Ninety-Nine Failures and a Sold-Out Miner: What the 2026 Crypto Graveyard Says About the Cycle

Ninety-nine crypto projects have failed in 2026, according to blockchain research firm Root Data, as reported by Cointelegraph on 26 July 2026 UTC. The headline number is the kind of statistic that gets pasted into cycle-bottom threads without much follow-up. The more telling companion data point sat further down the same Cointelegraph wire: Bitdeer, the publicly traded Bitcoin miner, mined a record-breaking 990 BTC in June and sold every coin. Two pictures, same week.

Taken together, the threads describe a market that is not collapsing so much as sorting. Failures are concentrated where they have always been concentrated: in the token layer. Survivors are concentrated where the cash flow is. The interesting question is whether the gap between the two is widening into a permanent two-tier market, or whether the current failures are simply the next cohort of the same Darwinian churn that has been running since 2018.

The 99, and what they actually are

Root Data's count is a tally of token projects that have effectively failed in 2026, per Cointelegraph's 26 July 2026 UTC wire. The number 99 is striking only by virtue of the pace: at the seven-and-a-half-month mark, it puts the year on track for a final tally that exceeds the lower-bound annual readings seen in recent cycles. The pattern is familiar. A fundraising peak in 2021, treasury drawdowns through the bear market, and a residual layer of zombie tokens waiting for someone to flip the switch.

What the failure count does not capture is the dollar-weight of what is collapsing. The cited posts do not specify the market capitalisation of the projects in Root Data's 2026 tally, nor how it compares to prior years. The available source items do not break the 99 down by category, so this article cannot say whether the cohort is heavy with launchpad tokens, structured-yield wrappers, or something else.

The counter-narrative is that the headline rate is the surface of a deeper problem. Monexus assessment: the more useful figure would be the ratio of failed projects to total active projects, which the cited posts do not specify, and which this article has not independently calculated. The aggregate count of 99 could be either a healthy cleansing rate or a leading indicator of a worse quarter, depending on which denominator you trust.

Bitdeer selling the lot

Bitdeer's June numbers are the more consequential signal. Mining 990 BTC in a single month is a record for the firm, according to Cointelegraph's 26 July 2026 UTC wire. Selling all of it is the policy choice, not the business outcome. The available source items do not specify the motivation behind the sale. The most natural reading is the simplest one: a miner monetising inventory into a recovery, rebuilding optionality, and betting that future extraction will be cheaper in fiat terms than current extraction is.

The implication is not that Bitcoin is being sold. Of course it is, every day, by every miner, every exchange, every authorised participant. The implication is that the marginal public miner is willing to print zero net accumulation even at a record monthly production rate. That is the behavioural shift. Through the 2020-2021 cycle, miners were accumulator-proxies. Through 2024, they became flow-through vehicles. In 2026, the largest public miners are running closer to pure harvesters than to long-term holders.

A counter-narrative runs the other way. Bitdeer, like other public miners, is reporting in a quarter where investors reward discipline over HODL conviction. A miner selling into a price recovery is not necessarily a bearish signal on Bitcoin. It is a bullish signal on Bitdeer's cost of capital relative to its cost of production. Monexus analysis: the two readings are not mutually exclusive, and the data point on its own does not adjudicate between them.

The Apollo footnote, and what it isn't

Separately, Cointelegraph's 25 July 2026 UTC wire carried an Apollo research note flagged as "interesting": initial public offerings have underperformed the broader market since 2019. The reference sat in the crypto desk's source folder because the relevance is structural. The 2021 IPO cohort put Circle, Coinbase, Robinhood, Bakkt, and a roster of miners and custodians into public hands. The 2026 vintage is doing the same with stablecoin issuers, tokenisation platforms, and the next generation of Bitcoin treasury companies. The Apollo figure, if accurately relayed, suggests that the public-equity bid for exposure to crypto is competing with vehicles that have, on average, produced less return than the indices they were meant to beat.

The framing here is not that crypto IPOs are bad. The framing is that going public in this cycle has been a worse deal for shareholders than the alternative: staying private, listing via a SPAC, or vesting in venture. For the broader market, the structural implication is that the public-equity float of crypto-adjacent companies is being repriced lower even as the underlying asset class sits near cycle highs. The disconnect between the two valuations is the story, not the IPO count.

Where the cycle actually sits

Look at the three data points together: 99 failures, a public miner selling 100% of monthly production, and a referenced Apollo note that IPOs have underperformed since 2019. Monexus assessment: the pattern is not a bear-market signal. It is a maturity signal. The wash-out is happening at the layer where wash-outs always happen. The miners are running disciplined balance sheets. The public-market vehicles are being marked down for the privilege of being public. None of this is the shape of a market that is about to break; it is the shape of a market that is sorting itself into permanent structure.

The forward-looking question is the next cohort. If the 2026 IPO vintage performs the same way Apollo documents, the public-equity channel for crypto companies closes in 2027 and capital stays private. If the failure count at Root Data accelerates through the second half of 2026, the marginal token-launch economics tip negative. If Bitdeer and its peers keep dumping production into rising prices, the available float expands just as ETF demand normalises. Each of these is a distinct forward path. The cited posts do not specify which dominates, and this article does not have first-party data to forecast it. Watch the third-quarter Root Data count, the next Bitdeer treasury disclosure, and the first two post-listing quarters of the 2026 cohort. That is where the cycle will be written.

Desk note: the Cointelegraph wire carried the failure count and the Bitdeer sale as adjacent data points. Monexus treats them as a single signal about capital discipline, not as two unrelated headlines. The Apollo reference is included as a structural footnote, not as a primary claim about the crypto market.

Wire provenance

This editorial synthesis draws on the following public wire/social posts:

• https://t.me/cointelegraph/71268
• https://t.me/cointelegraph/71264
• https://t.me/cointelegraph/71245
• https://t.me/Cointelegraph/71268
• https://t.me/Cointelegraph/71264
• https://t.me/Cointelegraph/71245