I agree with the other comments in this thread that Palantir isn't really sensibly called a startup, given its ability to scale and its current size. But isn't 1-2 years worth of salary a pretty typical funding round?
Probably much less. The organization is very flat and for the first several years they offered only 3 salary options close to 100K (something like 80, 100, 120) since they were still offering equity. Last I heard this system is still in place. So let's just say the typical overhead is around 150K and multiply that by 1500 employees (NYTimes/Wikipedia, 2015) and that gives us around $225 million.
Yeah, at a headcount of 1,500 employees (wikipedia) plus perks, expenses, etc. it's about a year's worth of operating budget.
Considering they just raised about $500m in the last 12 months before this...they're either buying expansion at an insane rate, or they're not making any money and are trying to figure out how to float this many people.
Their careers page says they have 73 openings across 19 offices...places that include Abu Dhabi, Canberra, London, LA, NYC, Tokyo, D.C., Copenhagen, Helsinki, Oslo, Tel Aviv, Singapore, Sydney, Wellington...so they're no longer just a U.S. concern.
Now this is one company I don't understand. From everything I read, it sounds like a technology consulting company (which is going to have huge scaling issues). Yet it raises money like a software startup. What's going on?
> CEO Alex Karp announced in 2013 that the company would not be pursuing an IPO, as going public would make “running a company like ours very difficult.”
I'm curious what they are telling investors about possible liquidity events. It's not like there are many companies out there that can do a $20B acquisition and if they're not planning to go public I wonder how the investors are planning to get enough money out to cover their risk and beat their expectation from other investments.
Their revenues are growing incredibly fast. Yes, they started out in the intelligence sector, but since then they have typically been taking only one client in each new industry they throw their solutions at, e.g. finance, oil & gas, utilities. They look for some of the hardest problems in each industry and solve those and then move on to another industry trying to push their valuation as high as possible. Palantir actually rejects would-be customers who reach out to them if the potential customer is in an industry they have already tackled because they don't have the manpower to take on new challenges and also repeat solving of problems they have already solved. Most stories out there just cover their early work whereas they have been scaling by leaps and bounds. Some estimates put this amount of funding at 3% of their valuation.
That makes no sense. Surely there are some efficiencies in "repeat solving" problems that they have "already solved".
A cynical thought would be that perhaps they don't want to have to sell to a company in an industry where they already have a reputation. That may by no means be true, perhaps instead they just don't have the attention span to really product-ize a solution instead of making one-offs. Or something else, either way, it smells of their typical cultish marketing.
They are primarily consulting for the Government, which has caps on how much profit you can make (at least my company was capped). They are probably trying to diversify their tech into market verticals that have less restrictions.
How much of this is cash-out of previous investors & employees? AFAIR Palantir employees have compensation packages ridiculously overweighted to illiquid and hard-to-value stock.
Are you serious? A "good" cybersecurity company? Why on earth can you make that bold statement? Unless proven otherwise, any big data companies in bed with the agencies are not considered good these days.
Have you forgotten, that your metadata is still being scooped up at large and propably sifted through thanks to companies such as this very exact Palantir?
I believe they also have a big data BI software platform. Think about how much work has gone into big data solutions like hadoop, spark and storm which requires technical data scientists/BAs/Devs to gather insights. If you could create specialized BI solutions for big data there's a lot of money in that
Some people care on what other folks work on - for ethical reasons. Remember the hacking team hack recently? A lot of folks noticed on how they sold their tech to opressive systems.
In the same line, some of us are watching when tech which can be used to single out targets (yes, that what big data is used for in intelligence) is featured in the news or here on HN.
We shall not work for them or the target's blood will be at our very own hands.
I asked what was more or uniquely interesting about Palantir, and even if I did think what they did wasn't interesting (I don't think that), what's so absurdly surprising about that?
As you're well aware, I'm constantly baffled by how little positive intent is assumed on my behalf around here, but no, I didn't mean to suggest the company isn't generally interesting to me.
It's a source of frustration for me that this line of comments is even getting any attention at all, considering its personal nature (the topic seems to be what I find interesting).
