How WhatsApp Sends Your Photos and Videos Without Sending the Actual File Through the Chat Message

How WhatsApp Sends Your Photos and Videos Without Sending the Actual File Through the Chat Message

When you send a 20 MB video on WhatsApp, the app doesn’t simply put that 20 MB file inside a chat message and push it to your friend. What happens behind the scenes is much more interesting.

Every day, billions of people use WhatsApp to send photographs, videos, voice notes, documents and other files. To the user, the experience looks incredibly simple:

Select a video → tap Send → your friend receives it.

But technically, several different systems work together to make that happen.

Your message travels through WhatsApp’s messaging infrastructure, while the actual media is handled separately as an encrypted object stored on WhatsApp’s media infrastructure.

This separation is important because it allows WhatsApp to handle enormous volumes of large files without turning every chat message into a massive data transfer.

More importantly, the media is encrypted before it leaves your device.

So, what actually happens when you send a video to your friend?

Let’s follow the journey.

1. It starts with your phone

Imagine you record a 20 MB video and send it to your friend Kofi.

Before the video leaves your phone, WhatsApp prepares it for transmission.

Depending on the media and your settings, WhatsApp may process or compress the file to make it more suitable for sharing. WhatsApp also supports different media-quality options, so the exact amount of compression can vary.

The important point is that your phone does not simply upload the original video as readable plaintext.

The media is prepared for encrypted transmission.

Think of it this way:

Your phone → prepares the video → encrypts the video → uploads the encrypted version

This is the first important security boundary.

2. Your phone generates media encryption material

WhatsApp uses end-to-end encryption to protect messages and media.

For large attachments such as videos, images, audio and documents, WhatsApp’s technical documentation describes the sender generating encryption key material specifically for the attachment.

The media is then encrypted on the sender’s device before being uploaded.

The result looks something like this conceptually:

Original video

VIDEO_DATA.mp4

↓

Encryption

AES / authentication / key derivation

↓

Encrypted media

RANDOM_BINARY_DATA.enc

If somebody obtains the encrypted object without the corresponding key material, they should not be able to simply open it as a normal video.

WhatsApp’s technical documentation describes encrypted media being uploaded to a blob store and a separate encrypted message containing the information needed by the recipient to retrieve and decrypt it.

3. What exactly is a blob store?

The term blob stands for Binary Large Object.

A blob store is essentially infrastructure designed to store large pieces of data such as:

• Videos
• Photos
• Audio files
• Documents
• Backups
• Other binary objects

Instead of putting a 20 MB video directly inside the messaging system, the application can store the large object separately.

The messaging system can then reference that object.

This is a common architectural pattern in modern cloud applications.

You see similar ideas in systems built around:

• Object storage
• Content Delivery Networks (CDNs)
• Distributed storage
• Cloud storage
• Media servers

In WhatsApp’s case, the important security property is that the media uploaded to the storage infrastructure is encrypted.

The storage system therefore does not need to receive the video in readable form.

4. Your phone uploads the encrypted video

Your phone now uploads the encrypted media to WhatsApp’s media storage infrastructure.

Conceptually:

Your phone

↓

Encrypted video

↓

WhatsApp media storage / blob infrastructure

The storage system holds the encrypted object.

Your phone then receives information that identifies or points to where that encrypted media can be retrieved.

This is sometimes described as a media URL, pointer, or direct path, depending on the implementation.

But there is an important distinction:

The pointer is not the video.

It is simply information that helps the receiving device locate the encrypted object.

5. Now comes the interesting part: the chat message

Your phone still needs to tell Kofi that you sent him a video.

So WhatsApp sends a message through its messaging infrastructure.

But the message doesn’t need to contain the entire 20 MB video.

Instead, the encrypted message can carry information associated with the media, including information needed to retrieve and decrypt it.

WhatsApp’s published technical documentation describes the media message as containing key material, a hash used to verify the encrypted blob, and a pointer to the blob store.

There may also be a thumbnail or preview associated with the media message.

So conceptually, the communication looks like this:

Encrypted chat message

• Media pointer
• Media encryption information
• Integrity information
• Media type
• Size and other metadata
• Thumbnail/preview information where applicable

The actual 20 MB video is not sitting inside that ordinary message payload.

