It has been around a year since the project started, and I would like to share a brief summary of what has happened so far, where the project is at and what’s coming next. I’ll only capture the highlights and key outcomes here, presented in rough chronological order.
December 2024#
In December 2024 I have officially transitioned out of Autonomys Labs (previously known as Subspace Labs) and started working on “Project Abundance” as an independent R&D project with the sponsorship from Subspace Foundation (for which I’m very grateful).
The first commit in the repository was made on December 27, 2024.
January 2025#
I had a conviction that the way smart contracts development is done in existing blockchain ecosystems is suboptimal and is not designed for performance. So my first goal was to think from the first principles what a good, high-performance design could look like.
I researched existing blockchains and did a bunch of prototyping and experimentation. By the end of the month, contracts were almost running with a lot of initial infrastructure and some early documentation available. There was no execution environment available yet, there was definitely no blockchain, but it was sufficient to get an idea of what it would look like to write contracts for a future system, what benefits, limitations and challenges it would have.
February 2025#
At the beginning of February I finally implemented the first version of the native execution environment, which
finally allowed running basic contracts with some basic documentation for the #[contract] macro.
Same month I was talking to some developers to collect initial feedback about the contract design, which helped with the APIs, documentation improvements and generally was a great opportunity to get some external feedback.
I also spent some time tuning the performance of the core data structures related to execution of contracts to get an idea of the low-level overhead required before any contract logic even begins executing, which turned out to be tiny, especially for the native execution environment.
March 2025#
In March, with contracts being somewhat usable, I went to the next layer in the stack: transactions. I spent a lot of time thinking about how transactions should work in a sharded blockchain and came up with a generic design, where the verification of transactions is done by a “wallet” contract. This way, there is a huge design space for implementation of contracts, with the blockchain itself not being aware of the signature verification logic or similar concepts, usually hardcoded into the core protocol of other blockchains.
This was also the month when Subspace Foundation sponsored Alfonso to help me with research of the sharding design.
By the end of the month there was a way to build contract files in ELF format. We also had productive discussions about the way hierarchical sharding might work with Proof-of-Archival-Storage consensus.
April 2025#
April continued with the sharding design research and preparations towards having an actual blockchain one day. Eventually I started seeing trees everywhere, and I must admit, I continue seeing them everywhere to this day.
By the end of the month I have refactored and massively improved the performance of some components forked from Subspace.
May 2025#
In May, I continued tinkering with forked Subspace components with the goal of shaping them closer to what I wanted and not depending on Substrate since I was not planning to use it.
A big and nice accomplishment of this month was the user-readable formatting for contract addresses. It was a lot of research and is inherently linked to the sharding design, I’m quite happy with the way it turned out.
By the end of the month, Alfonso also started seeing trees everywhere in the design. We were also working on a way to organize the hierarchical shards in a way that allows having a reasonable confirmation of blocks and archived history, such that we can actually have efficient cross-shard communication.
As a result of a lot of brainstorming and all the discussions, I finally designed a shard-native block structure for the beacon chain, intermediate shards and leaf shards, such that they form a hierarchical 3-level tree. It was a long and tedious process, but crucial to any progress down the line.
June 2025#
In June, we discussed how to assign farmers to shards, although the decision and especially implementation materialized much later.
I started making steps towards having initial block production and stumbled upon many opportunities for performance optimizations, which was quite rewarding to explore and continues to be the trend to this day. Modern hardware can do so much when software is designed with performance in mind!
We explored the topic of sector expiration extensively, but it has proven to be hard to come up with something robust that is also reasonably simple to understand and implement.
July 2025#
July marked the start of a lengthy process of tinkering with an awesome rust-gpu after exploring a few alternatives that ended up being much less awesome. Performance improvements continued to pour in from various fronts, making everything much faster still.
Exploration of sector expiration and a way to commit farmers to shards continued, though nothing better than the previous design was ultimately found.
I continued building foundational components like Sparse Merkle Tree and client database.
Unfortunately, this was the month when Alfonso’s contract was terminated by Subspace Foundation, so I was the only one working from this month onwards.
August 2025#
In August, I finally managed to put together a very custom client database prototype and worked on transaction processing design. I even created a node prototype, finally, even though it wasn’t capable of doing much just yet.
As I was tinkering with rust-gpu all this time and working on implementing sector encoding on the GPU, I kept discovering ways to make it faster on the CPU too. In the end, CPU plotting is more than 2x faster than Subspace repo at the time and even today.
September-October 2025#
I think it is safe to say that I spent most of September and October on GPU programming. I learned a lot more ways still to accelerate Proof-of-Space both on CPU and GPU. Ultimately, I was able to get GPU plotting to work.
November 2025#
Then there was almost a month of work gap as I was trying to get GPU plotting to work at a reasonable speed since the initial version wasn’t very fast. I didn’t post any updates during this time as it was probably not as interesting to anyone following.
What was interesting is that by the end of November I reached a key milestone, the node prototype mentioned before started producing beacon chain blocks!
Unfortunately, this was also the month when Subspace Foundation sponsorship ended for me too, so at this point I’m alone and living off my savings.
December 2025#
The first produced block achieved in November was nice. However, there were plenty of errors shortly after that, so my December started with fixing a bunch of things to the point that I was able to produce 10k blocks, which involved more consensus processes like solution range adjustment, Proof-of-Time entropy injection and block archiving.
I then started taking steps towards multi-shard farming and implemented a permissionless shard assignment designed back in June.
By the end of the month I decided to work on something more rewarding again related to contracts. After some research and experimentation I landed the initial version of RISC-V interpreter, CLI for building contracts and defined a contract file format no longer based on ELF, but one that can be converted from (and eventually to) ELF.
Today#
And today is January 1st 2026, so I can’t tell you how this year ends until we’re through. But I am sure there will be key developments with contract execution, I expect block production of intermediate shards and leaf shards at some point later this year, and maybe we’ll even get some networking between the nodes to sync from each other in a small devnet.
Definitely a lot of exciting stuff ahead, can’t wait to see where we end up one more year from today.
Acknowledgements#
I’d like to thank again the Subspace Foundation for their financial support, Alfonso for helping with research, Shamil and Liu-Cheng for fruitful developer feedback and everyone else I had a chance to discuss things with in some capacity.
I also want to thank Zulip for providing free access to “Zulip Cloud Standard” and Graphite for free access to their AI code review service.
And certainly to Subspace Labs (now Autonomys Labs) for original research and implementation of the Subspace protocol, none of this would have been possible otherwise, especially to Jeremiah and Dariia.
If you want to find me, Zulip Chat is the best way to start a conversation on topics related to “Project Abundance.”