JAM and the JAM Grid: The Subsequent Phases of the Polkadot Cloud

11 min readJan 22, 2025

The article was written by Christian Casini, a Permanence DAO member.

Note: This article is not meant to be technical. I did my best to abstract away complicated technical terms and prepare graphs with different performance metrics. If anyone believes some of the content or metrics need correction, please get in touch with me. I will be happy to correct the article.

Outline:

JAM: Unprecedented Scale and Computing Power
The JAM Grid
The Comparison Chart
What Applications Would be Possible With JAM?
What Applications Would be Possible With the JAM Grid?
How the JAM Grid Empowers Developers
How DOT Fuels the JAM Grid
Final Thoughts

1. Introduction

Polkadot aims to be the Web3 Cloud ( Polkadot Cloud) for building highly scalable and extremely resilient Web3 applications and services. Although there are differences, this is similar to what AWS does for Web2 applications and services.

Here is a brief timeline of the Polkadot cloud:

Polkadot Cloud — Genesis: May 2020: This was when Polkadot launched, and the first block was produced. It is when the Polkadot Cloud went “live.”
Polkadot Cloud — Milestone I (Parachains), November 2021: This was when the first cloud service was built. The first service allowed developers to build custom blockchains and connect them to the Polkadot Cloud for shared security.
First Cloud Service Deployed — Milestone II, December 2021: The first cloud service (parachains) went live, and different teams started to use it to deploy their parachains.
Polkadot Cloud — Milestone III (JAM), TBD, perhaps by the end of 2025: JAM, which stands for Join-Accumulate Machine, is a trustless supercomputer that forms the basis of a Web3 cloud.
Polkadot Cloud — Potential Milestone IV (JAM Grid), TBD: The JAM Grid will be a hypothetical cluster of interconnected supercomputers — each running the JAM protocol.

Today’s blockchains, even the “fast” ones, struggle with real-time, data-heavy workloads. By contrast, the JAM Grid could theoretically do one billion TPS, exabyte-scale storage, and hundreds of GB/s in bandwidth, which could open the door for Web3 services on par with modern cloud data centers.

Brief intro to JAM and the JAM Grid

JAM is a single supercomputer that enables people to build scalable and resilient Web3 applications and services. JAM’s primary goal is to provide a flexible and efficient framework for managing data and computations within a network. It aims to streamline data integration and maintenance while ensuring the network’s integrity and security. You can learn more about JAM in the gray paper and this presentation from Gavin Wood.

A new concept, the JAM Grid, proposes the next leap: a multi-supercomputer network that could reach 1 billion transactions per second (TPS), exabyte-scale data availability, and High-Performance Computing (HPC) level bandwidth.

Regarding the TPS metrics, it is important to note that in the medium article from Gavin Wood describing these metrics, the author stated that the compute power may be one quadrillion EVM equivalent gas per second. There is no universal “transactions per second” (TPS) figure for a given gas throughput because individual transactions can consume vastly different amounts of gas. A simple ETH transfer on Ethereum can use 21,000 gas, whereas a more complex smart contract interaction might use hundreds of thousands (or even millions) of gas. Here is a table with a comparison (it is important to note that all these are approximations).

The chart above shows that the upper bound for gas-heavy transactions is about 1 million TPS for JAM and 1 billion TPS for the JAM Grid. However, it is important to note that these are approximations.

2. JAM: Unprecedented Scale and Computing Power

What is JAM?

JAM, which stands for Join-Accumulate Machine, is a trustless supercomputer that forms the basis of a Web3 cloud. It is a new computational model and protocol designed to enhance Polkadot’s capabilities and address scalability challenges in blockchain technology. JAM will enable 1 million TPS, 2 petabytes of data availability, and 857 MB/s bandwidth.

Why Does it Matter

JAM is the next-step upgrade that would bring Polkadot closer to its original vision of being a permissionless world supercomputer.

How do other chains compare?

This scalability targets large-scale, real-time applications that traditional blockchains can’t handle.

3. The JAM Grid

What is the JAM Grid?

A hypothetical cluster of interconnected supercomputers — each running the JAM protocol — promising 1 billion TPS, 1 exabyte of data availability, and 600 GB/s bandwidth.

Why Does It Matter?

These numbers are orders of magnitude beyond today’s blockchain norms, hinting at a High-Performance Computing (HPC) like scale. If realized, it could handle global-scale applications while maintaining decentralized security.

How Do Other Chains Compare?

Projects like Solana, Aptos, Sui, and Avalanche push higher throughput than older blockchains but still don’t target exabyte storage or HPC-level bandwidth.