6. Kofi receives the message

Kofi’s WhatsApp receives the encrypted message.

His phone can immediately update the conversation interface.

He may see:

🎥 Video

along with a thumbnail or preview.

This is why the chat can show you that someone sent a video even before the full video has finished downloading.

The small preview and the large media object have different roles.

Think of it like receiving a movie poster before downloading the actual movie.

7. Kofi’s phone retrieves the encrypted video

Now Kofi’s phone knows where the encrypted media is stored.

His device can request the encrypted object from WhatsApp’s media infrastructure.

The process is approximately:

Kofi’s phone

↓

Media request

↓

WhatsApp media storage/CDN

↓

Encrypted video

↓

Kofi’s phone

Notice something important here.

The storage system is still providing the encrypted object.

The receiving device is responsible for processing the encrypted media using the key material delivered through the end-to-end encrypted message.

8. The video is decrypted on Kofi’s device

Once Kofi’s phone has both:

1. The encrypted video
2. The required cryptographic information

it can verify and decrypt the media.

The WhatsApp technical documentation describes the receiving device verifying the encrypted blob and its authentication information before decrypting the plaintext media.

Conceptually:

Encrypted video

↓

Integrity verification

↓

Decryption

↓

Original video

↓

▶️ Play

This is why end-to-end encryption is so important.

The encryption and decryption operations happen at the endpoints rather than relying on the storage system to decrypt the media for the recipient.

The Complete Journey

Here’s the entire process in one picture:

Your phone
Record/select video
↓
Prepare/compress media
↓
Encrypt media
↓
Upload encrypted media
↓
WhatsApp media storage
Encrypted blob
↓
Return media pointer
↓
WhatsApp encrypted message
Pointer + cryptographic information + metadata/preview
↓
Kofi’s phone
Receive message
↓
Fetch encrypted media
↓
Verify
↓
Decrypt
↓
▶️ Watch video

Why Doesn’t WhatsApp Just Send the Video Through the Chat Server?

This is primarily an architecture and scalability problem.

Imagine WhatsApp has to process millions of simultaneous video transfers.

If every messaging connection had to carry the complete media payload, the messaging infrastructure would become much more difficult to scale.

Instead, the system can separate:

Messaging traffic

Small, latency-sensitive information such as:

• Text
• Media references
• Delivery information
• Message metadata

from:

Media traffic

Large objects such as:

• 20 MB videos
• 5 MB photographs
• Audio recordings
• Documents

This separation is a fundamental concept in distributed systems.

It allows different infrastructure to be optimized for different workloads.

Think of It Like a Courier Service

Here’s an analogy that makes the architecture easier to understand.

Imagine sending Kofi a large physical package.

You don’t put the entire package inside the text message you send him.

Instead, your message says:

“Your package is stored at Warehouse A. Here is the information needed to retrieve it.”

The messaging system carries the instructions.

The warehouse stores the package.

In WhatsApp’s architecture:

Chat system = message/instructions

Media storage = encrypted package

Encryption key material = information needed to unlock the package

This separation is what makes the system efficient.

What Does the Media Storage System Actually See?

This is where things become particularly interesting from a cybersecurity perspective.

Suppose the original file is:

KOFI_BIRTHDAY_VIDEO.mp4

After encryption, the storage system receives something more like:

8F3A91D7C...encrypted binary data...29BC

The storage system can store and deliver the object.

But it is not receiving the video as ordinary readable video content.

WhatsApp states that media stored on its servers for delivery purposes is stored in encrypted form and that WhatsApp cannot see the media itself.

This is a critical distinction:

Storage of encrypted data is not the same thing as storage of readable data.

But Does This Mean WhatsApp Knows Nothing About the Message?

No.

This is an important nuance.

End-to-end encryption protects message and media content, but it should not be interpreted as meaning that absolutely no information exists outside the endpoints.

Messaging systems still need to operate.

Depending on the service and feature, systems can process various forms of metadata and operational information.

