The question “what crypto should I buy” has a useful answer and a useless one. The useless answer is a ranked list of tickers, which is what most of the internet will hand you this month. The useful answer is that a crypto allocation has two halves that behave nothing alike, and almost every mistake people make here comes from managing them as if they were one thing.
This page is updated monthly. It is not a set of recommendations, and it deliberately contains no price targets — for reasons the last section explains.
The market you are actually buying into
“Crypto is back” is half true, and the half that is false is the expensive half.
August was genuinely strong: US spot Bitcoin ETFs took in $3.52 billion, the best month of the year, and price followed. September opened with a $236 million outflow before flows turned positive again in the first week. Both facts are true, and neither is a trend yet.
- Spot Bitcoin ETF inflows, August
- $3.52bn
- Spot Bitcoin ETF outflow, early September
- −$236m
Millions of dollars of net flow. The second bar is an outflow, drawn by size rather than direction. Flows turned positive again in the first week of September; one month is a data point, not a trend.
That shape — a violent recovery month inside a flat-to-down year — is the one most likely to be misread. It is what a bottom looks like. It is also what the middle of a long drift looks like, and you cannot tell which from the chart.
A 25% month does not tell you whether you are early or late. It tells you volatility is back.
The boring half
“Boring” is the honest word, because “safe” is not on offer: this is an asset class that routinely halves. What the first half means is: assets that can lose 60% without the position becoming unrecoverable or unsellable.
Three tests, and they are all structural rather than predictive:
Can you custody it properly? A regulated spot ETF, or a hardware wallet you control, or a custodian with real segregation. If the only way to hold it is leaving it on an exchange account, that is a second risk stacked on the first.
Is there liquidity on a bad day? Every asset is liquid in a rally. The question is the spread and the depth during a 30% drawdown, when you might actually need to act.
Can you state the thesis in one sentence without using the word “potential”? Bitcoin’s is straightforward: a fixed-supply bearer asset with an institutional access route that did not exist three years ago. Ethereum’s is that it is the settlement layer most real on-chain activity and most stablecoin volume runs on. You do not have to agree with either sentence. You have to be able to say it, and then notice if it stops being true.
Two assets clear all three comfortably. Two more clear the first and are worth naming, because the test is only useful if it is applied honestly when the answer is inconvenient.
Bitcoin, around $78,300. The thesis: a fixed-supply bearer asset with an institutional access route that did not exist three years ago. What would break it: sustained ETF outflows, which is the one thing that route changed and the one thing that can reverse. It rose 24.95% in August and is still down on the year — that combination is the whole argument for owning it as a position rather than as a trade.
Ethereum, around $2,470. The thesis: the settlement layer most on-chain activity and most stablecoin volume runs on. What would break it: that activity moving somewhere cheaper and staying there. Watch settlement volume, not price.
Solana and XRP now pass the custody test too, and I had left them out. Both have spot ETFs with around $1.4bn in assets and — the part that matters — cumulative net creations of roughly the same size, meaning real money went in rather than an old trust being relabelled. Eight issuers compete on Solana. That is a genuine access route, and the reason it is worth saying is that the same test disqualifies Chainlink, Litecoin, Avalanche and Dogecoin, so it cannot be quietly skipped for the two it passes.
What neither they nor Bitcoin and Ethereum have been tested on is the second limb. Quoted spreads of 0.07–0.08% are rally-period figures, and no issuer publishes depth on a bad day. Nobody can run that test from outside, including me, and an article implying otherwise would be doing the thing this one is supposed to be against.
You do not have to agree with any of these sentences. You have to be able to say them, and then notice if one stops being true.
What is actually moving, and whether that is a reason
The honest answer to “what is trending” this month is one asset, and it is not close.
Zcash, around $1,150, up more than twentyfold over the year and at its highest since 2016. Three things happened in sequence rather than at once: the SEC closed its multi-year investigation of the Zcash Foundation in January without enforcement, removing an overhang; the Ironwood upgrade shipped on 28 July; and Grayscale listed the first US spot ETF for a privacy coin on 25 August.
The float is the part usually cited and it is more interesting than the headline. The share of supply held in shielded addresses is 28.8%, against 23.1% a year ago and 7.6% five years ago — a real multi-year trend that removes coins from the order books. But it peaked near 31% in May and has drifted down since, so the float is not thinning this month, and anyone using shielding to explain this month’s price is reaching.
A thinner float amplifies buying and amplifies selling by exactly the same mechanism. A twentyfold move is not evidence that the next one is up.
