Blockchain sleuth ZachXBT has accused employees at crypto trading platform Axiom of abusing internal tools to spy on users and trade using private wallet data, according to a detailed investigation posted on X today.

The alleged activity dates back to early 2025 and involves a senior business development employee based in New York.

According to ZachXBT, the employee, Broox Bauer, is heard on recordings claiming he could track “any Axiom user via ref code, wallet, or UID” and could “find out anything to do with that person.” In one clip cited by ZachXBT, Bauer also describes gradually increasing the number of wallets he monitored “so it does not look that suspicious.”

Axiom was founded by Henry Zhang — also known as Mist — and Preston Ellis — also known as Cal — in 2024 and later went through Y Combinator before quickly becoming one of the most profitable crypto platforms. The web-based trading terminal pulled in tens of millions of dollars in fees just months after launching in late January 2025, The Defiant reported earlier.

‘No Monitoring or Access Controls’

Screenshots shared in the X thread also show internal dashboards listing users’ private wallets, linked accounts, and transaction history, data that sources contacted by ZachXBT said appeared accurate.

The group also allegedly maintained shared spreadsheets mapping wallets tied to well-known traders and memecoin promoters. In another recording, Bauer lays out a plan to help a colleague make $200,000 by abusing this access, saying he would send over “the full list of wallets.”

“Regardless of whether Cal or Mist were aware, there was little to no monitoring or access controls in place to mitigate this abuse from happening in the first place,” ZachXBT wrote.

Because Bauer is based in New York, he added that the case could fall under the Southern District of New York, even if no criminal charges are ultimately filed.

It’s unclear, however, how much profit, if any, was made using the alleged insider information.

‘Shocked and Disappointed’

As the details came out, Axiom said in an X post, “We are shocked and disappointed to hear that someone on our team abused internal customer support tools to look up user wallets.” The company added that it “removed access to these tools and will continue to investigate and hold the offending parties responsible.”

Days before ZachXBT publicly named the firm, an alleged Axiom affiliate using the alias devininsider was already pushing back on speculation around Axiom. “We are simply a terminal that allows people to trade open market memecoins, what could we be possibly insider trading lol,” they said.

Blockchain tracker Lookonchain noted in an X post that just hours before ZachXBT named Axiom as the company accused of insider trading, two new anonymous wallets bet $59,800 on Axiom on Polymarket, and within three hours, turned it into $109,000.

Blockchains and artificial intelligence are complementary technologies, according to crypto asset manager Grayscale, even as markets have recently treated them as part of the same trade.

Zach Pandl, Grayscale’s head of research, said that while disruptive technologies tend to produce clear winners and losers, the relationship between AI and blockchain is more symbiotic than competitive. Rapid AI adoption is expected to reward some industries, such as chipmakers, while pressuring others, including segments of professional services.

“Although crypto valuations have been tightly correlated with the drawdown in software stocks, we think blockchains and AI are complementary from a fundamental standpoint,” he said in the Wednesday blog post.

U.S. equity markets have lately focused on the downside. The S&P 500 software index has fallen roughly 20% year to date, and crypto valuations have moved closely with the selloff. But Pandl maintains that the parallel drawdown obscures a more constructive long-term dynamic between the two technologies.

Investor anxiety about artificial intelligence’s disruptive potential has sparked a broad sell-off in tech and software stocks, erasing significant market value as traders reassess long-held valuations.

U.S. software and services shares have plunged sharply, wiping out roughly $1 trillion in market capitalization, as fears mount that fast-advancing AI tools could upend traditional business models and revenue streams.

The S&P 500 software index has slumped as investors rotate out of high-flight tech names amid heightened volatility and skepticism over how quickly and profitably AI adoption will play out.

Pandl contends that blockchains are likely to become the financial rails for AI agents. Today’s chatbots operate largely outside the financial system. But if AI agents are equipped with digital wallets, he expects them to transact over blockchains rather than traditional bank infrastructure.

Blockchains offer transparency, near-instant settlement, 24/7 availability and global reach with an internet connection, he said. While opening a bank account requires a human intermediary, any user, including a bot, can create a blockchain address. Pandl said rising volumes of low-value stablecoin transactions would be an early signal that this thesis is playing out.

At the same time, he argued that blockchain technology could help mitigate some of AI’s risks. As large language models proliferate, concerns around data provenance, deepfakes and the concentration of control over resources and decision-making are likely to intensify. Public blockchains, Pandl said, can provide verifiable records and more decentralized infrastructure to counterbalance those trends.

The report acknowledged AI may also introduce new challenges for crypto networks. Advanced tools could make blockchain surveillance more effective, potentially eroding user privacy. AI agents may also uncover new vulnerabilities in smart contracts; OpenAI recently launched EVMbench, an initiative aimed at using AI to identify and patch such risks.

Read more: Crypto isn’t losing to AI, its just ‘capitalism doing its job,’ says Dragonfly

Crypto markets dipped slightly on Thursday, with the total market cap dropping by about 2% over the past day to around $2.39 trillion.

