And when even the institutional bulls are publicly calling it a “full‑blown crypto winter” regime, you can feel the temperature shift. Less dopamine, fewer true marginal buyers, more “career + capital elsewhere” energy.
Key Takeaways
- BTC fell to the lowest since late 2024, with liquidations >$1B.
- No premium for crypto optionality, no patience for leverage, no bid for middlemen. De-risk first, ask questions later.
- Everyone wants to ship new products. Nobody wants to buy the old ones.
- RWA is not a narrative anymore. TradFi distribution rules.
- Prediction markets are the “new social layer” (Solana gets the distribution).
- CT mood-shift into hostility, and the whole valuation meta is turning into “revenue now or you’re dead” linear thinking.
Macro: Capex Up, Confidence Down
After pressing pause and holding policy steady at 3.50%–3.75%, the Fed is not in a rush to cut while inflation remains “somewhat elevated” and the labor market “stabilizing”. The U.S. jobs report came out ahead of next week’s CPI. Next FOMC is Mar 17-18 (not immediate, but the path matters).
Equities are rotating away from expensive growth and into “defensive/value,” with the market explicitly taxing the AI capex arms race. Tech is still the marginal driver of risk sentiment (global tech sell-off tied to fears around the soaring cost of AI investment, with ~$830B wiped from tech stocks since Jan 28). The week can be resumed as “earnings are fine; multiples aren’t”.

US growth is not rolling over cleanly. Services are steady but not accelerating: ISM services held at 53.8, but new orders slowed and prices paid rose to 66.6, which keeps the “services inflation won’t die” narrative alive. Governor Lisa Cook is on the “wait-and-see / don’t cut until inflation behaves” side, explicitly highlighting upside inflation risks and tariff uncertainty. Markets are still pricing additional cuts after Powell’s term ends on May 15, with Trump nominating Kevin Warsh as successor.

For equities, earnings aren’t “bad.” They’re just not bailing out multiples. Big Tech is still spending hard on AI, but the market is increasingly asking “who gets paid?” (capex up, margins down, and a lot of AI optimism already priced).
- Alphabet/Google: $175–$185B capex for 2026 (near-double vs 2025) as it builds AI infrastructure; cloud growth cited at +48% in Q4.
- Meta: posted 2025 revenue $200.97B (+22% YoY) but guided 2026 capex $115–$135B and 2026 total expenses $162–$169B, explicitly framing infrastructure and talent spend as the growth engine.
- Apple: just printed a monster fiscal Q1: $143.8B revenue (+16%), EPS $2.84 (+19%), with iPhone + Services at all-time highs.
Meanwhile, crypto stocks got margin-called: MSTR -80% to ~$100, COIN -68% sub-$150, HOOD -54%, Galaxy -64%, Circle -84%, Bullish -80%, MARA -70%, Gemini -85%, SBET -95%, BMNR -90%, Twenty One Capital -90%, DJT -80%.

The next two catalysts are the only ones that matter: U.S. employment (Feb 11, shutdown-delayed chatter) and January CPI (Feb 13) set the rate/dollar impulse into the March 17–18 FOMC and the May Chair transition. Base case is choppy rotation (defensive/value up, long-duration down) while markets demand visible ROI on AI spend; bear case is hotter inflation/earnings rigidity pushing real yields + USD higher and forcing another leg of deleveraging; bull case is cooler inflation + easing financial conditions sparking short-cover. What flips the playbook is DXY rolling over—until then, preserve ammo.
Crypto: BTC Resets the Deck
The week started with BTC dropping under $80k for the first time since April 2025. It finished closing under ~$65k, pulling off the first real threat to MicroStrategy’s key level. This was a -5.65σ move using a 200-day lookback. For context, that volatility has only been seen 4 times in the history of BTC.

Overall, dangerous trading conditions ahead as wicks keep breaking structural support zones. As expected, alts multiply BTC’s pain. That’s the regime: BTC is the balance sheet; alts are leveraged residual. Until BTC stabilizes, alt bounces are trades, not investments.

The market is getting used to forced purges and deleveraging. Dips are no longer just driven by spot selling, and tracked liquidations continue showing OI overwhelm across the board. When BTC sneezes, alts react like they always do, panicking in classic bear microstructure. Bounces do and will happen, but leadership usually doesn’t return to alts until BTC stops trending down. There is barely any appetite for bidding alts without a solid BTC base. After flushes like this, you often get a sharp bounce, but unless spot demand shows up, it’s usually a dead-cat pop into resistance.

This is what apathy looks like: the loudest believers are de‑risking and the best communicators are reallocating their time. Vitalik is selling, ~2,961.5 ETH (~$6.6M) over the past three days (and still selling), including clips swapped into stables and routed out (donation motives aside, it’s founder sell‑flow). Kyle Samani is stepping back from Multicoin after a decade to go explore “other tech” (AI/robotics), i.e., a marquee crypto VC brain rotating attention away from the arena. Eigen’s DevRel Nader Dabit also posted he’s joining Cognition (AI agents), another high‑signal builder choosing the AI compounding curve over “wait for the next narrative.”
AI Digest: Build > Buy
- OpenAI released GPT‑5.3‑Codex (faster, more agentic, and positioned as a “do the whole job on a computer” model).
- Anthropic fired back with Claude Opus 4.6, pushing harder into long‑horizon work and knowledge‑work outputs (docs/spreadsheets/presentations) plus “agent teams” in Claude Code; it’s also rolling out a 1M‑token context (beta) and 128k output, with premium pricing beyond 200k tokens.
- SpaceX acquired xAI. Musk wants to roll its AI stack—capital, distribution, compute—consolidating narrative + compute + distribution under one private cap table ahead of IPO ambitions, with the “space-based data centers” angle as the headline hook.
- Evals: METR flags GPT‑5.2 “high” with a ~6.6 hour task time horizon. Not AGI, but enough to automate meaningful work.