Of what relevance is it to the submission whether or not I personally find what Palantir does to be interesting? Wouldn't a much more valuable conversation be about the apparent-but-perhaps-not-actual uniqueness of what Palantir does?
They make software for non-technical users to explore relationships in unstructured data, and they do a lot of in-house consulting with their clients in terms of migrating data and building custom software add-ons. It's really not anywhere near as sinister as people make it out to be.
Contrary to the WSJ article, I don't think Palantir does any data mining. Instead, Palantir's core competence is taking a customer's existing data and appropriately modeling it in their software.
"A document leaked to TechCrunch revealed that Palantir's clients as of 2013 included at least twelve groups within the US government, including the CIA, DHS, NSA, FBI, CDC, the Marine Corps, the Air Force, Special Operations Command, West Point [...]"
Plenty of CIA analysts are non-technical, for starters, and I would imagine the same holds for the other agencies you mentioned. One does not need to be a programmer or especially well-versed in technology to hold those jobs. I think we're splitting hairs on the definition of "non-technical," but we can both agree that parsing large amounts of data isn't the easiest thing in the world to do, especially without formal training. Palantir's software does help with the data exploration process, as their website examples and demos should show.
I've done money laundering and fraud investigations for financial firms. Never got to use Palantir, but I know a guy who did. He said it's basically for people who don't know how to query or code on any level. I'm not surprised either. I can count the people in my field who want to learn SAS / SQL / R / VBA / Whatever on one hand.
Every big company that sells enterprise software has those same clients. Those are Oracle clients, IBM clients, Microsoft clients, and Cisco clients. Why is Palantir special for having a GSA price list?
Personal attacks are not allowed on HN. MaxScheiber posted a fine comment, which implied that he had worked there (or had inside knowledge somehow), so your comment doesn't "explain" anything, it merely poisons the discussion.
The last thing we want at HN is to discourage people who know things from sharing what they know. We don't need you to like Palantir, but we do need you to refrain from commenting disrespectfully here.
You don't do consulting and not understand what your client is doing with the data they have. Palantir doesn't need to do any data mining of their own. Making a model of unethical data is still unethical.*
*I don't know if Palatir is an unethical company or not, just responding to the accusations.
Microsoft and Dell aren't touching the data and building PCs hand-in-hand with the company to ensure the computer run exactly as the company needs them to. Microsoft and Dell don't need to know what anyone is doing with their PCs. Palatir needs to know what people are doing with their software so they can structure models around that exact set of data.
From my experience with companies that sell consulting/technology to the government, 'blowing the whistle' is more likely to lead to people questioning how the government spent so much money on something so rudimentary, rather than discovering something sinister.
According to crunchbase [1], they've raised over $1.6 Billion, last raise was $50m in December and $444.2m last September! For a software startup making knowledge management software with a nice interface!
I feel like they'd be pretty easy to disrupt technology-wise. There's enough big-data, open source graph, visualization and mapping front-ends to probably build a reasonable competitor pretty quick.
If they ever have to run as a real business, it's going to hurt real fast. Either that or I'm horribly wrong and something else is going on, and hiding it all behind private equity markets is making it hard to see.
One thing is for sure, most of the options the employees have are probably diluted beyond all recognition at this point.
At the least, it is not easy to break into government contracting - the expense of the security clearances necessary, having to bid for contracts, and working with the complicated systems with all of the typical bureaucracy attached to them makes it not a simple thing to disrupt. Connections can matter a lot as well.
There's a reason why most of the companies in the government contracting world in the US are large established companies.
Actually, the SBIR/STTR programs[1] streamline this process quite a bit. My company has been living in this space for its entire 20-year existence, but it offers an "in" for small companies. In fact, fostering small business creation and supplier diversity are specific goals of the SBIR/STTR programs.
Furthermore, the vast majority of government work does not require any sort of formal security clearances.
Steady (and growing) cash flow from contracts with large organizations (whether you like their clients or not), very strong technical staff and consequently, inventions, prominent brand and public awareness, etc..
