Stop Guessing Your DVR Storage: A Real-World Surveillance HDD Sizing Guide

You set up your recorder—a DVR (Digital Video Recorder) or NVR (Network Video Recorder), which are the boxes that capture and store footage from your security cameras—popped in whatever hard drive the installer suggested, and everything seemed fine. Then, three weeks into a lease dispute or after a parking-lot incident, you go looking for footage and find out the system only kept seven days of video before automatically overwriting the oldest recordings. Seven days. The whole point was to have a buffer. That gap is almost always a storage sizing problem, and it’s one of the most preventable mistakes in any surveillance build—whether you’re running four cameras at a small office or speccing forty channels across a multi-site commercial deployment.

This guide gives you the actual math, the hardware decisions that follow from it, and clear rules for which tradeoff to pick when your budget and your retention requirements don’t perfectly align. No guessing, no vendor-supplied estimates that assume you’ll compress everything to the point of uselessness.

---

Why Drive Size Alone Tells You Almost Nothing

Here’s the conversation that leads to bad installs: the customer says “I want two weeks of footage,” the installer quotes an 8 TB drive, and nobody does the math. Two weeks of footage from eight 4K cameras in continuous mode can burn through 8 TB in under six days. Two weeks from eight cameras running at 1080p with H.265 compression and motion-only recording might fit in 2 TB with room to spare. The difference is enormous, and it lives entirely in four variables:

1. Resolution — how detailed each frame is (1080p, 4MP, 4K/8MP, etc.)
2. Bitrate — how many megabits per second the camera is actually sending, which depends on resolution, frame rate, and compression codec
3. Codec — H.264 is the old standard; H.265 (also called HEVC) cuts file size roughly in half at equivalent quality, per IPVM’s bitrate comparison research
4. Recording mode — continuous 24/7 versus motion-triggered only versus scheduled windows

These four variables interact. Getting one wrong blows up the estimate. Below is the formula that integrates all of them, and it’s the same one storage calculator tools are running under the hood.

---

The Storage Formula (And How to Actually Use It)

The core equation:

Daily storage (GB) = (Bitrate in Mbps × 3,600 seconds × recording hours per day) ÷ 8,000

Then multiply by number of cameras, then multiply by your retention target in days.

Walk-through example:

• Camera count: 8
• Resolution: 1080p at 15 fps (a typical “budget” IP camera setting)
• Bitrate: 2 Mbps per camera (H.264 codec)
• Recording hours: 24 per day (continuous)
• Retention: 14 days

Daily per camera: (2 × 3,600 × 24) ÷ 8,000 = 21.6 GB/day Total per day: 21.6 × 8 cameras = 172.8 GB/day Over 14 days: 172.8 × 14 = ~2.4 TB

Now flip the codec to H.265 and the bitrate drops to roughly 1 Mbps per camera at equivalent visual quality. Same math: you’re down to ~1.2 TB. That’s the codec dividend—real, and worth the minor camera/NVR compatibility check it requires.

By the Numbers: Resolution vs. Storage at 24/7 Recording, 8 Cameras, 14 Days

Resolution	Codec	Bitrate (per cam)	Total Storage Needed
1080p / 15fps	H.264	~2 Mbps	~2.4 TB
1080p / 15fps	H.265	~1 Mbps	~1.2 TB
4MP / 15fps	H.264	~4 Mbps	~4.8 TB
4K / 15fps	H.265	~4 Mbps	~4.8 TB

Estimates based on published codec efficiency ratios; actual bitrates vary by scene complexity, manufacturer firmware, and VMS encoding settings.

---

Choosing the Right Drive: Surveillance-Grade vs. Desktop

This is where a lot of first builds go sideways. Standard desktop hard drives—the kind that ships in most consumer NAS boxes or gets grabbed off a shelf because it’s on sale—aren’t built for the write patterns surveillance recording creates. A security camera system writes data constantly, often to multiple streams simultaneously, for years without a meaningful rest cycle. Desktop drives are optimized for reads; they tolerate relatively light, bursty writes.

Surveillance-grade drives handle this differently at the firmware level. Western Digital’s Purple series (rated for up to 64 cameras simultaneously per their published datasheet) and Seagate’s SkyHawk line both use what the manufacturers call “ATA streaming” and vibration compensation tuning for multi-drive rack environments. PCMag’s comparison of NAS and surveillance drives notes that the firmware tolerance for write-heavy workloads is the functionally meaningful difference, not just marketing. For a 4-camera home setup, a desktop drive might survive two years without issue. For a 16-channel commercial install writing continuously, Security Sales & Integration’s 2024 coverage of HDD reliability in surveillance deployments found premature drive failures disproportionately tied to non-surveillance-grade media in always-on systems.

Practical rule: If the recorder runs 24/7 and has more than four channels writing simultaneously, budget for surveillance-grade drives. The cost delta between a surveillance Purple and a comparable desktop WD Blue is typically $15–$30 per drive in the 4–8 TB range (as of Q2 2026). That’s not where you save money on a $5,000–$50,000 install.

