Anza26

Starting this year, I’ve stepped into the CEO role at Anza. I’ve spent my career building large-scale, high-performance systems and leading engineering organizations, and I’m excited to carry that focus forward as we enter our next phase.

Anza’s mission is to build and ship software that enables Solana to keep scaling without compromising on performance, decentralization, or security.

Last year, we laid out a clear mission: increase bandwidth, reduce latency, and do it without compromising liveness. We hardened the Agave validator, expanded adversarial testing, and experienced zero down time. We shipped the first block limit increases and maintained slot times under 400ms, unlocking levels of throughput never seen on Solana mainnet before. New and improved schedulers were deployed, gossip network overhead was slashed, and Alpenglow core has been running on test clusters for several months.

This year, that mission is evolving.

In 2026, we are doubling down on the performance and liveness mission with an increased focus on user and developer experience.

Below is a look at the major initiatives we’re shipping or driving to meaningful progress this year and how they fit together into a coherent future for the Agave client and the Solana network.

Alpenglow: The New Consensus Engine

Alpenglow is our long‑term consensus solution. It provides tighter timing enforcement than TowerBFT and leverages BLS cryptographic primitives to dramatically reduce finalization latency while preserving safety.

In 2026, our focus is on shifting Alpenglow out of development clusters and onto mainnet in Q3 by:

• Stress‑testing network‑fault and equivocation scenarios

• Hardening rewards and incentives mechanisms

• Tightening up performance and end‑to‑end latency budgets

Multiple Concurrent Proposers

MCP introduces a foundation shift to the Solana market structure and crypto as a whole. Today, ordering is decided by the consensus leader. MCP brings ordering in-protocol by enforcing it in the replay stage. Censorship resistance from multiple proposers ensures that this replay ordering cannot be bypassed through omission.

Why this matters:

• Multiple proposers break the single leader monopoly and increases censorship resistance

• Predictable ordering unlocks execution control possibilities for applications

In 2026, we’ll ship an initial MCP version focused on enforcing transaction ordering within a batch in-protocol.

Increase Bandwidth, Reduce Latency

IBRL is the umbrella term for improvements that remove hard ceilings imposed by today’s software and kernel paths.

This year’s key initiatives include:

• Rolling out XDP shred transmission by default to dramatically increase turbine bandwidth

• Increasing block limits to 100M Compute Units

• Enabling direct mapping to reduce memory copy costs inside the Solana Virtual Machine (SVM)

• Reducing slot times below 400ms

Together, these changes remove multiple independent bottlenecks across networking, execution, memory, and time.

Unblocking Transaction Ingress

As the network scales, artificial transaction sending limits become bottlenecks and constrain economic activity.

We’re raising transaction sending limits in Agave to better reflect modern hardware and software capabilities. This allows for more transaction ingress to everyone and will only throttle users based on stake weight during periods of global congestion.

Operator Experience

Validators are the beating heart of the Solana network. Too often, when defaults don’t deliver, operators are forced into out-of-protocol solutions. We are adding new capabilities to address some common problems.

Block Revenue Distribution (SIMD‑0123)

We’re improving how block rewards and fees are accounted for and distributed, increasing transparency and reducing operational ambiguity. This allows users to participate in block rewards directly and also allows validators to automatically redirect a percentage of rewards to designated partners such as infra providers.

Scheduler Bindings

This allows for a clean separation between transaction scheduling logic and packing mechanics. This is intentional groundwork ahead of MCP, not a replacement for protocol-level ordering.

Before MCP arrives, there is market incentive for operators to differentiate with unique scheduling rules. However, this imposes risk today because of complex interdependencies between these components that can lead to entire blocks being rejected by the cluster. This cleaner separation creates a lower risk sandbox for experimentation.

Developer Experience

Developers often feel protocol constraints long before users & operators do. In 2026, we’re removing some of the most limiting ones.

Rent Reduction

Lowering long‑term storage costs so applications can scale the number of users without prohibitive capital overhead.

Larger Transaction Sizes

Unlocking more expressive transactions and reducing the need for awkward, multi‑step execution patterns without atomicity guarantees.

These changes directly expand what’s possible to build on Solana.

Optimized Onchain Programs

Juicing performance and cutting costs when interacting with the Associated Token Account (ATA) program by moving to p-ATA.

Moving Forward

2026 is about setting the foundation for the next phase of scaling.

I’m excited to lead Anza through this phase alongside an exceptional team of engineers, researchers, and operators and to continue working closely with the core developers and broader Solana ecosystem to ship impactful, production-ready protocol improvements.

If you’re building, operating, or pushing the limits of what Solana can do, we’re building this future with you.