That's pretty vague. Of course they can hire good people if they have lots of money. But what exactly is it that makes the company valuable? This seems to be going round in circles. People keep giving them money, so they must be worth lots of money.
i feel like you're being obtuse? don't confuse money raised in investment rounds (such as in the OP) with money that customers pay to palantir.
a company's value is the discounted value of its cash flows, future and present. palantir should expect to make money from its contracts, sale of software (i'd lump the value of their patents, etc here), and whatever else they have dreamed up.
since they have lots of customers paying them a lot, palantir has value. investors want to invest in a company that's growing. hence they give palantir money so palantir can keep growing. if palantir doesn't keep growing, they'll lose (part of) their investment.
I understand that people are giving Palantir a lot of money. We don't know (do we?) how much they are actually getting from customers. My question is: what does Palantir actually do to create so much value?
What is their revenue, then? I thought that as a private company they don't have to declare it. Now, I've asked the same question a bunch of times and instead of just saying "I don't know", which would be the honest answer, people are just getting grumpy.
You severely underestimate the value of large govt. contracts. Govt. contracts are very sticky. The govt. doesn't often change the software it uses, and when it does change the software it uses it signs up for ludicrously huge sums. You need terrific political connections to get these govt. contracts. Even with better technology, a start-up can't win a govt. contract. Ask one of your friends that works at Amazon about how long it took them to get govt. to use AWS.
Also, if you have worked with a govt. institution you'll quickly learn that the processes and human capital used by the govt. are quite poor compared to the private sector. As a result, govt. contractors quickly figure out special tools/processes/training to serve the govt. which are not easy to replicate without the experience of actually serving a govt. dept.
Many things, which in the private sector you might leave to your intelligent counterpart at the customer, with the govt. you have to handle yourself, or develop a special training program. The govt. can budget for these training programs or additional assistance, so it will not affect the bottom line ... but you have to develop those programs and processes, and you can't underestimate the importance of it.
In the private sector if on the customer side an engineer cannot figure out how to go through a simple 3 step process to configure some software, he may be fired by his supervisor. In the govt., you will be asked to break the 3 step process down into 10 smaller, simpler steps, and to create a special video tutorial for govt. workers. Even after that, you may need to staff a help-line that is open during East Coast working hours.
It is very easy in Silicon Valley to overestimate the capabilities of the people who will be using your software, because you may be dealing with very good talent at your early customers in California. The minute you get your first govt. contract you'll see that you were badly mistaken about what kind of resources you need for customer support.
So in the government contracting space there's effectively two kinds of work that are relevant here:
- products, which Palantir claims to be
- services, which Palantir appears to be by offices and head count
You are right that product contracts (and associated services like installation and training, O&M, etc.) can be fairly sticky, but the government has been on a trend towards open-standards, open-source etc. and proprietary software products are under lots of pressure to be incomparably better than anything else out there.
But "cheeks-in-seats" service contracts are not. The government is notorious at tossing out incumbents on service contracts and bringing in new companies, who then simply hire all the people who worked on the previous contract.
Palantir seems to push themselves out there as a two-product company with some sales-staff. It's obviously not going to be turn-key software, there's likely to be some configuration and support stuff up-front. But once you, as a product company, get big enough, even those tasks can be handled competently by other companies. [1] But their product sales don't look to be a huge dollar amount. fbo has a couple awards from 2011 for ~$750k licenses. This is four years ago, but we're not looking at dozens of millions of dollars.
edit here's an award for $20m for Palantir support to the war in Afghanistan. Note that DoD only reports contracts over $6.5m. So the bulk of their DoD awards must be falling under that threshold.
So it's likely most of their growth trajectory has been in specialized consulting services that direct customers to their products, but also use other tools and solutions. But again, once the ecosystem is big enough, the government can just try to compete that work out and hire another company to handle that work...companies that might not direct the solution-space towards Palantir's products.
Looking at Palantir only as a software company is incredibly misleading. They sell high-ticket custom solutions to massive organizations that they leverage many of their internal software tools and libraries to build. This involves large amounts of custom product and engineering development, ingesting complex data sources, integrating with operational processes, and communicating results and insights back to the customer in a way that changes the customers behavior and adds value.
