Hacker News Reader

Comment by rsync | Hacker News Reader

"Spruill told TechCrunch that DigitalOcean will scale to $1 billion in revenue in the next five years, and it will become free cash flow profitable (something the CEO also referred to, loosely, as profitability) in the next two."

I find this to be incredible. DO is not a speculative e-business ... they are not a social network. They are the proverbial sellers of picks and shovels during the gold rush:

"The way to get rich during the gold rush isn't mining gold - it's selling the picks and shovels."

Here is a pick and shovel seller that can't make a profit and is going into debt ...

View on HN

If they had raised VC funding you wouldn’t have written this comment.

When the reality is raising debt instead of VC funding implies that capital providers think they are far more stable than the majority of VC funding is. Debt is cheaper capital precisely because the capital providers are convinced that you are a much safer bet.

I think a lot of people on HN (and Americans in general) don’t really understand debt and have just made debt = bad into a mantra, when in fact it’s one of the best ways to finance growing companies, and usually only available to profitable, or close to profitable companies.

It’s VC funding that is a sign of a lack of profitiability and no immediate possibility of profitability.

Most picks and shovels sellers fail and go bankrupt. That is absolutely not the way to make money during a gold rush.

There is an immense supply of picks and shovels sellers during gold rushes, which caps pricing power. That's why DigitalOcean's prices are low, they have no pricing power or lack of competition; it's also why they can't make a profit despite massive demand. You're selling a non-differentiated, low value product and only temporarily doing well because of extreme demand. The same exact thing happens to those people during oil rushes (stories abound in recent times in oil boom markets after the crash, most of the picks & shovels sellers get massacred every time). They get mauled when the tide goes out as most of them are heavy on inventory and often debt at the wrong time, and then stuck with huge amounts of low value inventory in products nobody wants. Few of them see the bust coming, they go down with the bust and rarely make enough money to get rich and walk away. They spend most of the short boom building up enough capital to afford to keep up with demand, which always overwhelms during the boom phase, then extremely rapidly deflates (the bust phase).

The people who get rich(er) during a gold rush, are the mine owners with existing capital and normal, profitable operations that are not dependent on the rush in the first place.

Amazon is the one getting rich during the gold rush (they own the mine), not DigitalOcean (the picks & shovels seller that will be forced into a sale to a larger party, IBM perhaps, as the boom fades / saturates / normalizes). DigitalOcean, Linode, Vultr, and 37 others like them, are future casualties (they'll sell, most won't literally go under) of the cloud boom in one form or another. And I say that as a customer and big fan of DigitalOcean.

They're renting our resources, so there is an implied debt up front for that. Servers and datacenters are very expensive compared to the monthly revenue extracted from them. In order to expand to $1b ARR, you've got buy multiple billions of dollars worth of servers. Debt financing is how you do that.

It's not 1999 anymore. There will be increasing demand for cloud services going forward. It's more of a bread and butter rush than a gold rush.

It's an amazing (good) achievement that they were able to debt finance, and really improves their credibility as a business in my estimation.

> Here is a pick and shovel seller that can't make a profit and is going into debt ...

That's one way to see it. Another is that they need money to expand the business.

Did they miss their window? Would this business make more sense during the 2001 dot-com craze? Are startups currently afraid to go with anyone who isn't AWS/GCE/Azure because they understand the cost of moving platforms is high?

No. Compare Linode and DigitalOcean. Linode bootstrapped, took very few financial instruments to aid the journey, had a few missteps along the way, completely reinvented the entire business more than once, and still serves a niche that makes them a successful (and profitable, as in real profitable, not imaginary profitable) company. Their margins are quite good. Slicehost had a solid business when Rackspace bought them, too, despite Rackspace subsequently burying that business in their wandering-through-tech mass grave.

DigitalOcean, on the other hand, took a Sand Hill approach and is throwing money at becoming an AWS unicorn without realizing they are never going to be AWS. Read this carefully, DO: you will never, ever, ever be AWS. Full stop. You are not in the conversation, nor is Linode, nor is prgmr (but lsc knows that), nor is Vultr, and so on, and so on. Ask IBM about trying to take on AWS.

