TLDR:

• Dragonfly’s Qureshi compares today’s AI agents to the 1964 mouse, warning adoption will take far longer than expected.
  
• OpenClaw remains buggy and unreliable for financial tasks as models operate outside their training distribution today.
  
• The x402 protocol processes only around one million dollars daily, confirming the market is still in its tinkering phase.
  
• Qureshi expects a new model generation within months, but says reaching the early majority will still take several more years.

Agentic payments are gaining momentum as a talking point across crypto and fintech circles globally. Yet a senior voice from one of crypto’s most recognized investment firms is urging caution on timelines.

Haseeb Qureshi, a managing partner at Dragonfly Capital, recently shared what he called his “most bearish take” on the subject.

While he believes agents will eventually reshape how money moves, he argues the technology remains far from ready for mainstream use.

Dragonfly’s Qureshi Points to History as a Cautionary Benchmark

Qureshi grounded his warning in a well-known piece of technology history. He referenced the computer mouse, which was first invented in 1964, as a parallel to today’s AI agents.

That invention clearly pointed toward mass personal computing, yet widespread adoption took many additional years. His point is that spotting a transformative technology early does not mean it arrives on schedule.

OpenClaw sits at the center of his current skepticism about agentic readiness. The Dragonfly executive described the tool as buggy, complicated, and unfit for managing real financial assets.

It regularly makes poor decisions and, in his words, “goes bankrupt doing stupid shit.” These are not minor rough edges — they reflect a structural gap between agent capability and real-world task demands.

The core problem, according to Qureshi, is that current models are handling tasks well outside their training distribution. That mismatch produces the erratic and unreliable behavior users routinely encounter.

No major lab has yet applied reinforcement learning directly to OpenClaw interaction traces. However, those traces carry strong training signal that labs have not yet tapped.

Once a lab trains purpose-built models on agentic task data, a major performance improvement is expected. Every major AI laboratory is working toward this, Qureshi noted, because the commercial prize is clearly visible.

That model release will likely arrive within months, not years. Still, even that milestone will only mark the close of the tinkering era, not the start of mass adoption.

Live Payment Data Backs the Dragonfly Partner’s Cautious Stance

Qureshi pointed to real protocol data to support his position on where the market currently stands. The x402 protocol is processing roughly one million dollars in daily volume at present.

The Machine Payment Protocol is recording even smaller figures than that. Together, those numbers confirm the current user base consists almost entirely of early experimenters.

The Dragonfly executive also drew on a widely cited framing from investor Chris Dixon. The idea is that what technically curious people do on weekends today, the broader public will be doing within ten years.

That pattern has played out consistently across major technology waves, from the internet to mobile. Agentic payments appear to be sitting at the very beginning of that same cycle.

Qureshi mapped out the full adoption curve to give context to what comes next. After the tinkering phase closes, the market enters early adopter territory, which itself will take time to mature.

The early majority follows that, and then comes the late majority and eventual late adopters. Each phase carries its own timeline, and none of them collapse quickly.

For now, the Dragonfly partner sees agents as a long-term story that the industry should not rush. The technology direction is clear, and the destination is not in question.

What remains uncertain is how long each phase of adoption will actually take. That uncertainty, he argues, is precisely what crypto has a habit of underestimating.

We began our new Outset Data Pulse analysis expecting 12 years of headline data to confirm a familiar belief in crypto: that news moves markets, and that faster headlines give you an edge.

But what the findings showed instead was more unsettling: most of the time, price seems to move first, and the headline comes later to explain it.

That’s not to say that “news doesn’t matter.” It’s closer to saying we’ve been treating it as the trigger when it often behaves more like the explanation after the move. And it’s easy to see why that belief survived for so long. 

Anyone who spends enough time around crypto starts to notice the same thing: something moves, the news feed lights up, and then the dots get connected. When Bitcoin dumps or soars, coverage multiplies. When a major decision hits, whether it’s an ETF approval, an exchange collapse, or a legal victory, headlines also explode.

But the part of that belief which really matters – the part that turns news into a tradable edge – is directional. If headlines genuinely cause price movement, then reading faster makes you earlier. If price movement causes headlines, then reading faster mostly just makes you better informed about what already happened. 

That was the real question here: not whether news exists in the sequence, but whether it consistently comes early enough to matter in the way traders often assume.

The part where the data got harder to argue with

The core dataset powering this Outset Data Pulse report includes 63,926 CoinDesk headlines spanning January 1, 2014 through December 30, 2025, matched to daily Bitcoin closing prices from the TradingView composite index. 