For example, a messaging service may need information associated with:

• Account identity
• Device connections
• Delivery status
• Timestamps
• Group membership
• Network connections
• Service diagnostics
• Security systems

WhatsApp itself distinguishes between user content and other categories such as account information, connections, group information, device/connection information and troubleshooting information in its privacy documentation.

Therefore:

End-to-end encryption ≠ zero metadata

Instead, it means that the protected content is encrypted so that the service provider is not supposed to be able to read the message or media content itself.

What Happens When You Send a Video to a Large Group?

Now let’s make the example more interesting.

Suppose you send a video to a large WhatsApp group.

You might think:

“If there are 200 people in the group, WhatsApp must receive 200 copies of my video.”

That’s not the architecture you should imagine.

The media can be uploaded as an encrypted object, while the message is distributed to the intended recipients with the information needed for their devices to retrieve the encrypted media.

WhatsApp’s technical documentation describes mechanisms for efficient group-message distribution, including Sender Keys, while its media architecture allows recipients to access the same encrypted media object rather than requiring a completely separate uploaded copy for every recipient.

Conceptually:

Your phone

↓
1 encrypted media upload

↓
Encrypted media storage

↙ ↓ ↘

📱 Kofi
📱 Ama
📱 Kwame
📱 Adwoa
📱 …many other recipients

Each recipient can retrieve the encrypted media when needed.

This is a major bandwidth and infrastructure optimization.

Why Can WhatsApp Store the Media for Up to 30 Days?

This is one of the most interesting parts of the system.

According to WhatsApp’s current privacy documentation, when a user sends media in a message, WhatsApp may store that media for up to 30 days in encrypted form on its servers to help with efficient delivery, including when recipients choose to forward the media.

Why is that useful?

Imagine Kofi receives your video today but forwards it tomorrow.

If the encrypted media is still available, the system can potentially reuse the existing stored object rather than requiring the entire original media file to be uploaded from Kofi’s device again.

That can save:

• Upload bandwidth
• Download bandwidth
• Processing resources
• Storage operations
• Time

This becomes especially important when the same media is shared repeatedly.

But What About Forwarding?

Forwarding is where the architecture becomes even more useful.

Suppose you send:

Video A → Kofi

Kofi forwards it to:

Ama

Ama forwards it to:

Kwame

The underlying media infrastructure can make efficient use of the existing encrypted media object rather than treating every forwarding action as a completely new physical upload.

WhatsApp explicitly says it temporarily stores media in encrypted form to aid more efficient delivery, including additional forwards.

However, forwarding is also a major misinformation and abuse vector.

That’s why WhatsApp has introduced restrictions around forwarding certain messages and media.

The goal is to make mass distribution more difficult without preventing normal person-to-person sharing.

Why Does WhatsApp Have Forwarding Limits?

Imagine a piece of false information being forwarded to:

10 people

Those 10 people forward it to:

100 people

Those 100 forward it to:

1,000 people

And those 1,000 forward it to:

10,000 people

The problem is no longer simply technical.

It becomes a network propagation problem.

A messaging platform has to balance:

Fast communication

with

Abuse prevention

with

User privacy

with

System scalability

Forwarding restrictions are therefore one of several mechanisms WhatsApp uses to slow large-scale viral distribution.

Why Does WhatsApp Let You Download Media Only on Wi-Fi?

If you’ve ever configured WhatsApp’s media settings, you may have seen options such as downloading media only when connected to Wi-Fi.

Now the architecture makes that behaviour easier to understand.

The message can arrive even when you are on mobile data.

But the large media object doesn’t necessarily have to be downloaded immediately.

Your phone can receive the message and preview information while postponing the large media download until the appropriate network condition exists.

For someone with limited mobile data, this is extremely useful.

Imagine receiving:

100 videos × 20 MB

That’s:

2,000 MB

or roughly:

2 GB

of media.

You probably don’t want your phone automatically downloading all of that over mobile data.

The separation between messaging and media retrieval makes this kind of behaviour practical.

What Happens If Kofi Is Offline?

This is another important advantage of the architecture.

Suppose you send Kofi a video while his phone is turned off.

WhatsApp can retain the encrypted message temporarily while attempting delivery.

