- Solana is holding just above its latest weekly low after another sharp drop.
- The weekly chart still shows falling peaks and a price stuck below both long-term averages.
- Anza wants to double network capacity and make token transactions far lighter by the end of 2027.
- Most of the roadmap is still in development and not yet live on mainnet.
Solana (SOL) traded near $73 on Thursday, just above this week’s $69.93 low, after a slide that erased most of its September rally. The selling comes just as Anza, the core development firm behind Agave, Solana’s main validator software, used the Breakpoint 2026 conference in London to present a 2027 roadmap that aims to double the network’s compute capacity, a plan the market has so far ignored.
Every Rally Since 2025 Peaked Lower
On the weekly chart, SOL’s problem is a staircase of falling peaks. After the January 2025 top near $295, the token set a second high around $250 in September 2025, then rebounded to roughly $140 early this year and to about $120 in late September, with sellers stepping in at a lower price each time.
The most damaging rejection came in October, when SOL briefly traded above the 0.236 Fibonacci retracement at $115.76, drawn between the 2025 peak and this year’s low, could not hold it, and then broke through $87.94, the zone where the price moved sideways for most of the first half of 2026. That former floor is now the first ceiling any rebound has to clear.

The 50-week exponential moving average stands at $102.55 and the 200-week at $110.16, so SOL now trades close to 30% below its one-year trend. With the shorter average also sitting under the longer one, the $102 to $110 band is where a recovery would most likely run into renewed selling.
RSI Has Room to Fall Before Oversold
The weekly Relative Strength Index has dropped to about 38 from above 60 in late September. Buying pressure has faded fast, yet the reading is still above 30, the usual oversold threshold, so the indicator does not yet suggest the selling has run its course.
This week’s candle still has more than three days to run. A weekly close below $69 would put $60.13 back in view, the early-summer low that also anchors the Fibonacci grid, while holding the zone leaves room for a bounce toward $87.94 that would still sit inside the pattern of lower highs.
| Level | Price | What it means |
|---|---|---|
| 200-week EMA | $110.16 | Top of the main resistance band |
| 50-week EMA | $102.55 | One-year trend line, now pointing down |
| 0.118 Fibonacci | $87.94 | Old consolidation floor, first resistance |
| Weekly low | $69.93 | Support being tested this week |
| Fibonacci base | $60.13 | Early-summer low, next major support |
Source: SOL/USDT weekly chart, TradingView
What Anza Put on the 2027 List
Brennan Watt, Anza’s chief executive, presented the plan on November 17, and Solana highlighted it in an announcement two days later. The headline goal is to lift capacity from around 250 million compute units per second to 500 million by the end of 2027. Compute units, or CUs, measure how much computational work a transaction or smart contract consumes, so the target describes raw processing headroom.
After Alpenglow, @bw_solana laid out @anza_xyz roadmap for Solana in 2027
Performance: network compute capacity doubles, transactions cut 80% CU units for token accounts
Expressivity: vector processing for models running fully onchain, 64 → 96 accounts per transaction, read…
— Solana (@solana) November 19, 2026
For developers, the number of accounts a single version-1 transaction can touch rises from 64 to 96 with the Agave 4.5 release planned for this quarter, while legacy and v0 transactions keep the old limit. Anza also plans direct access to account data and vector processing for computation-heavy apps such as on-chain models, with no benchmarks published yet.
At the network level, Rotor is meant to replace Turbine, the mechanism that spreads new blocks across validators, and a proposed architecture called Constellation would let several validators propose transactions at once and write ordering rules into the protocol. Today a single leader per slot decides which transactions enter a block and in what order, so this part matters most to trading firms.
Aug 2026 – Agave 4.2 halves slot times to 200 ms
Oct 21, 2026 – Alpenglow targets ~150 ms finality, down from ~12.8 s
Nov-Dec 2026 – Agave 4.4 and 4.5, 96 accounts per v1 transaction
2027 – 500M CUs per second, Rotor, Constellation
Cheaper Token Activity Is the Bigger Win
Anza aims to cut CU consumption for transactions involving token accounts by 80%, so an operation burning 10,000 CUs today would need about 2,000. Because Solana caps each block by total compute units, the same space could fit roughly five times as many of those operations, from stablecoin transfers to DEX swaps.
On the fee side, the flat base fee of 5,000 lamports (a lamport is one billionth of a SOL) per signature stays the same, while priority fees that users add to get processed faster are priced per compute unit requested and could fall for those operations. Twice the CU capacity also does not mean twice the transactions per second, because transactions vary widely in how much computation they need.
Storage Rent Falls Before 2027 Arrives
The SIMD-0437 proposal is phasing in a roughly 90% cut in storage rent, from 6,960 lamports per byte toward 696, and Solana’s upgrade tracker lists it as partially activated, with Watt suggesting at the event that a further sevenfold drop could land within about two weeks. That shrinks the refundable deposit needed to open a token account, a cost apps often cover when they onboard new users.
Anza has not published a binding schedule for the larger 2027 items, and its own record gives a way to measure them as they come due. Its January 2026 roadmap, the first Watt published as CEO, set out slot times below 400 milliseconds and direct account mapping inside the Solana Virtual Machine; slots were cut to 200 milliseconds in August, while direct mapping is still listed as pending activation with no mainnet date.
Credit: Source link