Look it, they mine big data and then draw conclusions from it. They're like a software version of human precognition, and have made intuition into an algorithm. Hard to say if these investments are justified when a lot of the problems posed by Big Data can be solved with things like R and filtering. I can see the need to buy more storage for such datasets, and 'elastic' storage sounds like the hard problem here - not data mining.
This does not work, but the US government is fond of snake oil and security theater. Real security comes from stable people with stable ideologies. The US instead often as a result of bad policy destabilizes ideology and people in certain regions and then tries to use technology to make up for the lack of wisdom. Which leads to security theater and other security snake oil.
I don't think the term startup means that you have only a few employees. Rather, it means you are still researching your business model, still in a growth stage, etc. Certainly Tesla and SpaceX are startups, despite having thousands of employees. Tesla is experimenting with home batteries, licencing tech, SpaceX considering internet satellite constellations, etc. They are both very far away from achieving their goals, and stability as revenue-generating, competive companies. Definitely still startups.
Is there a name for the (anti?)pattern where a company pushes hard to be considered a "product" company, but has to scale its hiring to its customers pretty linearly?
Such a company might have a general software package but that package has to be changed for every customer.
84 comments
[ 3.2 ms ] story [ 148 ms ] threadok, two years?
Considering they just raised about $500m in the last 12 months before this...they're either buying expansion at an insane rate, or they're not making any money and are trying to figure out how to float this many people.
Their careers page says they have 73 openings across 19 offices...places that include Abu Dhabi, Canberra, London, LA, NYC, Tokyo, D.C., Copenhagen, Helsinki, Oslo, Tel Aviv, Singapore, Sydney, Wellington...so they're no longer just a U.S. concern.
> CEO Alex Karp announced in 2013 that the company would not be pursuing an IPO, as going public would make “running a company like ours very difficult.”
https://en.wikipedia.org/wiki/Palantir_Technologies
11 years old, 1500 employees, never planning to go public. At some point you have to stop calling it a "startup".
A cynical thought would be that perhaps they don't want to have to sell to a company in an industry where they already have a reputation. That may by no means be true, perhaps instead they just don't have the attention span to really product-ize a solution instead of making one-offs. Or something else, either way, it smells of their typical cultish marketing.
If you dislike Palantir, though, you should also dislike Sqrrl.
Just seems like there's a weird (and generally incorrect) fascination with Palantir that I don't understand.
In my field specifically, Palantir has this reputation for being a good cybersecurity company that simply doesn't come from anywhere concrete.
It's just confusing.
Have you forgotten, that your metadata is still being scooped up at large and propably sifted through thanks to companies such as this very exact Palantir?
In the same line, some of us are watching when tech which can be used to single out targets (yes, that what big data is used for in intelligence) is featured in the news or here on HN.
We shall not work for them or the target's blood will be at our very own hands.
It's a source of frustration for me that this line of comments is even getting any attention at all, considering its personal nature (the topic seems to be what I find interesting).
Of what relevance is it to the submission whether or not I personally find what Palantir does to be interesting? Wouldn't a much more valuable conversation be about the apparent-but-perhaps-not-actual uniqueness of what Palantir does?
Contrary to the WSJ article, I don't think Palantir does any data mining. Instead, Palantir's core competence is taking a customer's existing data and appropriately modeling it in their software.
From https://en.wikipedia.org/wiki/Palantir_Technologies#2013
I wouldn't call these fellas "non-technical users" myself.
Edit: Ah you're a former employee of Palantir, that explains a lot.
The last thing we want at HN is to discourage people who know things from sharing what they know. We don't need you to like Palantir, but we do need you to refrain from commenting disrespectfully here.
*I don't know if Palatir is an unethical company or not, just responding to the accusations.
Yes. https://en.wikipedia.org/wiki/IBM_and_the_Holocaust
According to crunchbase [1], they've raised over $1.6 Billion, last raise was $50m in December and $444.2m last September! For a software startup making knowledge management software with a nice interface!
I feel like they'd be pretty easy to disrupt technology-wise. There's enough big-data, open source graph, visualization and mapping front-ends to probably build a reasonable competitor pretty quick.