That gave us 4,381 days where both a closing price and a headline count were available – enough to test the relationship from several angles, including causality, price behavior around major news spikes, headline sentiment, and topic clustering on the busiest coverage days.

It is also broad enough to cover nearly every “surely news mattered there” event worth testing, including bull and bear cycles, the FTX implosion, the COVID crash, and the start of the spot Bitcoin ETF era.

News volume didn’t forecast price

One of the first things we looked at was whether yesterday’s information helps forecast today’s movement.

We inspected five time horizons, from one day out through five days out. What kept standing out was that the news did not predict Bitcoin’s price across those lags.

Then there’s the kind of number you can’t really argue with because it’s too small to appear important: the correlation between daily changes in article volume and daily Bitcoin returns was 0.019, which means only 0.04% of daily price action was explained. For practical purposes, this is effectively zero.

The longer-term picture points in the same direction. Year by year, article volume and Bitcoin volatility moved on very different rhythms, with no stable relationship between heavier coverage and more explosive price behavior.

That doesn’t mean news and volatility never overlap. They obviously do. But over time, the relationship stays too loose and inconsistent to treat headline volume as a dependable signal on its own.

Price started showing up before the coverage

We also looked in the reverse direction: whether price moves tended to show up before headline volume did, and the most interesting pattern appeared around a two-day lag.

But the part that felt closest to actual market experience was looking at the 50 biggest news days and tracking Bitcoin’s price three days before and three days after each spike.

What stood out was the shape of the move. In the three days before a major coverage spike, Bitcoin’s price was already elevated, around 1% above the event-day baseline. Then after the spike, price drifted down by roughly 0.8% by day three.

That is not a “news moves markets” narrative. It’s a “markets move, then news catches up” story. And once you see that shape, you start noticing how many famous crypto moments feel like they rhyme with it.

Even the biggest headlines didn’t behave like clean signals

These are the kinds of moments we all remember because they felt like turning points for crypto. For example, the U.S. Securities and Exchange Commission approved the spot Bitcoin ETF on January 11, 2024. CoinDesk published 51 articles that day while Bitcoin dropped 7.67% the next day and was down 10% by day three.

Compare that with December 4, 2023, when speculation was running hot but nothing had been confirmed. CoinDesk published 81 articles, and Bitcoin rose 5% the next day. 

The same inconsistency showed up elsewhere: after the FTX collapse produced the busiest news day in the dataset, Bitcoin barely moved, while the January 2017 break back above $1,000 was followed by an 11% drop the next day and nearly 20% within three.

Across the ten biggest news events in the dataset, price reactions never settled into a usable pattern – some produced strong gains, others sharp losses, and many no clear follow-through at all.

That inconsistency matters because it’s what breaks the tradability story. If “news moves markets” were a stable indicator at the daily level, the largest news spikes would be where you’d expect the relationship to show up most clearly, certainly not where it dissolves into randomness.

We tried sentiment too

At that point, the obvious pushback is that volume is noisy, but sentiment might still hold the edge. Surely, bullish vs bearish headlines should matter, right?

So every headline was run through FinBERT, a financial-language sentiment model. It labeled each headline as positive, negative, or neutral. It also averaged sentiment across each day.

The dataset’s distribution was nearly perfectly balanced, with 58% neutral, 21% positive, and 21% negative. The part that matters for trading is the next step: did daily headline tone correlate with daily returns?

The reported correlation was 0.07, with sentiment explaining about 0.5% of price movement. Again, this is close to nothing for anyone trying to systemically time entries. Worse (or maybe more revealing), the relationship wasn’t stable. In rolling three-month windows, the correlation flipped between positive and negative with no consistent pattern.

There’s also something that feels obvious once you say it out loud: headline sentiment can end up ‘grading’ language that is already racing to price. A headline like “Bitcoin falls below $70,000” gets a negative score, but the fall is already in the same day’s price data.

So we’re back in the same place: the headline is describing the move, not front-running it.

The reframing that made everything make sense

None of what we have seen so far lands in the “ignore news” category. That’s not true, and it’s not useful.

The more hopeful shift is this: by the time a headline hits a major publication, the information has often already moved through faster channels. This includes order flow, on-chain data, social layers, insider networks, and other forms of positioning and interpretation that don’t wait for editorial cycles.

That’s the line that changes how we read the market. The media isn’t where the signal starts. It’s where the signal becomes legible. Headlines are pretty much the “last mile,” representing the moment when a move that has already begun gets named, packaged, debated, and turned into a story people can repeat.