WhatsApp’s privacy documentation states that undelivered messages may be kept in encrypted form for up to 30 days while delivery is attempted.

Media is similarly retained temporarily in encrypted form for delivery optimization.

When Kofi comes online, his device can receive the message and retrieve the associated media.

The important concept is:

The sender doesn’t have to remain online until the recipient downloads the video.

The messaging infrastructure and media storage infrastructure help bridge that gap.

A Technical View of the Architecture

For engineers, we can simplify the architecture into several logical components.

1. Client application

The WhatsApp application on the sender’s device.

Responsibilities include:

• Selecting media
• Preparing media
• Encrypting media
• Generating cryptographic material
• Creating the media message
• Uploading encrypted media

2. Messaging infrastructure

Responsible for moving encrypted messages between clients.

It handles things such as:

• Message delivery
• Device connectivity
• Message routing
• Delivery state
• Group-message distribution

3. Media storage

Responsible for storing encrypted media objects.

Examples of objects include:

• JPEG images
• MP4 videos
• Audio
• Documents

4. Media delivery infrastructure

The recipient needs an efficient way to retrieve the encrypted media.

This is where media-serving infrastructure/CDN-style delivery becomes important.

5. Recipient client

The recipient’s WhatsApp application:

• Receives the encrypted message
• Extracts the media information
• Retrieves the encrypted object
• Verifies it
• Decrypts it
• Displays or plays the media

The Architecture in One Diagram

                SENDER
                   │
                   │
             Select video
                   │
                   ▼
          ┌─────────────────┐
          │ Prepare /       │
          │ Compress Media  │
          └────────┬────────┘
                   │
                   ▼
          ┌─────────────────┐
          │ Encrypt Media   │
          └────────┬────────┘
                   │
          Encrypted video
                   │
                   ▼
          ┌─────────────────┐
          │ Media Storage   │
          │ / Blob Store    │
          └────────┬────────┘
                   │
              Media pointer
                   │
                   ▼
          ┌─────────────────┐
          │ Encrypted Chat  │
          │ Message         │
          └────────┬────────┘
                   │
                   ▼
                KOFI
                   │
                   ▼
          Receive encrypted
             chat message
                   │
                   ▼
          Retrieve encrypted
                media
                   │
                   ▼
          Verify + Decrypt
                   │
                   ▼
              ▶️ VIDEO

This separation between the message plane and the media plane is one of the most useful ways to understand the architecture.

Message Plane vs Media Plane

For IT professionals, this distinction is worth remembering.

Message Plane

Deals primarily with communication between clients.

It is optimized for:

• Small messages
• Low latency
• Routing
• Delivery
• Encryption
• Synchronization

Media Plane

Deals with large binary objects.

It is optimized for:

• Large files
• High bandwidth
• Object storage
• Caching
• Content delivery
• Scalability

This is not unique to WhatsApp.

Similar architectural principles appear throughout modern distributed systems.

Video streaming platforms, cloud applications, social networks and collaboration platforms all separate different workloads so that each part of the system can scale appropriately.

Where Does Encryption Actually Happen?

This is perhaps the most important concept in the entire article.

Imagine three locations:

Your phone

→ WhatsApp infrastructure

→ Kofi’s phone

The desired security model is:

Your phone
   │
   │ ENCRYPT
   ▼
WhatsApp infrastructure
   │
   │ ENCRYPTED DATA
   ▼
Kofi's phone
   │
   │ DECRYPT
   ▼
Original media

The infrastructure in the middle can transport and store the encrypted content without needing to turn it back into the original readable video.

WhatsApp says its personal messages, photos and videos are protected by end-to-end encryption, meaning the intended participants — not WhatsApp — hold the cryptographic access needed to read the protected content.

What Happens If Someone Steals the Encrypted Blob?

This is where encryption provides its most important benefit.

Suppose an attacker somehow obtains:

encrypted_video.enc

Possessing the encrypted object alone should not be equivalent to possessing:

original_video.mp4

The attacker still needs the appropriate cryptographic material and a valid decryption process.

This is one reason encryption is so important for cloud storage and network communications.

A storage breach involving encrypted data is fundamentally different from a storage breach involving plaintext data.