DigitalOcean is an incredible waste of capital and appears almost as if they didn’t bother to look at Linode’s story at all. The companies are nearly identical behind the scenes and are pursuing the exact same addressable market, but one is a VC darling intent on burning capital on a completely solved problem and therefore gets VC ecosystem attention and expands to fill that attention. Seriously, you can address this market with Excel and some Python scripts; I exaggerate, but not much. The amount of money DigitalOcean spends given my intimate familiarity with the problem space has been baffling to me for the last decade.

DigitalOcean should buy Linode with this debt so the Linode people can come in and fire the right people at DO and shed about seven hundred pounds of weight. When, not if, when DigitalOcean collapses, Linode will continue on and enjoy a sudden inflow of DO’s market share. Linode long predates DO and will postdate them, too.

I have a poor opinion of Linode in a lot of ways, by the way, so.

Our original business was bootstrapped with no outside investment so we know that growth model very well. In fact that bootstrapping allowed us to build DigitalOcean when no VCs were interested in funding us by self-funding through the profits from our original business.

The problem with the approach you detailed is that it is based on growth rate. If you have more customers coming to you than you have cash on hand to buy servers, you will be forced to turn customers away.

So if you are growing rapidly you will need outside investment, whether equity or debt, in order to grow the business.

In our case we raised equity that helped us to secure additional debt terms and also due to our high growth after product market fit we also needed additional cash to continue to build out our operations.

I'm a fan of bootstrapping businesses and not raising outside capital unless it is necessary, but in our case the choice was to raise capital or turn away customers.

I’ll reiterate the margins. Based on your account, my suspicion is only reinforced, actually: if you had big enough capacity problems to need a $3 million round to buy gear at the size you were in 2013, I’m mystified that your margins were that low. Was that the $10/month decision biting you (notice Linode waited) or the far bigger headcount? How far Linode got on basically two technical employees, including the founder driving gear to the datacenter in the back of his Ridgeline, would seemingly surprise you, as would when the revenue was sufficient to sustain ongoing server spend. Linode ran out of capacity ALL THE TIME. It’s a HUGE market. Turning away a customer is not fatal in the slightest.

One server, three months, paid off. Two months profit, next server. Linode leased in the beginning! I’m not arrogant enough to assume I could do it better, but I do know the technical side very well, and with an American Express and a decent funnel of people you know you can be profitable in under a year at this. The fundamentals of what is basically a rack-and-stack game, and what is possible with $400 million of capital... it just doesn’t align with how I’d expect a VPS provider to operate, but you’ve convinced the checkbooks to keep it going, I guess.

I actually compare DO and Linode’s approach all the time as an example of VC methodology versus patience. The $400 million raise game is not appropriate for the VPS space. It’s a market literally defined by bootstrappers.

If you hit $1 billion ARR in that market, by the way, I’ll shut up. Knowing what I know about the market, that sounds about as likely as DigitalOcean colonizing Mars, but I’ll applaud you if you do it.

I apologize. I'm confused - is the point that we should have raised less debt? Or not used debt?

Linode was founded in 2003 and grew to $100MM in revenue in 16 years.

DigitalOcean was founded in 2011 and launched in 2012 and grew to $250MM in revenue in 7 years.

Stands to reason we would need more money over a shorter period of time to achieve that.

The $3MM seed round wasn't used to buy equipment but to fund the business. It improved our balance sheet which allowed us to obtain more leases from vendors which were getting worried about how much exposure they had to us without much of a financial history.

Secondly, we went from signing up 5 customers a day to 250 customers a day after product market fit. When we signed up 5 customers a day I could do most of the customer support myself along with one employee and some backstop from our original company. But at 250 customers signing up every day we obviously needed a dedicated support team, so there were immediate necessities to hiring more people as we went from an "idea/product" to a complete company. We didn't need to hire one more support person, we needed to hire an entire support team over night so that we could have 24/7/365 coverage. Again hard to do that if you don't have capital available to pay salaries. And that's just one team/function of the company that underwent tremendous stress pre and post product market fit.

