The Cost of Storage Is Too Damn High

Cloud storage prices were supposed to go down. They went up. And the production industry is eating the cost while AI companies hoard every chip on the planet.

Abstract geometric composition on warm graphite of stacked horizontal blocks compressing upward, transitioning from cool steel at the base to stress orange and warning amber at the top, with hairline fracture marks, scattered particle debris, and faint engineering grid lines evoking the rising pressure of infrastructure costs on production workflows.

I went to buy a Samsung T7 last month. The 4TB. Same drive we've been using on every shoot for the past two years. Fast, reliable, fits in your pocket, perfect for on-set backup and offline editing.

It was $349 the last time I bought one. Now it's $799. Same drive. Same capacity. More than double because... reasons.

So I started looking around. And it's not just Samsung. It's everything. Every layer of storage that a production company touches got more expensive in the last 12 months and nobody really talked about it.

So I'm gonna talk about it.

What Happened

The short version is AI ate the storage market.

Every major tech company on the planet decided they needed to train models on everything, all at once, right now. That means they need chips. Lots of them. The same NAND flash chips that go into the SSDs we use on set are getting redirected to data centers. NVIDIA can't make GPUs fast enough. Samsung and SK Hynix are prioritizing high-bandwidth memory for AI over standard NAND production. The whole supply chain shifted.

And the people who got squeezed? Small and mid-size businesses that just need to, you know, store stuff.

Production companies. Architecture firms. Healthcare providers. Anyone who generates large files and needs them accessible. We didn't suddenly start using more storage. The cost just... went up. Because the companies with the biggest checkbooks got to the front of the line.

It's Not Just the Drives

If it were just hardware I could stomach it. Buy the drives, eat the price increase, move on. But it's everything.

Frame.io. We use it for every client delivery. Review links, approvals, final exports. It's genuinely great software and we're not switching. But their pricing tiers keep shifting and suddenly the plan that worked for our volume six months ago doesn't anymore. Need more storage? That's another tier. Need more team members? Another tier. It adds up fast when you're pushing terabytes of 4K deliverables through it every month.

Google Drive. We used to keep shared project folders with clients there. Simple, everyone has a Google account, easy access. Then Google Workspace changed their storage policies. The "unlimited" plans aren't unlimited anymore. The per-user storage caps got tighter. And if you want to actually store production-level files in Google Cloud? That's a whole different pricing conversation that makes your eyes water.

Dropbox... I don't even want to talk about Dropbox. They went from a tool we used daily to something we actively avoid recommending because the pricing relative to what you get is genuinely insulting at this point.

Every platform found a reason to charge more, and the reason is always vague enough that you can't really argue with it.

Our Clients Feel It

We've had to have conversations this year that we never expected to have. Clients asking why their archival storage bill doubled. Clients reconsidering how much footage they want us to keep on hand. One client, a fashion brand we've worked with for three seasons, asked us to delete everything older than 18 months because the monthly cost wasn't worth it to them anymore.

That hurts. Because six months later they're gonna need that footage for a rebrand or a lookbook or a pitch deck and it won't exist. But I get it. When your storage line item goes from $200 a month to $450 and your production budget didn't change... something's gotta give.

We've started building storage estimates into our proposals now. It used to be an afterthought. "Yeah we'll keep your files for a year, no big deal." Now it's a real cost and we treat it like one.

The Math Nobody Wants to Do

A single shoot day for us generates anywhere from 500GB to 2TB of raw footage. That's before proxies, exports, project files, graphics, audio. We clone everything to Samsung T7s on set for immediate backup, then transfer to our edit drives, then push deliverables through Frame.io, then archive to cloud. Every step has a cost now.

Multiply that across 8 to 12 active clients and you're looking at 30 to 50 terabytes of managed storage at any given time. That's not wild. That's just... running a production company in 2026.

But the cost per terabyte isn't what it was two years ago. It's not even what it was last January. SSDs are up. Cloud tiers are up. Delivery platforms are up. And the enterprise plans that used to make it manageable keep "adjusting" their pricing in ways that somehow always cost more.

Who's Actually Winning

Not us. Not you. Not the dentist who needs to store patient imaging files or the photographer backing up 15 years of RAW archives.

NVIDIA's market cap crossed $3 trillion. Microsoft spent $80 billion on AI infrastructure in one fiscal year. Google, Amazon, Meta, they're all in an arms race for compute and storage capacity. And they're paying whatever it costs because the AI upside is theoretically infinite.

Meanwhile my upside is making sure a client's brand activation footage from October doesn't accidentally get archived to a cold tier that charges $90 to retrieve.

There's something absurd about it. The technology was supposed to get cheaper. That was the whole promise of cloud everything. Economies of scale. Competition drives prices down. More capacity, lower cost per gig. Remember that pitch? I remember that pitch.

Instead we got a gold rush where the gold is silicon and the prospectors have unlimited venture capital.

What We're Doing About It

We restructured everything internally. Samsung T7s stay on active duty for on-set backup and fast editing access. Completed projects get moved to mid-tier cloud within 30 days. Frame.io gets used for active deliveries only, not long-term storage (that was an expensive lesson). After 6 months, everything shifts to cold archive with retrieval fees clearly communicated to clients upfront.

It's not sexy. It's annoying, honestly. But it cut our monthly storage bill by about 35% and our clients know exactly what they're paying for and why.

We also started having a conversation at the start of every project about deliverables and asset management. What gets kept, for how long, in what format, and where. Google Drive for lightweight client access. Frame.io for review and approval. Cold archive for everything else. It sounds boring but it saves everyone money and prevents that awful phone call eight months later when someone needs a file that got deleted.

This Is a Production Problem

I don't think enough people in this industry are talking about infrastructure costs. We obsess over cameras and lighting and edit software and none of that matters if you can't afford to store what you make.

Storage is production. It's not overhead. It's not IT's problem. It is the thing that determines whether the work you did six months ago still exists in a usable form. And right now, the economics of that are broken. From the SSDs in your kit bag to the Frame.io plan on your credit card to the Google Cloud bucket holding three years of client archives. All of it got more expensive and none of it got better.

They'll probably stabilize eventually. New chip fabs are being built. NAND production will catch up. Maybe. But "eventually" doesn't help with this quarter's invoice.

So yeah. The cost of storage is too damn high. And if your production partner isn't talking to you about it, they're either eating the cost themselves or they haven't noticed yet. Neither is great.

Anyway.

Jump to Section

Icon

0%