear winners in the AI race. At least not as far as the models/products go.
Maybe a wild round of mergers & acquisitions, combined with regulatory capture and some monopoly will be what settles everything. Probably with a crash in the middle of it all.
The market has not settled yet. We're in the late 90s internet era when it comes to "AI". As with the dotcom bubble, real value will only become evident after the crash.
"OpenAI’s compensation as a percentage of revenue was set to reach 46% in 2025"
At least the revenue is large enough to cover the payroll. That's a good milestone.
Not really a fan of Altman, but I don't mind the competition he brings to the landscape.
Does that include stock? I bet it’s just cash.
The opposite. The article is just about stock-based compensation, and that 46% number is explicitly just that, not cash compensation.
I'm not surprised they have to pay higher, since no amount of money could convince me to work towards human irrelevance, or BigAdTech.
When I see this technology improve and free the lives of those whose salary is akin to slavery, then I might reconsider.
Context: I've been reading about the Mondragon Corporation, and it seems a much better model than this maximum extraction economy we are building. I'll submit a story for it, although I discovered it through a HN book recommendation (Kim Stanley Robinson).
> I'm not surprised they have to pay higher, since no amount of money could convince me to work towards human irrelevance
The way Altman and others want AI to develop, this is what they’re working toward too
Do people who contribute to YouTube Shorts get paid a lot? It's a very nasty thing in the current form.
GPT-5.2 has radically changed my outlook on OpenAI. Head and shoulders above others.
The excellence is there.
It hasn't done great in the head-to-head comparisons I've seen, and has trailed in several areas on the leaderboards (though not all).
Based on the 'code red' they declared, this model seems to have been rushed a bit.
I'm also a happy customer.
But, one thing has been consistent for the past 3 years: After every release from all the serious competitors, the hype can go either way.
As far as the hype cycles go, OpenAI is oscillating between "Best model ever" and "What a letdown, it's over" at least twice a year.
The competition is fierce, and a never-ending marathon of all the players getting ahead just a bit. No clear long-term winner.
5.2 is good. But at this point every few months company A trumps company B with a new “SOTA” (for some definition of SOTA).
OpenAI has no real moat. Anthropic is focusing on developers as a clear target, and Gemini has the backing of Google.
I don’t see OpenAI winning the AI race with marginally better models and arguably a nicer UI/UX (ymmv, but I do like the ChatGPT app experience).
That said, my usage decreases month over month.
Really, I found 5.2 to be a rather weak model. It constantly gives me code that doesn't work and gets simple APIs wrong. Maybe it's just weak on the domain I'm working in.
[dead]
Looking at the stick compensation of companies in 2000 doesn’t seem particularly relevant today?
I heard that the environment there is 996 with high turnover. So you might be paid double in comparison to a FAANG job but you work double as well. (This was about dev positions not researchers)
Anyone know if that’s true? I only heard it second hand.
I would expect so. They have no moat and are in a race to survive.
Employees also have no moat, and that is not unique to OAI. You either work as much as the other people on your team or you get replaced by one of the many people eager to work harder than you for that money.
This assumes a constant stream of an available workforce. Meanwhile, in the US where OpenAI is based, scrutiny and pressure from the current administration is making it harder to hire at their largest locations.
no paywall: https://www.wsj.com/tech/ai/openai-is-paying-employees-more-...
OpenAI is exactly what happens when a company finds itself in such a far, far away blue ocean strategy that there are no more traditional "economic anchors" (to call it that) to reason with.
It usually ends in blood and tears, for both employees and investors.
BUT: the SOTA has been greatly advanced, which matters a great deal more than the destiny of a particular corporation or the social status of sam-i-am.
So, overall: good news.
I like this viewpoint - it basically casts VC-backed AI startups as privately-subsidized applied R&D projects, which largely seems to be the case for foundational model companies.
Definitely. If VCs want to fund expensive salaries, so what?
The funniest part is there is no amount of money that would get me back in the office again
OpenAI doesn't offer traditional rsus (at least to regular employees?), but instead profit sharing units.
https://www.levels.fyi/blog/openai-compensation.html
Might change how you evaluate the value here.
The value of any traditional RSU is to treat it as a nice bonus if you get it, so not so much different from any other stock or option package.
I don't think this way at all. RSU (for public companies) is real, tangible money. Sure, there is some amount of risk involved with it, but it is WAY more tangible than options than can just be diluted if the person with the cap table is feeling ornery that day. The key that makes RSU's real money is their immediate liquidity on award, so you can literally just treat them as cash compensation in this regard if you decide not to hold them.
When you work at BigCorp for an extended period of time, your salary often ends up being majority by RSU as the vest rolls start to stack up
Yeah, agreed. It was options I always thought of as a "nice bonus", but RSUs I always thought of like cash.
The tax situation around options also makes for many more variables in the risk calculation (i.e. do I exercise as soon as I vest to reduce potential AMT, but then I may need to find sizable cash to execute), while RSUs are just much simpler
Most private companies don't offer RSUs.
This is not true. Most fast-growing private tech companies with valuations a fraction of OpenAIs switch from ISOs to RSUs because exercise costs start to become prohibitive.
Of course they aren’t liquid yet, but they’re still RSUs, and it’s much more common than OpenAIs profit shares structure
“Paying” is a relative term here.
Anyone that works for startups knows that it’s not really “compensation” until it’s cash in your bank account. Until then it’s just a theoretical number on paper, which tends to end up being worth a lot less than originally advertised/hoped.
I’ve lost track of the number of times that someone’s startup got acquired for (insert what sounds like a big number) and everyone is like “wow the employees must all be rich” only to find out later that after preferred cap tables and other terms the employees got very little.
A lot could happen here, but history says “watch this space” on this stock-based comp. Some options on the secondary markets but that only works as long as OpenAI can convince more people to dump money on the burning pile of cash they have going at the moment.
this is a completely false and outdated take, you can easily sell openai shares
Citation needed. Please provide more information.
The private secondary markets are extremely liquid if you’re a household name
The user experience is nearly the same as cash if you have an ounce of interest in having cash
You need company approval to sell yeah? That could be a major issue.
the industry secret is "not really", no, you don't
way before private secondaries got big, there were boutique funds and lawyers that set up the contracts and structures to circumvent this friction
and most companies do not care about that covenant, they care about the optics around their cap table to attract other investors and also satisfy securities regulators
oh no, that might put upward pressure on Amazon employee salary demands!
The best people have all seemed to fled Amazon the last 24 months. LinkedIn is flooded with long tenured top tech folks leaving. I doubt that gives those left a great deal of leverage and Amazon was never known for market-leading comp.
Adjusted for inflation?
From the article:
> The compensation figures have been adjusted into 2025 dollars to account for inflation.
I skimmed but missed that, thanks! I have seen so many times where comparisons are breathlessly made without adjustment in the media, so I’m pleasantly surprised it was done here.