We’ve covered BackBlaze many times in the past, most recently in April. They’re fun to watch because they expose a lot of the guts of their hardware operation (all those custom-designed pods) and publish tons of drive stats. Their B2 service is also among the cheapest per-GB pricing you can get.
However, their finances have been awful. Their stock has been oscillating between $5 and $10 forever, and since 2019 they’ve never managed to actually earn a profit.
I’ve been expecting an announcement that they’ve been acquired. They have $192 million in assets, zero long-term debt, and $26 million in cash. They employ less than 400 people. So for some hyperscaler out there who wants access to their IP or some private equity firm who wants to roll them into a portfolio, they’re pretty attractive.
However, it looks like BackBlaze management has found a lifeline, and it’s a rather creative one. They’re going to sell object storage not to consumers or small business, but rather to clouds:
Today, Backblaze announced a $335 million, multi-exabyte agreement with CoreWeave to deliver object storage capacity supporting key elements of CoreWeave’s storage infrastructure. The agreement spans more than five years.
So let’s assume that’s $335 / 5 = $67 million a year. It’s probably not structured quite that simply but let’s use that number. It’d be enough to push BackBlaze into the black. They’re losing about $25m/year at present.
This deal makes a lot of sense, assuming BackBlaze can deliver. Hyperscalers consume tons of storage and CoreWeave – which essentially is a middleman – is probably desperate to improve margins because CoreWeave is also losing money. Supplying storage to one big cloud business is a lot easier than dealing with a million consumer PC backups.
This deal does have a lot of risk:
- When I say CoreWeave is losing money, I don’t mean the bush league millions that BackBlaze is losing. CoreWeave is losing $1.5 billion a year. While they have $55 billion in assets, they also have $50 billion in debt and other obligations. Will long will they be able to pay BackBlaze?
- Five years is an eternity in this space. CoreWeave in its present iteration hasn’t been around that long. It was founded in 2017 as a crypto company, but then pivoted in 2022 when it realized it had a lot of GPUs in datacenters it could rent out to AI companies.
- “Multi-exabyte” is a lot. That could be 2 EB but let’s assume for the sake of discussion that it is 3 or 4 EB. That could be roughly what BackBlaze has on the floor now. They don’t disclose figures, but in 2019 they disclosed 750 PB (or .75 EB). Has consumer back and their B2 business doubled or tripled since then? Let’s say it’s quadrupled to 3 EB. This CoreWeave deal would mean providing 2x or more of their existing storage. If they have that much space capacity, great…if not, this could mean a huge round of capital to provide service, and then to maintain it for 5 years.
Still, considering the doldrums BackBlaze has been in, this is good news. The market likes it…their stock finally broke $10/share.




















Leave a Reply