What this changes

Reading faster doesn’t necessarily make you earlier. The market absorbs information before the newsroom has even agreed on the framing. Headlines are often better at telling us what just happened than what happens next. That’s not an insult to journalism. It’s a statement about timing.

And using media as a timing tool can put you behind the market, because the thing you’re reacting to may already be reflected in flows and positioning by the time you are ready to move.

Even the report puts it plainly: headlines are not a clean signal feed. On peak-coverage days, about 61% of headlines fell into broad industry noise – partnerships, fundraising, product launches, stablecoin developments, NFT and gaming updates – with no obvious link to Bitcoin’s next move. Even regulation, the strongest plausible category, still failed to produce a reliable signal at the daily level.

One of the stranger findings was that even Bitcoin halving did not emerge as a distinct cluster on extreme-news days, suggesting that some of Bitcoin’s most important forces do not operate through the daily headline cycle at all.

Where we have to be honest about the exceptions

News could matter at much shorter timeframes, specifically minutes rather than days. A breaking headline can still move the market in the moment, even if that effect gets muted once you zoom out to daily closing prices.

At the same time, longer and slower narrative shifts, the kind that build over weeks, may still influence price in ways this approach can’t fully capture.

There are limits to this too: one publication, even a highly trusted one, does not represent the whole information universe. Crypto’s fastest information often travels through social platforms and private channels that this dataset can’t track. Also, some patterns may only show up in specific conditions, not in the cleaner daily relationships these tests can pick up.

So we’re not left with a simple “news is useless” mantra. Rather, we’re left with something more actionable: most of the time, the headline is the market becoming explainable, not the market beginning to move.

Bitcoin’s on-chain picture remains centered on profitability dynamics, with the total supply in profit holding near a historically significant zone. As of Thursday, CryptoQuant data show about 60.6% of BTC supply in profit, placing the market in a band (roughly 50% to 60%) that has repeatedly framed cycles and potential accumulation phases. The metric briefly dipped to 50.8% on Feb. 5—the lowest since Jan. 2, 2023—leaving a sizable portion of holders at or near breakeven and at a potential loss.

Historical echoes are often cited by traders when profitability enters this range. In January 2023, BTC traded around $16,682 with profitability near 51%, just before a pronounced rally that CryptoQuant’s analysis notes as mirroring a pattern later seen in a multi-hundred percent upmove. A separate moment in March 2020 saw the total supply in profit slip below 50% as BTC hovered near $6,500, ahead of a bull run that pushed prices toward $69,000 in 2021. While past patterns can offer context, they do not guarantee future outcomes; profitability alone does not pinpoint price bottoms, but it does sketch zones where long-term accrual has been strong and selling pressure historically eased.

Key takeaways

• Bitcoin’s supply in profit stands around 60.6%, a level within the 50–60% zone historically linked to market-cycle resets and renewed accumulation.
• Long-term holder profitability remains meaningful: the long-term holder net unrealized profit/loss (LTH-NUPL) sits near 0.40, suggesting holders remain in profit even as overall profitability tightens.
• Institutional and corporate participation has grown, with entities holding roughly 15.8% of circulating BTC (about 3,319,677 BTC), potentially dampening short-term price sensitivity to swings.
• Short-term holder (STH) inflows to Binance have fallen to about 25,000 BTC on March 25, indicating less reactive selling from newer market participants.
• Valuation-based on-chain signals (MVRV, NUPL, Puell) are flashing zones associated with stress for retail demand but not definitive bottoms, highlighting a balance of risk and upside potential ahead.

Profitability baselines and market structure

The 50–60% profitability corridor has been a recurring feature across several cycles. When a large share of supply sits in profit, unrealized gains on the network compress, which can reduce the incentive for holders to sell into weakness. In this framework, the market’s current 60.6% profitability suggests a still-robust share of the supply that could weather minor downturns without triggering acute downside selling pressure. Yet the same metric also shows that a meaningful number of investors remain in the red or near break-even, underscoring the persistence of volatility and the potential for renewed demand when risk appetite shifts.

Crucially, the composition of who owns BTC is shifting. The rise of corporate entities and exchange-traded products (ETFs) as significant holders means a portion of the market is increasingly dominated by entities with longer time horizons and lower sensitivity to short-term price swings. In aggregate, these participants are estimated to control around 15.8% of the circulating supply, or roughly 3.32 million BTC. This dynamic tends to flatten peak-forcing selloffs that can accompany prolonged drawdowns, contributing to a market where profitability compression does not necessarily translate into a wave of distressed selling from veteran investors alike.