4. The Comparison Chart

TPS (Transactions Per Second): A measure of raw throughput — how many transactions a chain can handle in a given second. Higher TPS often means more scalable dApps and a better user experience.
Data Availability: The capacity of the blockchain network to store and manage data on-chain. It includes the total amount of data and its accessibility across all network nodes. Robust data availability ensures consistent access to and verification of transaction records, application states, and user data. It also upholds a blockchain’s integrity, security, and decentralized nature.
Bandwidth: Reflects network capacity — how quickly data can move across the system. For blockchains that aim to process huge volumes of transactions (and associated data), high bandwidth is essential to prevent bottlenecks.
Architecture: Highlights the design approach — single-chain vs. sharded or subnet-based, for example. Architecture influences TPS and how easily a blockchain can grow and evolve.

Now that we understand why these metrics matter, let’s see how the JAM Grid vision stacks up against other leading projects. (Important to note that many TPS figures are testnet or theoretical maxima — real-world performance may vary. The data availability and bandwidth metrics are also ranges/approximations since there aren’t exact figures. I did my best to present a fair and accurate comparison, but if you find inconsistencies, please email me. I am happy to update the article with updated information.

You can find the Google Doc that has the chart with the footnotes here.

Again, it is important to emphasize that metrics in the above charts are ranges/approximations. I encourage anyone to reach out to me if they have more accurate figures.

Key Takeaways

Exabyte Storage & HPC Bandwidth:
The JAM Grid envisions 1 exabyte of data and 600 GB/s throughput — on par with large data centers, not typical blockchains.
Contrast with Today’s Networks:
While chains like Solana, Avalanche, and Aptos boast higher TPS than older platforms, they cannot approach HPC-level data availability or sustained bandwidth.
Multi-“Supercomputer” Approach:
There will be multiple supercomputers, each running the JAM protocol. This creates a network where all the supercomputers can communicate with each other and potentially share resources. The way these supercomputers will be able to communicate will be one of the key challenges.

5. What applications would be possible with JAM?

With around 1 million TPS, 2 petabytes of data availability and 857 MB/s bandwidth, the JAM supercomputer brings a massive leap beyond most existing chains. Here are some hypothetical examples of what JAM could unlock:

(1) Real-Time Gaming & Virtual Worlds (City-Scale or Country-Scale)

Why Not Now? Current blockchains can’t handle frequent in-game state changes for thousands (let alone millions) of players.
Why JAM Helps: 1 million TPS means the fast-paced interactions of a city- or country-level virtual world could remain on-chain, reducing reliance on centralized servers.

(2) Real-Time IoT & Automation (Enterprise or City Scale)

Why Not Now?
IoT devices generate millions of events each second across city or industrial deployments. Most blockchains can’t ingest that many on-chain without massive congestion.
Why JAM Helps: ~1 million TPS throughput plus petabyte-level data availability means each device can reliably post updates on-chain at millions of events per second, perfect for a large enterprise or smart city. Hundreds of MB/s bandwidth ensures sensor data syncs quickly among global nodes — enough for city or enterprise scopes, though not the entire planet’s IoT devices.

(3) High-Volume Stablecoin or Payment System (Country-Level)

Why Not Now?
Traditional blockchains face network congestion and high fees under heavy load, stifling efforts at mainstream, on-chain micropayments.
Why JAM Helps: ~1 million TPS capacity (assuming simpler transactions) allows for a stablecoin or payment network servicing a country or regional economy. Fast throughput keeps fees low and transaction finality quick. Not quite the global scale of billions of daily micropayments, but enough to surpass most existing L1 limitations.

With JAM implemented, the Polkadot Cloud will greatly outperform many current L1s on throughput, storage, and bandwidth, allowing real-time, data-heavy dApps that are currently impossible on typical blockchains.

6. What applications would be possible with the JAM Grid?

Current blockchains, even the “fast” ones, aren’t designed for real-time, data-heavy workloads. By contrast, a JAM Grid supporting billions of TPS and exabyte-scale storage could enable:

Today’s blockchains — even the “fast” ones — struggle with real-time, data-heavy workloads. By contrast, a JAM Grid that supports billions of TPS, exabyte-scale storage, and hundreds of GB/s in bandwidth could open the door for Web3 services on par with modern cloud data centers. Below are some example applications that might finally be viable under these conditions:

(1) Massively Multiplayer Metaverse & Gaming

Why Not Now? Current blockchains can’t process the real-time updates of millions of concurrent players, let alone store large volumes of in-game assets, states, and logs on-chain.
Why JAM Grid Helps: With billions of TPS and colossal bandwidth, the JAM Grid could handle frequent state updates (e.g., player actions, world changes) all on-chain, removing the need for centralized game servers, exabyte-scale storage enables massive 3D asset repositories, character inventories, and item histories to remain verifiable and persistent.

(2) Real-Time IoT & Automation

Why Not Now? IoT devices generate constant data streams — temperatures, vehicle telemetry, sensor updates — often millions of events per second. No mainstream blockchain can ingest and verify that volume in near real-time.
Why JAM Grid Helps: High throughput and scalable data availability allow each device to record updates directly on-chain without overwhelming the network. High bandwidth ensures the rapid propagation of data across globally distributed nodes.

