Earn 5.18% APY staking with Solana Compass + help grow Solana's ecosystem

Stake natively or with our LST compassSOL to earn a market leading APY

Jito Announces BAM Direct Bundle Ingest, Routing Solana Bundles Past the Block Engine's 50ms Hop

Solana ๐Ÿงญ Compass By Solana ๐Ÿงญ Compass

Jito's BAM now takes Solana bundles straight from searchers, skipping the Block Engine. Jito reports 16.9% more landed bundles; access is by whitelist only.

Jito Announces BAM Direct Bundle Ingest, Routing Solana Bundles Past the Block Engine's 50ms Hop
A tied bundle of paper slips dropping down a brass mail chute into a box printed with the BAM cube logo, beside a capped side pipe, on a dark green nautical chart.

Jito JTO$0.538+4.4% has announced Direct Bundle Ingest (DBI), a route that lets trading firms send transaction bundles straight to BAM (Block Assembly Marketplace) nodes without passing through the Jito Block Engine, and it has moved existing Block Engine traffic onto the same route. The BAM Team's post of 13 October 2026 says DBI "has been live on mainnet for an early cohort of searchers over the past month". The post gives no start date and no cohort size. Jito's X account published the announcement as an article at 20:20 UTC the same day.

Keep up to date with the Solana eco
Follow us on Google News

A bundle is a group of up to five transactions that execute together, in order, or not at all. Searchers are the automated traders, such as arbitrage and liquidation bots, who pay tips to have their bundles included in a block. BAM is Jito's block-building system: its nodes order transactions and bundles, then pass them to the validator producing the block. Jito's changelog of 9 October 2026 says BAM-connected validators hold 164 million SOL, 37% of Solana SOL$106.77-0.1% stake as of epoch 1050, so a change to the bundle route reaches validators holding more than a third of the network's stake.

How Direct Bundle Ingest changes the bundle route into BAM

Since BAM launched in September 2025, its nodes have depended on the Block Engine, Jito's older bundle service, to authenticate, simulate and filter bundles before BAM ordered them. Jito says that arrangement "added an extra hop of roughly 50ms". Solana's slot time target has been 200ms since 9 October 2026, so by our arithmetic the hop was worth a quarter of the time a validator has to build a block.

With DBI, a searcher submits a bundle directly to a BAM node over QUIC or JSON-RPC. The node checks every signature, then simulates the bundle and drops it if any transaction fails. Surviving bundles enter BAM's auction, ranked by total fees (priority fees plus Jito tips) relative to the compute they request. Winners go to the BAM validator currently producing blocks as a single batch, and if any transaction fails at execution the whole bundle reverts.

The second change needs nothing from users. Bundles and sendTransaction traffic that still arrive at the Block Engine "are now directly forwarded to BAM through DBI, forgoing the Block engine's 50ms auction", the post says. Jito does not publish a measured end-to-end latency figure for either route.

Jito's early results: 7.5% fewer bundles received, 16.9% more landed

Jito compares two periods of about two and a half weeks, 3 to 20 September and 20 September to 7 October 2026. According to its table, BAM received 7.5% fewer unique bundles in the second period (21.97 billion against 23.76 billion) and landed 16.9% more of them (124.3 million against 106.3 million). The end-to-end land rate rose from 0.45% to 0.57%, and total tips on landed bundles roughly doubled, from 14,279 SOL to 28,765 SOL.

Bundles received
21.97B
-7.5%vs 3-20 Sept, per Jito
Bundles landed
124.3M
+16.9%vs 3-20 Sept, per Jito
End-to-end land rate
0.57%
Landed bundle tips
28,765 SOL

A land rate of 0.57% still means that about one bundle in 175 reaches a block. Most of the rest fail before the auction: Jito reports that simulation success rose from 11.4% to 16.1%, and that the share of compute spent on failed simulations fell from 50% to 41%. Jito attributes the gains to "better filtering logic on Block Engine bundles", which leaves more capacity for bundles that can pay.

Jito also reports a 72% rise in average reward per landed bundle. That figure and the doubled tip total describe one change. Dividing each period's tips by its landed bundles gives 0.000134 SOL and 0.000231 SOL, the per-bundle averages in Jito's table.

What the before-and-after comparison leaves open: filtering and a rising Solana tip market

Every figure comes from Jito's own before-and-after comparison, and the post does not separate bundles sent directly through DBI from those forwarded by the Block Engine. Jito does not say on which date either change took effect, and it describes the direct senders as "only an early cohort".

Routing is also one of several things that changed. The 9 October changelog credits "pruning of low-quality Block Engine search keys", a filtering step, with raising BAM tips per block by 15.5%.

The tip market was also growing across the whole network. Solana Compass data shows tips paid to Jito tip accounts on all validators came to about 21,750 SOL from 3 to 19 September and about 34,940 SOL from 20 September to 6 October 2026, a rise of roughly 61% between two 17-day windows that approximate Jito's. Our series counts every tip to a Jito tip account before Jito's fee, on BAM and non-BAM validators alike, and Jito's post does not define its BAM tip total, so the two totals are measured differently and only the direction compares.

Daily tips to Jito tip accounts, all Solana validators (SOL), 3 Sept to 6 Oct 2026

Network-wide Jito tips averaged about 1,280 SOL a day from 3 to 19 September 2026 and about 2,055 SOL a day from 20 September to 6 October, per Solana Compass.

View on Solana Compass โ†’

DBI access is by whitelist, and Alpenglow Maintenance Mode pauses it

Searchers and applications request DBI access through a form linked from the post, and "once whitelisted, you can start sending immediately". Jito publishes no criteria for approval. Its stated next steps are smarter filtering of low-quality bundles, further spam controls, and capacity for a larger cohort, "eventually enabling searchers and applications who mainly rely on the Block Engine to fully transition to DBI on BAM leader slots".

Jito's case for validators is a loop: more direct flow raises rewards for BAM validators, which draws more stake to BAM, which gives searchers more BAM blocks to target. The post presents that as the expected direction and offers no stake figures for it yet.

The announcement does not mention Alpenglow, the consensus upgrade now dated for Solana mainnet, which Anza says is scheduled to go live on 21 October 2026 at about 14:50 UTC. Jito's BAM readiness post of 5 October 2026 says Jito plans to put every mainnet BAM node into Maintenance Mode in the days before activation, that DBI will be unavailable for the duration, and that DBI users should "use the legacy Block Engine bundle-submission path". Neither that post nor the 9 October changelog gives a start or end time for the maintenance window. Jito says BAM nodes will come back "gradually" after activation, with timing to be announced on its Discord.

Solana ๐Ÿงญ Compass
Solana ๐Ÿงญ Compass
@SolanaCompass

Solana Compass is an independent Solana analytics and staking platform, operating a validator on Solana mainnet since September 2021. Its network statistics and...


Comments

Please login to leave a comment.

Related tokens Open token โ†’

Solana tokens

Solana Token Markets

Explore all tokens โ†’