Robinhood expects roughly $650 million from prediction markets this year. Bernstein thinks that reaches $1.7 billion by 2028, which would edge past the broker's crypto revenue for the same year. That forecast anchors this 24 hour crypto recap of a Monday session where bitcoin sat just under $80,000 and the more interesting news came from the edges of the industry.

Prices first, briefly

Bitcoin hovered around $80,000 all day and slipped under it at points as traders trimmed risk ahead of Wednesday's Fed decision. Ether was near $2,300, XRP around $1.28, SOL close to $101. I see no conviction in that tape. It looks like a market waiting on a calendar: the Fed, the ECB and BOJ rate decisions, and Coinbase earnings, all this week.

The stories below matter more than the candles.

Prediction markets edge toward the crypto line

Bernstein's analysts, in a forecast shared with Cointelegraph on Monday, put Robinhood's 2028 prediction market revenue at $1.7 billion. Crypto-derived revenue for that year is pegged at $1.6 billion. Prediction markets would then be about 18% of a projected $9.4 billion in total revenue.

A product that barely existed on the platform a couple of years ago is projected to outearn the asset class that made Robinhood a fixture of retail trading. The gap is $100 million, which is thin, and forecasts three years out carry wide error bars. I'd treat the crossover date as a guess and the direction as the actual claim.

The forecast landed a day before Robinhood's third-quarter report. Bernstein expects revenue up 13% from the prior quarter to about $1.5 billion, helped by equity and options volume, event contracts and a rebound in crypto trading. Strong event contract volumes tomorrow would make the 2028 thesis a little less speculative. Weak ones, and expect an argument about seasonality.

Musk and the neocloud squeeze

Elon Musk said on X on Sunday that lenders have mostly stopped financing companies whose revenue doesn't extend beyond AI data centers. SpaceX, by contrast, lined up $40 billion in debt for GPUs, thanks to its investment-grade standing and cash from Starlink and rockets. His forecast for the 200-plus specialist AI cloud providers: "guppy-level death and consolidation." Some will survive, he added, but not many.

Musk is no impartial observer, since he competes for the same chips and capital. He isn't alone, though. Gartner's own projection, issued this month, is harsher in number than in tone: by 2030, over four in five neocloud operators will no longer be standing as independent businesses. The stated reason is that payback on AI infrastructure spending is lagging the cost of building it.

Why does a crypto desk care? Because a slice of the mining sector has rebuilt itself around this business. IREN now rents computing power to AI customers. Core Scientific hosts data centers for neocloud clients, CoreWeave among them. A miner that pivoted to AI to escape halving math has swapped one cyclical bet for another, and this one depends on credit markets that, per Musk, are closing the door. Nobody has shown that any specific miner is in trouble. The exposure is real all the same.

China sets a date for its blockchain backbone

China's Central Cyberspace Affairs Commission, a body inside the ruling Communist Party that coordinates digital policy, released a new national informatization blueprint on Friday, covering the 15th Five-Year period. Its headline item for our purposes: a national blockchain network should be largely built by 2030.

The plan gets specific about geography. The Yangtze River Delta is expected to push blockchain deeper into trade and finance. The Chengdu-Chongqing region in the southwest is slated for large-scale deployment in manufacturing and agriculture. The target itself is old news: a national development plan released in March included the project, and a joint policy from party leadership and the State Council on Oct. 9 called for building it. Friday added a deadline.

This is infrastructure for state use, and public chains shouldn't read it as an invitation. Beijing's ban on crypto trading hasn't moved, and nothing in the plan suggests otherwise. I'd file it under enterprise and government ledgers, and under how seriously China takes owning the plumbing.

Who holds the coins nobody can name

The quietest item of the day may be the most useful. Galaxy Research's October snapshot, as CoinDesk reported, classifies about 11.1 million BTC as unattributed. That is some 55% of the roughly 20.1 million mined. For these coins, Galaxy found no link to a trading venue, a fund, a state, or any other identifiable holder.

Galaxy's head of research, Alex Thorn, took it as proof that self-custody is still widespread, writing on X that "maybe even most" bitcoin sits with individuals. Xapo Bank replied that retail has been in the market for over a decade. It's a tidy story: retail came first, institutions followed.

The caveat matters, though. Unattributed isn't the same as personally held. The bucket likely includes coins lost to forgotten keys or dead owners, early-era mined coins outside the Patoshi pattern, institutions Galaxy hasn't identified (funds, companies, OTC desks), and custody setups that scatter holdings over many addresses to resist tracing. So 55% is a ceiling on what retail might own, and nobody has measured it.

The named 45% includes exchanges, BlackRock-led ETFs, the U.S. government and the largest single holder, Satoshi Nakamoto's Patoshi-pattern stash at 1,123,658 BTC. For all the ETF flow headlines, the biggest holder is still a category called "don't know." I find that oddly reassuring and oddly unsettling in equal measure.

What to watch

Tuesday's Robinhood report is the first hard data point on whether event contracts are scaling the way Bernstein assumes. Wednesday's Fed decision will settle whether bitcoin gets its move back above $80,000 or keeps drifting. On the AI side, the next test is whether another neocloud, or a miner with AI exposure, reports trouble raising debt. Musk has said where he thinks the stress is. Lenders will say whether he's right.