Capacity tiers to know for 2026:

• 2–4 TB: Adequate for small installs (4 cameras, 7–14 days, 1080p/H.265). Entry point for WD Purple and SkyHawk lines.
• 6–8 TB: The practical sweet spot for 8–16 camera mid-range builds. Highest value per TB in the surveillance-grade segment based on current shelf pricing.
• 10–18 TB: For 16–32 channel commercial recorders or any deployment mandating 30+ day retention. Seagate SkyHawk AI drives extend into this range with on-drive analytics support.
• Above 20 TB: Enterprise SAS or high-density SATA for large NVR arrays; typically specified by integrators rather than purchased off the shelf.

---

Retention Requirements: What Your Client Actually Needs vs. What They Say

“I want a month of footage” sounds reasonable until you show a retail client the cost difference between 14 days and 30 days of 4K retention across a 24-camera system. More importantly, many clients don’t know their retention requirement is actually set externally—by insurance, by local ordinance, or by industry regulation.

Common retention mandates worth surfacing early in a deal:

• Banking / financial services: 90 days is common in U.S. state-level requirements for ATM and lobby coverage
• Cannabis retail (state-licensed): Many state licensing bodies mandate 30–90 days of continuous recording; some require specific frame rates
• Multi-tenant commercial real estate: Typically 14–30 days as a lease or insurance baseline
• Healthcare (HIPAA-adjacent physical security): No federal mandate on camera retention specifically, but accreditation bodies often push 30-day minimums for internal areas
• General commercial (default): 7–14 days covers most insurance claim timelines per common policy language

When you’re closing a deal and the client says “two weeks,” confirm whether that’s a preference or a contractual floor. If it’s contractual, build the storage to meet it with 20% headspace—drives that run at 100% capacity degrade faster, and you don’t want to re-open a contract to explain why retention slipped from 14 days to 11 because you left no margin.

---

Motion-Only Recording: The Honest Math

Motion-triggered recording is the most common way integrators and IT managers try to stretch storage, and it works—sometimes dramatically—but it’s worth being precise about when and why.

In a high-traffic retail environment, cameras near entrances may trigger 60–80% of the time during business hours. Motion-only recording might reduce that camera’s daily storage consumption by 30–40% versus continuous. In a low-traffic server room or storage area, the same logic might reduce daily consumption by 80–90%. IPVM’s research on storage methodology notes that motion recording estimates in vendor calculators typically assume 25% activity, which is conservative for retail and aggressive for low-traffic areas. You need to know your actual scene before you apply the multiplier.

The risk with motion-only: gaps. Any event that happens without triggering the motion threshold—camera tampering, a slow intrusion, a lighting-triggered false negative—won’t be recorded. For loss-prevention or forensic applications, the liability of a gap can outweigh the storage savings. For many commercial clients, the right answer is a hybrid: continuous recording at reduced frame rate (e.g., 5 fps) for baseline coverage, plus motion-triggered recording at full frame rate. Some VMS platforms—Blue Iris, for example, handles this natively with its dual-stream recording mode—make this straightforward to configure. Others require manual workarounds.

---

If X, Then Y: Decision Rules for Sizing Your Storage

This is the part you can actually use when you’re staring at a quote or configuring a recorder.

If you’re speccing a 1–8 camera install at 1080p with H.265 and need 14 days: → A single 4 TB surveillance-grade drive almost certainly covers it. Start there. Add a second for redundancy if the client has any tolerance for data-loss risk.

If you’re running 8–16 cameras at 4MP or above and need 30-day retention: → Calculate using the formula at the actual bitrate from the camera spec sheet (not the marketing “up to” number). Budget for 6–8 TB drives in a 2–4 drive array. Verify the NVR’s maximum internal bay count before you spec the cameras.

If your client has a regulated retention requirement above 30 days: → Price out a RAID array (RAID 5 or 6 for fault tolerance) rather than a single large drive. Factor in the parity overhead—RAID 5 on four drives gives you approximately 75% usable capacity. A four-drive array of 8 TB drives yields ~24 TB usable, not 32.

If the client wants to cut costs and asks about smaller drives with motion-only recording: → Show them the hybrid model instead. Continuous at 5 fps to protect against gaps, motion-triggered at full frame rate for evidentiary quality. It usually recovers 40–60% of the storage difference without the forensic liability of pure motion recording.

If you’re evaluating a cloud-hybrid VMS (Verkada, Eagle Eye) and the storage conversation comes up: → These platforms typically store a rolling buffer on-device (usually 7–30 days depending on the camera tier) and push events to cloud. The local storage spec is fixed in the camera hardware—you don’t right-size it the same way. Evaluate the cloud retention tier pricing separately; per IPVM and Security Sales & Integration coverage of cloud VMS cost modeling, per-camera monthly fees for extended cloud retention can add $5–$15/camera/month over baseline, which becomes meaningful math at 40+ cameras.

---

One More Thing Before You Order

Whatever number your sizing math produces, add 20% before you finalize the drive order. Scene complexity varies by season (summer foliage changes motion-detection hit rates), firmware updates sometimes change default encoding settings, and clients add cameras after the initial install far more often than they remove them. The drive that’s exactly right today is the drive that runs out of headroom in month eight. Surveillance-grade 8 TB drives are not an area where over-speccing by one tier breaks a reasonable budget—it’s the kind of decision that saves a callback.

Hardware recommendations for this use case are available in our [NVR Storage Drives buying guide]; look for surveillance-rated options in the 6–8 TB range as the practical starting point for most mid-size commercial builds.