(3) Global-Scale Social Networks

Why Not Now? Social platforms see billions of daily interactions (likes, posts, comments, messages), requiring massive throughput and storage. Existing blockchains can’t handle such a firehose of activity on-chain.
Why JAM Grid Helps: Billions of TPS can accommodate near-instant posting, while exabyte storage can preserve rich media and user histories without relying on off-chain solutions. Built-in decentralization ensures no single company controls user data or moderation.

(4) Decentralized AI & Machine Learning at Mass Scale

Why Not Now? Training and inference for large AI models require massive computing and vast data sets, which are typically handled by centralized HPC clusters.
Why JAM Grid Helps: The JAM Grid’s HPC-like throughput and bandwidth can host distributed AI workloads on-chain, ensuring auditable data provenance, verifiable training results, and democratized access to HPC resources.

(5) Global Stablecoin Payment Systems

Why Not Now? Traditional stablecoins on L1 blockchains already face congestion and high fees under heavy load. Billions of daily micropayments (e.g., IoT microtransactions and routine purchases) are off the table.
Why JAM Grid Helps: High throughput + fast finality enable stablecoins to scale to global volumes without sky-high gas fees or bottlenecks. This lays the groundwork for universal, low-cost payment systems that are more open than any corporate provider.

In short, JAM’s extreme throughput and storage unlock new Web3 services that rival traditional cloud solutions in performance while retaining decentralized ownership.

7. How the JAM Grid empowers developers

While the previous section showcased examples of global-scale applications the JAM Grid could support, developers might wonder: How does this open new doors for me? Here are a few ways the JAM Grid transforms the building experience rather than just unlocking theoretical workloads:

Massive On-Chain Data Handling

Today’s Bottleneck: Even “fast” L1s and many L2s impose strict data or state limits, forcing developers to rely heavily on off-chain servers, IPFS, or private storage. This complicates architecture, adds trust assumptions, and limits transparency.
JAM Grid Breakthrough: With exabyte-scale data availability, developers can design truly data-heavy dApps where logic and large data sets reside on-chain. This ensures auditability and composability without gluing in an off-chain database.

Easier Path to Global Adoption

Today’s Bottleneck: Even successful Web3 projects often hit performance ceilings, alienating mainstream users when fees spike or throughput lags.
JAM Grid Breakthrough: By supporting billions of TPS, devs can aim for mass-market adoption — like genuinely mainstream social apps or multinational supply-chain solutions — without the system buckling under the load.

The JAM Grid doesn’t just enable significant, flashy use cases; it dramatically reduces the friction between an innovative idea and a performant, trustless solution. Suppose you’ve spent years hitting scaling walls or stitching together partial solutions. In that case, the JAM Grid’s HPC-level design opens a future where developers can stay fully on-chain, scale horizontally, and still rely on one shared security layer.

Think of it as the “cloud moment” for Web3 development: You code, you deploy, and the network (via Polkadot’s DOT-secured HPC layer) ensures performance and security — all while preserving the decentralized ethos at the heart of blockchain.

8. How DOT Fuels the JAM Grid

DOT is the lynchpin securing Polkadot’s entire ecosystem — from the Polkadot Cloud to each JAM supercomputer in the JAM grid.

Security via Staking: To secure the Polkadot network — and thus the JAM Grid — validators stake DOT. As the JAM Grid expands, more stake is required, potentially increasing DOT’s role as collateral securing these supercomputers.
Access & Participation: To build or deploy services on JAM-powered supercomputers, users and developers will interact with Polkadot’s infrastructure, paying fees or staking DOT. Essentially, DOT becomes the ticket to access the world’s best supercomputer (or grid of supper) computers for Web3.
Economic Alignment: If the JAM Grid thrives, it enhances Polkadot’s utility, reinforcing demand for DOT — creating a positive feedback loop between the supercomputers’ growth and DOT’s fundamental value.

By making DOT the asset that secures, accesses, and propels the JAM Grid, DOT becomes the core economic driver of a global-scale Web3 Cloud. That synergy positions DOT as more than just “another token” — it becomes the foundation for hosting and running the next generation of decentralized, high-performance applications.

9. Final Thoughts

Suppose the JAM Grid can truly demonstrate 1 billion TPS and exabyte-scale storage. In that case, it would prove that there is a way to build a genuinely decentralized infrastructure capable of handling global-scale demands. Rather than fostering a zero-sum contest, this vision highlights a collective push toward permissionless, globally accessible “supercomputers” governed by communities rather than tech giants.

These ambitious specs are a guiding star for what a genuinely Web3-native “cloud” might look like: globally accessible, economically incentivized, and free from centralized control. If the JAM Grid succeeds, it won’t just raise the bar for Polkadot — it will inspire the broader ecosystem to envision new possibilities for high-throughput, trustless computing. The ultimate goal is for billions of people to benefit from transparent, censorship-resistant services.

Call to action(s):

Check out the JAM website.
Follow Permanence DAO on Medium and Twitter