That doesn’t mean encryption makes a system invulnerable.

Endpoint security still matters.

If malware compromises your phone while the video is being viewed, for example, the attacker may potentially access information at the endpoint.

This leads to an important cybersecurity principle:

Encryption protects data, but endpoint security protects the places where encryption begins and ends.

The Hidden Complexity Behind a Simple “Send” Button

When you tap Send, you are triggering a surprisingly sophisticated distributed workflow.

Something as simple as:

“Send this video to Kofi.”

can involve:

• Media processing
• Encryption
• Key generation
• Hashing
• Authentication
• Object storage
• Message routing
• Device synchronization
• CDN/media retrieval
• Decryption
• Local caching
• Network optimization
• Delivery acknowledgements
• Storage lifecycle management

And all of this normally happens in seconds.

That is what makes modern messaging applications such impressive pieces of distributed software.

A Few Important Misconceptions

Myth 1: “The video is sent directly through the chat message.”

Not exactly.

The chat message can carry the information needed to retrieve the encrypted media, while the large media object is stored and delivered separately.

Myth 2: “WhatsApp stores my video as a normal MP4 file that anyone at WhatsApp can open.”

WhatsApp states that media stored on its servers for delivery purposes is encrypted and that it cannot see the media itself.

Myth 3: “End-to-end encryption means WhatsApp knows absolutely nothing.”

Not necessarily.

End-to-end encryption protects the content of the communication, but services can still process metadata and operational information required to provide the service.

Myth 4: “If I delete the chat, every copy of the media everywhere instantly disappears.”

That is too simplistic.

There can be copies or representations on:

• Your device
• The recipient’s device
• Backups
• Temporary server infrastructure
• Forwarded messages
• Screenshots or recordings

WhatsApp’s disappearing-message and media features also have specific limitations.

For example, WhatsApp notes that certain view-once media may still be subject to recording or other endpoint behaviours depending on the device and feature.

The Bigger Lesson for Technology Professionals

WhatsApp’s media architecture demonstrates several important principles of modern system design.

1. Separate workloads

Don’t force a messaging system to perform the same job as a large-object storage system.

2. Encrypt at the endpoint

If the storage layer doesn’t need to understand the content, don’t give it the plaintext.

3. Use object storage for large data

Large binary objects are better handled by systems designed specifically for large-scale storage and retrieval.

4. Keep control information separate from payload data

A small message can contain information describing where and how to retrieve a large object.

5. Design for reuse

Temporarily retaining encrypted media can reduce redundant uploads and improve forwarding efficiency.

6. Design for unreliable networks

Users may have:

• Slow mobile connections
• Wi-Fi only
• Intermittent connectivity
• Limited data
• Offline devices

A good messaging system must account for all of them.

7. Security and scalability must coexist

A system serving billions of users cannot simply choose between security and performance.

It has to engineer both.

The Architecture Behind One Tiny WhatsApp Button

The next time you tap Send on a WhatsApp video, remember that you are not simply transferring a file from your phone to another phone.

You are interacting with a distributed system.

Your phone prepares and encrypts the media.

The encrypted media is stored separately.

An encrypted message tells the recipient how to retrieve it.

The recipient’s device downloads the encrypted object.

The recipient’s device verifies and decrypts it.

And the whole process happens while WhatsApp manages enormous numbers of users, devices, connections and media objects simultaneously.

The most fascinating part is that all of this complexity is hidden behind one very simple interface:

Send.

That is the real magic of modern distributed systems.

Final Takeaway

The next time someone asks:

“How does WhatsApp send a 20 MB video without putting the whole video inside the chat message?”

The short answer is:

It separates the message from the media.

The media is encrypted on the sender’s device and stored as an encrypted object. The encrypted chat message carries the information needed for the recipient’s device to retrieve and decrypt that object.

So the architecture looks roughly like this:

Phone → Encrypt → Media Storage → Encrypted Message → Recipient → Download → Verify → Decrypt → Play

And that simple sequence combines cryptography, distributed systems, object storage, networking, caching, content delivery and security engineering into one seamless user experience.

That is why a five-second tap on Send can represent an impressive amount of engineering underneath the surface.