As for the financial health of the business:

The debt terms require repayment as you yourself know. Whether you use leases or have a single larger structure like debt, either way this isn't "burn" money. You need to repay it with interest.

With an equity raise you can "burn" the money because you never have to repay it, as investors received stock in exchange for the funds.

When you look at the equity side of the business we have raised a total of $123MM to date and the last raise was in July 2015. We haven't raised any outside equity capital to fund the business since then.

Meaning that we are capital efficient and not losing $50MM/yr or some outrageous amount. Otherwise we would have been forced to raise an additional round of funding.

> raise capital or turn away customers

If you have to borrow $100MM to provide the same basic service that you’ve been providing for 8 years, then aren’t these “customers” less customers and more recipients of free services?

Perhaps, that’s the point, no? Buy the remaining VPS market with unsustainable pricing, and then raise the price...

Let's say a server costs $10000 and gives you $500 a month in profit.

Borrowing money to buy that server is definitely not giving away free stuff. Nor is that pricing unsustainable.

No matter how profitable a business is, there's a limit to how fast it can expand debt-free based on how fast the profits accrue.

First,"original business". Second, there's no problem with the approach, it's literally validated as not having a problem in the wild. Right now.

>realizing they are never going to be AWS

I mean not real AWS, but I think there is a market between, say good old VPS ( which is what DO and Linode are before everyone are "Cloud" ) and AWS. Managed solution like Database, Object Storage, Backup, Simple CDN. Which is certainly what DO AND Linode are going into.

There is another trend I spotted, good old Bare Metal is coming back. Useful for base load.

I don't disagree with you on your premise that AWS has an entirely different, and much larger size and scale than DO. And has a competitive advantage due to that. But your comment also seems to make the assumption that everybody should be content to let AWS (or Azure, etc, entities with very deep pockets) become a literal monopoly, and everybody should be totally fine with that.

Nope. I didn’t say anything like that at all. You just don’t best AWS by trying to replicate them, since they have enough capital to create four of you several times a year (and do, i.e., Lightsail).

Saying that, they themselves really need to look at Baidu, or Alibaba.

In the approach of "throwing money on the problem until it works" few can beat them. Alibaba's losses were surreal when they tried to get even to the tenth of AWS scale, but they went way further, and now they are profitable.

Alibaba and Baidu have a totally other market segment, which is domestic Chinese companies that want and need to be fully compatible with Chinese law for domestic data hosting, retention, government compliance. There is a venn diagram overlap between the markets addressed by the (AWS, Azure, DO, etc) and the Chinese virtual machine providers, but it's nowhere near 100%.

No, they have tons of business outside of China, just take 5 minutes to google that.

People can of course challenge their bookkeeping, but still...

>No, they have tons of business outside of China

If not for China they wouldn't be where they are and wouldn't have succeeded. Having a large core market you are the most competitive in (due to government regulations) provides a massive upside that you can then leverage in other markets.

Linode tried to hide the fact they were compromised multiple times. I don't trust anything they say.

You may be right but it's hard to take throwaway accounts seriously.

Fine, ask lsc. I have tremendous respect for Luke given that he had, thanklessly and for years, tried to compete in this exact market singlehandedly. Even competing against him at Linode, that was remarkable to me.

I have receipts for my Linode employment and I am unconcerned about the veracity of my analysis (I’m actually paid for said analysis IRL). Think about that bar you’ve set for a minute: “this person doesn’t actively court a reputation on Hacker News, therefore the information is suspect.” If that sounds good to you, we are unlikely to agree on anything, and I’ve little interest in defending my methods to someone who values speaker over spoken.

I have other gripes with Linode, but it’s not exactly par for a ‘disgruntled’ former to speak positively about a business, no?

Thanks for whatever you contributed to at Linode <3

No, most startups using either AWS or GCE (at least in the market I'm around in, Europe-based) is choosing them because they are giving away free credits and you can usually run the first year for free before starting to think about costs, and at that point you've reached market validation and can afford start thinking about costs.

I think it's really two different markets. DO is for lower end sites that need little more than a VPS or two.

> during the 2001 dot-com craze?

The x86 virtualisation tech just wasn’t up to it back then.

Reply on news.ycombinator.com