Bitcoin (BTC) is trading near $67,000, down 2% over the past day but up 1% for the week, slightly below Wednesday’s peak.

Ethereum (ETH) slipped to $1,992, posting a 3% daily loss. Among other Top 10 assets, Solana (SOL) dropped 3.5%, XRP plunged 5%, and BNB fell 1.5%.

‘Constructive Return of Liquidity’

Analysts at glassnode noted in an X post today that “profit-taking continues to absorb momentum at the $70K threshold,” implying that this is consistent with a thin liquidity regime where even modest realization events are sufficient to suppress recovery attempts.

“Historically, breaks below 1 have persisted for 6+ months before reclaiming it, a recovery that typically signals a constructive return of liquidity to the market,” they added.

Paul Howard, senior director at crypto trading firm Wincent, said in commentary for The Defiant that stronger-than-expected earnings overnight had lifted tech stocks and risk assets more broadly.

He noted that “the short squeeze on Circle was notable, alongside the significant short interest in MSTR and the earnings beat from NVDA,” adding that these moves contributed to Bitcoin’s rally over the past 24 hours.

Howard added that the market is still looking for a clear catalyst that could push cryptocurrencies significantly higher, rather than just supporting them as a hedge trade.

Big Movers and Liquidations

Among the Top 100 assets by market cap, Pippin (PIPPIN) led gains with an 18.4% jump, followed by Internet Computer (ICP), which is up 8.5%.

On the downside, Cosmos Hub (ATOM) fell 7.9%, and Morpho (MORPHO) declined 3.6%.

CoinGlass reports that more than 157,000 traders were liquidated over the past 24 hours for a total of $560 million.

Shorts dominated with around $420 million liquidated, compared with nearly $148 million in long positions.

ETFs and Macro Conditions

Spot Bitcoin ETFs saw inflows of $506 million on Wednesday, Feb. 25, the largest single-day inflow since Jan. 5, bringing total net assets to $87.6 billion. On that same day, spot Ethereum ETFs added $157 million, bringing cumulative net assets to $11.8 billion.

On the macro front, U.S. Treasury yields were mostly flat. The 10-year note slipped slightly to 4.042%, the 30-year bond yield edged down to 4.687%, and the 2-year note ticked higher to 3.473%.

Thursday’s Labor Department report showed initial unemployment claims for the week ended Feb. 21 at 212,000, slightly above the prior week’s revised 208,000 but below the 215,000 forecast, CNBC reported.

On the geopolitical side, Iran’s foreign ministry said today’s nuclear talks in Geneva produced “very constructive” proposals, but didn’t give any details, according to the Associated Press. The U.S. and Iran are negotiating indirectly, with Oman’s foreign minister and the UN’s nuclear watchdog also present.

The AI and data center boom partly driven by Bitcoin miners is increasingly being financed through high-yield bond issuance, underscoring how lenders are pricing both risk and opportunity in the sector.

According to TheEnergyMag’s latest newsletter, companies tied to AI data center development have raised about $33 billion in long-term senior notes over the past 12 months, excluding convertible debt — bonds that can later be converted into equity and typically carry different risk dynamics.

The interest rate spread is notable: While regulated utilities and traditional energy companies generally borrow at 4% to 5%, AI- and crypto-linked issuers pay closer to 7% to 9%.

The average coupon on newly issued US dollar high-yield debt has was close to 7.2% in late 2025, from 8% to 9% in 2023, according to Janus Henderson Investors, citing BofA Global Research, average coupon, as of Nov. 30.

Those at the higher end of the spectrum are largely current or former digital asset mining companies that have pivoted into AI infrastructure, suggesting capital remains comparatively expensive for the group. 

TheEnergyMag cited recent raises, including CoreWeave at 9.25% and 9% in May and July 2025, Applied Digital at 9.2% in November, TeraWulf at 7.75% and Cipher Mining at 7.125% and 6.125%.

“The message from lenders is clear,” TheEnergyMag wrote. “Regulated load and contracted generation still get treated as infrastructure. AI and bitcoin, even when attached to long-term offtake agreements, are still treated as growth credit.”

Related: Canaan buys 49% stake in three Texas mining sites for $40M

AI infrastructure boom intensifies 

Despite concerns about overspending and potential overcapacity, the AI data center build-out remains one of the most visible trends in the economy, and a major driver of demand on Wall Street.

The scale of that momentum was underscored on Wednesday when chipmaker Nvidia posted blockbuster fourth-quarter results, with profit rising 94% and revenue climbing 73% year-on- year. The chipmaker reported $43 billion in net income and $68.1 billion in revenue.

Meanwhile, Bitcoin mining companies are planning about 30 gigawatts of new power capacity aimed at AI workloads, nearly triple the capacity they currently operate. Much of it remains in development pipelines or early-stage planning, but the industry has made clear that AI infrastructure is a strategic priority.

Related: The real ‘supercycle’ isn’t crypto, it’s AI infrastructure: Analyst