AI & Altcoins Push BTC to the Backseat

Prices diverged sharply across crypto last week. BTC fell to 21-month lows, before recovering. Overall the sell-off was narrow, altcoins largely held their ground and in some cases gained. SOL was the standout performer, up over 12% since June 28. ETH and HYPE added 4.4% and 3.5%, respectively.

The decline was likely limited to BTC due to idiosyncratic factors such as Michael Saylor’s MSTR pivoting its treasury approach. Strategy will now be open to selling BTC to fund dividend repayments under its new Digital Credit Capital Framework.

This flexibility is practical and makes sense for the firm, but for diehards following co-founder Michael Saylor’s “never sell” mantra, it may be harder to accept. The market, however, is indifferent to ideology, and MSTR has risen rapidly this week, likely driven by the $1 billion share repurchase program and a more flexible approach.

Despite being back over $100 on Thursday, MSTR is still down over 27% since the beginning of June as BTC prices trend lower and the leveraged play on the digital asset looks less attractive to most.

BTC's trading activity told a different story to its price action this week. Volumes surged as prices fell, peaking on Wednesday as BTC ground toward $58.2k, then stayed elevated through Thursday even as the selling pressure eased. That pattern matters. It isn't a clean flush where volume spikes and resets. It looks more like sustained attention, traders watching closely and staying engaged rather than walking away once the low was in.

Alts didn't need anywhere near that kind of flow to move. SOL was up around 12% on the week but volume had fallen back to 65% of Monday's peak by Thursday. Whatever bid showed up for alts this week wasn't backed by proportionally larger flow, and that's not encouraging for sustained gains.

The order book tells the same story. BTC's book is over 50x deeper than HYPE's in absolute dollar terms, but the more telling number is width. SOL's visible depth spans roughly 7% either side of mid, versus BTC's 1.2%. What liquidity exists in alts is spread thin across a wide price range rather than concentrated near the touch. BTC, ETH, and SOL all showed bid-heavy skew at the weekly low, consistent with buyers stepping in. HYPE was closer to balanced.

Thin volume can carry a rally just as easily as it amplifies a selloff, it just takes fewer sellers showing up to stop it.

Memory had a bad week, and a lot of capital was paying attention in digital asset markets. Micron Technology (MU), one of the largest memory chip makers in the world, became a focal point of the wobble. After trading above $1.2k around June 23, it spent much of the following week drifting lower and fell below $1k by July 2. The move reflected a legal case against Micron and peers, plus a broader re-rating as new compression and efficiency advances raised questions about how much memory AI systems will really need in the future.

NVDA shed 6.0% over the same period, but the two names told different stories. MU is a pure-play commodity memory name. NVDA is broader AI infrastructure. The market separated them, and that separation is the point this week.

Across Binance equity perps from 28 June to 1 July, memory and semi names accounted for 63.2% of all activity, or $10.4bn of over $16bn. MU alone traded $3.4bn. NVDA, by comparison, did $230mn. The view on Trade.xyz was the same: MU, SK Hynix, SanDisk, and Roundhill’s DRAM were four of the top six instruments by notional, with the memory and semi category driving the bulk of trading volume on the platform.

Traders were selling memory specifically, and the distinction matters, HBM and GPU demand held while commodity NAND and DRAM did not. Open interest on Trade.xyz gives us the clearest read on positioning through the move. MU OI sat around $217mn, or 193k contracts, at the start of the window and was broadly stable through Sunday, by Monday afternoon it was over $260mn, or 246k contracts.

Another interesting note during the memory sell-off this week was that Trade.xyz saw more activity relative to its total volume than Binance, and this was the same the weekend before too. However, the weekend accounts for only 6% of overall weekly volume, while on Binance this is just 4.1%.

The weekend lull raises an uncomfortable question: if on-chain equity markets are open 24/7 but most activity clusters around U.S. hours and weekdays, what is the always-on infrastructure actually for? And what is the differentiator, given that the CME will soon offer 24/7 WTI futures, with more products planned?

Right now the answer is mostly optionality, as we saw at the onset of the Iran war. The ability to trade at 2am on a Sunday has value in a crisis, even if most participants never use it. But with weekend share sitting at 4-6% of weekly volume, the case for round-the-clock equity and traditional markets exposure rests almost entirely on tail scenarios. That may be enough for now, but that means there might be only a couple of winners, and as we’ve seen with HIP-3 deployers sunsetting recently, the market is unforgiving and TradFi is coming.

Prices might continue to be suppressed for some time, and that’s not just based on short-term liquidity trends or AI-driven tumult. The outlook in the options market is equally dreary. When we map delta against strikes, we get a rough read on the market-implied chance that BTC expires above each strike.

That probability generally falls as you move farther out along the curve. So at 0.5 delta, (or 50% implied-prob) the level where the market sees a coin-flip, rises steadily with time. At 24 July it sits at $62.1k, essentially in line with spot. By 25 December it's at $65.6k, $3.7k above spot. That widening gap isn't a bullish signal though, it's just the futures basis. Positive carry means the forward drifts above spot over time and these curves price off the forward.

The market isn't getting more optimistic across expiries, it's getting less certain. That uncertainty is what’s interesting here. Looking at the $70k strike for the 24th of July, the curve puts implied-prob at 9%. The December curve puts the same strike at 41%.

Equally the downside widens in the same way, although the distribution isn't symmetric. By December the implied probability of BTC finishing below $50k sits around 20%, versus roughly 4.5% on the 24 Jul curve. This isn’t a revelation, more time = more uncertainty.

This week showed how quickly the narrative can shift. BTC fell on idiosyncratic factors while alts held, memory sold off while GPU demand stayed firm, and altcoin rallies ran on thin volume. The common thread is a market that is rotating fast and punishing single-factor exposure.

As we saw with options pricing, the market is rife with uncertainty. By December the range of outcomes is very wide, and BTC options are pricing that uncertainty rather than any clear direction. That, alongside the broader macro headwinds we've seen over the past few weeks, points to a volatile second half of 2026.

There’s plenty on the calendar in the next few weeks too. US CPI for June drops on July 15, earnings season has kicked off, and the Fed minutes from the June meeting are due July 9. The Fed minutes could give more insight into potential hikes, and this will certainly weigh on BTC and risk assets.

Bottom line, any surprise in either direction on inflation or AI capex guidance in earnings will matter for the whole market.