If they ever have to run as a real business, it's going to hurt real fast. Either that or I'm horribly wrong and something else is going on, and hiding it all behind private equity markets is making it hard to see.
One thing is for sure, most of the options the employees have are probably diluted beyond all recognition at this point.
1 - https://www.crunchbase.com/organization/palantir-technologie...
There's a reason why most of the companies in the government contracting world in the US are large established companies.
Furthermore, the vast majority of government work does not require any sort of formal security clearances.
[1] http://www.acq.osd.mil/osbp/sbir/
I think their success is a mix of massive investment in sales, marketing and BD, plus reasonably good execution on purchase.
This competitive intersection allows them to position themselves against all kinds of seemingly unrelated companies.
a company's value is the discounted value of its cash flows, future and present. palantir should expect to make money from its contracts, sale of software (i'd lump the value of their patents, etc here), and whatever else they have dreamed up.
since they have lots of customers paying them a lot, palantir has value. investors want to invest in a company that's growing. hence they give palantir money so palantir can keep growing. if palantir doesn't keep growing, they'll lose (part of) their investment.
Except that he didn't. And he never would.
Like saying that the Mona Lisa is a nice painting of a woman.
Also, if you have worked with a govt. institution you'll quickly learn that the processes and human capital used by the govt. are quite poor compared to the private sector. As a result, govt. contractors quickly figure out special tools/processes/training to serve the govt. which are not easy to replicate without the experience of actually serving a govt. dept.
Many things, which in the private sector you might leave to your intelligent counterpart at the customer, with the govt. you have to handle yourself, or develop a special training program. The govt. can budget for these training programs or additional assistance, so it will not affect the bottom line ... but you have to develop those programs and processes, and you can't underestimate the importance of it.
In the private sector if on the customer side an engineer cannot figure out how to go through a simple 3 step process to configure some software, he may be fired by his supervisor. In the govt., you will be asked to break the 3 step process down into 10 smaller, simpler steps, and to create a special video tutorial for govt. workers. Even after that, you may need to staff a help-line that is open during East Coast working hours.
It is very easy in Silicon Valley to overestimate the capabilities of the people who will be using your software, because you may be dealing with very good talent at your early customers in California. The minute you get your first govt. contract you'll see that you were badly mistaken about what kind of resources you need for customer support.
- products, which Palantir claims to be
- services, which Palantir appears to be by offices and head count
You are right that product contracts (and associated services like installation and training, O&M, etc.) can be fairly sticky, but the government has been on a trend towards open-standards, open-source etc. and proprietary software products are under lots of pressure to be incomparably better than anything else out there.
But "cheeks-in-seats" service contracts are not. The government is notorious at tossing out incumbents on service contracts and bringing in new companies, who then simply hire all the people who worked on the previous contract.
Palantir seems to push themselves out there as a two-product company with some sales-staff. It's obviously not going to be turn-key software, there's likely to be some configuration and support stuff up-front. But once you, as a product company, get big enough, even those tasks can be handled competently by other companies. [1] But their product sales don't look to be a huge dollar amount. fbo has a couple awards from 2011 for ~$750k licenses. This is four years ago, but we're not looking at dozens of millions of dollars.
edit here's an award for $20m for Palantir support to the war in Afghanistan. Note that DoD only reports contracts over $6.5m. So the bulk of their DoD awards must be falling under that threshold.
edit here's a 5 year contract from DoJ for ~$10m. http://www.govconwire.com/2014/04/doj-selects-palantir-softw... hardly $20billion dollar valuation contract sizes.
So it's likely most of their growth trajectory has been in specialized consulting services that direct customers to their products, but also use other tools and solutions. But again, once the ecosystem is big enough, the government can just try to compete that work out and hire another company to handle that work...companies that might not direct the solution-space towards Palantir's products.
1 - https://www.palantir.com/partnerships/service-providers/
The term is being diluted...
Same goes with Uber, etc.
As Paul Graham says, "A startup is a company designed to grow fast." and it seems like Palantir is still growing incredibly fast.
By that definition, Palantir is certainly a startup.
Such a company might have a general software package but that package has to be changed for every customer.
> Palantir is trying to raise another ~400mm to finance their operations.