On-chain signals and market stress zones

Beyond aggregate profitability, on-chain flow metrics add nuance to the picture. Short-term holder activity has shown a meaningful contraction in selling pressure on BTC. CryptoQuant data indicate STH inflows to Binance dropped to near 25,000 BTC on March 25, a low not seen during the February sell-off, according to comments from market analysts. Such a drop points to a cooling in reactive selling from newer market participants and a potential for steadier price action if selling pressure remains subdued.

Meanwhile, traditional valuation models that analysts watch—market-value to realized-value (MVRV), NUPL, and Puell Multiple—continue to illuminate where stress is most likely to surface. Analysts have observed that when MVRV falls below 1, NUPL slips under -0.2, or Puell Multiple approaches 0.35, those periods have historically coincided with heightened retail stress or undervalued conditions. While these indicators do not guarantee a local bottom, they map out zones where downside risk has often been bounded by prior upside potential, offering traders a probabilistic framework for assessing risk-reward dynamics in the near term.

Taken together, the current on-chain configuration suggests a market moving away from the kind of acute, long-term holder distress that punctuated bear markets in 2015, 2018, and 2022. The divergence between a modestly higher supply-in-profit reading and steady LTH-NUPL points to a market that could see renewed accumulation without triggering uniform, forceful capitulation among long-term investors. In other words, the landscape is shifting toward an ownership mix that may support more measured corrections rather than sharp, cyclical lows.

Related: Bitcoin in ‘later stages’ of bear market: Watch these BTC price levels

What readers should watch next

For traders and investors, the key questions revolve around whether the current on-chain balance can sustain a move higher without retesting lows. The persistence of a sizable profit pool coupled with a growing share of BTC held by institutions could support a gradual re-accumulation narrative, even if price swings remain volatile. Markets will likely respond to macro developments, policy signals, and shifts in risk appetite as much as to on-chain metrics.

Next steps to monitor include: the trajectory of MVRV, NUPL, and Puell readings as BTC moves through key price zones; any shifts in the distribution of BTC held by corporates and ETFs; and observed changes in STH and overall exchange flows that could presage larger moves in supply held by retail participants. While on-chain data cannot predict exact bottoms, it continues to offer a granular view of where investors are positioned and how that positioning might shape the path of least resistance for Bitcoin in the months ahead.

Nasdaq’s push to bring tokenization into capital markets could lead to a dual-market structure in which traditional US exchanges operate alongside blockchain-based trading venues, according to TD Securities — a shift that could split trading activity and lead to price differences across platforms.

In a recent note, Reid Noch, vice president of US equity market structure at TD Securities, pointed to plans by Nasdaq and the New York Stock Exchange to introduce tokenization into alternative trading systems (ATS), a type of venue that matches buyers and sellers outside traditional exchanges.

While both exchanges are exploring tokenization, Noch said Nasdaq is pursuing three parallel initiatives: upgrading how trades are settled after execution, enabling companies to issue tokenized shares and supporting trading on offshore platforms such as Kraken.

Together, these efforts could result in two distinct systems — one within the regulated US market, and another operating through offshore, blockchain-based platforms.

However, TD Securities warns the expansion into offshore platforms could introduce a separate venue for trading the same underlying assets. These tokenized shares would be backed by real stocks but operate outside the US regulatory framework, with potential differences compared to traditional holdings.

For investors, that could mean the same stock trading in different places at different prices, making markets harder to follow and potentially shifting activity away from traditional exchanges.

Cointelegraph reached out to TD Securities for additional insights but did not receive a response in time for publication.

Related: Crypto Biz: Kraken plugs into the Fed

Tokenized trading moves into the mainstream

The market for tokenized assets has grown quickly in recent years, with equities emerging as the next major focus.

As Cointelegraph recently reported, Kraken’s xStocks platform, which offers tokenized versions of publicly traded shares that can be bought on blockchain-based venues, has surpassed $25 billion in cumulative trading volume, reflecting roughly 150% growth since November.

For traders, this points to a shift toward round-the-clock markets, where stocks can be traded outside regular hours. However, it could also bring new risks, including lower trading activity and price differences across platforms.

Coinbase has also expanded into tokenized stocks as part of its push to build an “everything exchange,” signaling growing competition between crypto platforms and traditional exchanges for equity trading.

NYSE, for its part, has also been exploring other tokenization initiatives through a partnership with Securitize, aimed at developing a platform for tokenized securities that could support extended or round-the-clock trading.

Related: VersaBank expands tokenized deposits with cross-border FX use case