Nothing else has a story this month, and that is information too. Bitcoin is recovering from a weak year rather than breaking out, Ethereum is flat, and the rest of the large caps are trading on the same macro news as everything else — which is the ordinary condition, and a perfectly reasonable one to buy into slowly. A month with one genuine story is a month with one thing to think about.
Which of them is moving today is not a question a page like this can answer: it is rebuilt occasionally and the tape changes by the minute. Our crypto heatmap shows the whole set live and updates while you have it open. Use it for what is moving; use this for whether that should change anything.
The lottery half
Meme coins are not a smaller version of the first half. They are a different activity, and treating them as “the risky part of my portfolio” rather than “the part I expect to lose” is the specific error this section exists to prevent.
4.55%
The share of tokens launched on Pump.fun between January 2024 and June 2026 that were still trading after 90 days — 850,000 out of 18.67 million. Nearly 69% never traded past the day they were created. That is the base rate your pick has to beat.
Two things have genuinely changed in this corner in 2026, and they cut in opposite directions.
The upside, in the order it actually happened: Grayscale listed the first US spot Dogecoin ETF on 24 November 2025, 21Shares followed on Nasdaq on 22 January 2026, and only then — on 17 March 2026 — did a joint SEC and CFTC interpretive release name Dogecoin as one of sixteen digital commodities. Shiba Inu is on the same list. The ETFs did not follow the classification; they preceded it by months, which is a different and less tidy story than the one usually told.
The classification is the more interesting half. The same release treats meme coins as digital collectibles and names one as an example — so Dogecoin and Shiba Inu were sorted out of the meme coin category on functional grounds, not into it.
The downside: those flows are tiny. US spot Dogecoin ETFs showed a $419,000 net inflow on 3 September against a negative month overall. For comparison, Bitcoin’s ETFs moved $174.6 million in a single day that same week — roughly four hundred times the size. The institutional wrapper exists; the institutional demand largely does not.
- Bitcoin spot ETFs, 4 September
- $174.6m
- Dogecoin spot ETFs, 3 September
- $0.419m
Net inflows on single trading days, not directly comparable periods. The lower bar is drawn at a minimum width so it stays visible; the true ratio is roughly four hundred to one.
So the honest framing for this half: an ETF listing legitimises the access, not the asset. If you want exposure here, the correct question is never “which one” first. It is “how much can this be and still not matter if it goes to zero” — and that is an arithmetic question with an actual answer.
Our position sizing formula is the tool for it. The discipline that makes the lottery half survivable is that you size it at total loss, not at a stop level, because there frequently is no exit at the price you planned on.
How much of each — the real answer
This is the part the ranked lists skip, and it is the part that determines your outcome far more than your ticker selection does.
| Layer | What it holds | How to size it |
|---|---|---|
| Core | Assets passing all three structural tests | An amount you could watch fall 60% without changing your plan |
| Satellite | Larger alternatives with a real thesis and real liquidity | Small enough that being wrong on any one is an annoyance |
| Lottery | Meme coins and anything pre-revenue | Sized at total loss. If zero would hurt, it is too big |
The single most useful rule: decide the three numbers before you choose any asset. Choosing the asset first and the size afterwards is how a lottery position quietly becomes a core one, which is the mechanism behind most of the painful crypto stories you have heard.
Buying in stages matters more here than in equities, because the volatility that makes entry timing feel important is the same volatility that makes it unknowable.
Three catalysts in six days
An unusual cluster this month, and worth knowing before rather than after:
| Date | Event | Why it matters here |
|---|---|---|
| 11 September | US CPI report | The inflation print that shapes rate expectations five days before the Fed |
| 15 September | CLARITY Act cloture vote | The Senate’s procedural test for the crypto market structure bill |
| 16 September | Federal Reserve policy meeting | Rate decision; crypto has traded as a rate-sensitive risk asset all year |
The middle one is the crypto-specific one, and its effect is widely misunderstood — the bill classifies assets by a maturity test rather than picking winners, and the vote on the 15th is procedural rather than final. We covered the mechanism and who it actually helps in which crypto benefits from the CLARITY Act.
A cluster like this is a reason to have decided your sizing in advance, not a reason to trade the events. Positioning into a binary you cannot forecast is a coin flip with extra steps.
What this page will not do
No price targets. Nobody publishing a September target for any of these has information that supports one, and the numbers in the ranked lists exist to make an article feel actionable rather than because they mean anything. And no view on your entry: that depends on your tax position, your time horizon and what else you own, none of which is on this page.
Educational investment analysis, not personalised financial advice. Crypto assets are volatile and can lose all their value. Nothing here is a recommendation to buy, sell or hold any asset.