- Clawdbot (rebranded through “Moltbot” into OpenClaw) is the clearest symptom of AI stealing mindshare from crypto. Not only in terms of mindshare, but also in terms of adoption.

- ElevenLabs: reported $500M Series D at ~$11B valuation (Sequoia lead per the roundup). Voice is still a top-funded wedge for consumer agents.
- Cerebras: reported $1B at ~$23B valuation. Compute financing is not slowing; it’s getting institutionalized.
- Gemini distribution / cost curve: Google claims ~750M MAU, 10B tokens processed/min, and ~78% unit-cost reduction since Gemini 3.0 launch.
- “Agentic engineering” > “vibe coding”: the mainstream dev loop is shifting from “write code faster” to “delegate tasks to agents.”
- VS Code agent sessions + Copilot Agents are pushing Claude/Codex style workflows into default dev stacks (distribution wins).
Signals Radar
BTC — Lost $80k again and struggling to find stability after a leverage unwind. The market already did the “fast flush” part; now it’s trying to rebuild structure. A clean read would set $80k as the line between “range and reset” vs “trend stays heavy,” but price action is struggling to confirm a recovery. Meanwhile, MSTR is still the “levered BTC beta” for a lot of players. Most recently, BTC dipped below the company’s average purchase price (~$76,052) during the selloff. If BTC is weak and MSTR is weaker, that’s risk-off confirming. Similarly, if BTC is flat and MSTR catches a bid, that could signal that risk appetite is returning (or short-cover). But generally speaking you still need BTC >80k to trust it.

ETH & SOL — Both acting as beta and unlikely to lead until BTC stops being fragile. Optimism might climb back as soon as they start making higher lows vs BTC. Until they start outperforming on up days (and not bleeding harder on down days), this is BTC-dominance chop. Solana’s tailwind is now prediction markets and distribution, while Ethereum’s headwind is Bitmine’s ETH treasury reflexivity (now holding 4.285M ETH, ~3.5% of circulating supply introducing a single point of failure that might have cascading effects to other DATs).

HYPE — Hyperliquid keeps on stacking wins. After HIP-3, which opened the huge market of equity perps, it now introduced HIP-4 to unlock options and prediction markets. Unlock fears were erased as the team will only be receiving ~140k HYPE instead of the full headline ~1.2M.

ASTER — Buyback machine back on; watch the wallet + $0.60. They have now activated the Strategic Reserve Buyback Fund, with buys executed from the reserve wallet 0x5E49…a397. This translates into more of the venue’s daily fees getting routed into spot market support via automatic buybacks, plus an additional reserve bucket for opportunistic buys when liquidity is thin. Line-in-the-sand is still that ~$0.60 zone; if it loses and can’t reclaim quickly, buybacks won’t save you in a risk-off market.

OKB — OKX founder, Star Xu, publicly blamed Binance for the Oct 10, 2025 flash crash (“10/10”). His claim: Binance’s USDe yield promo + collateral treatment created systemic leverage loops. Binance ran a limited‑time 12% APR USDe reward program, and explicitly counted USDe “used as collateral” in eligible balances. Xu argues users looped collateral (USDe > borrow USDT > buy more USDe > repeat), creating artificial “low‑risk” yields and latent liquidation fuel. Binance’s response: macro shock + market makers pulling liquidity + network congestion, no platform‑wide outage, and they claim ~75% of the day’s liquidations happened before the USDe/BNSOL/WBETH depegs they reference later in the timeline.

ENA — Getting hit from three angles—risk-off beta, supply unlocks, and narrative overhang—while the team remains silent building ahead of a big announcement. ENA needs to reclaim the mid‑$0.15s area and hold to stop the bleed; otherwise it’s still “sell rips.”

LDO — Price still sitting on the <$0.5 trapdoor despite renewed efforts with v3’s vaults. Last week’s $0.41 represented an all-time-low. The market indicates that it’s still risky to front-run a falling knife, regardless of future value accrual pathways and market sector dominance. LDO today is still mostly governance optionality, and not a token with hard, automatic cash-flow to holders. That said, there is an automated buyback framework in the works (still in forum stage prior to implementation, with voting planned this month).

ONDO — This week’s NYC Summit and latest roadmap—perps on tokenized equities, “day one” tokenized IPO access, and MetaMask distribution for tokenized U.S. stocks, ETFs, and commodities—inspired some hope, but the large FDV and overall market conditions were primed for “sell the news”. A break above ~$0.3 will require lots of volume for buyers to push further with conviction. ONDO’s anemic price action aside, the conference served as yet another stark reminder that crypto is here to stay, but that the DeFi and TradFi are bifurcating. TradFi giants aren’t interested in using public crypto infrastructure, but in building their own walled gardens to extract rent acting as intermediaries.
Lots of chatter around stablecoins, tokenization, RWAs, yet very little investable opportunities in sight that can yield outsized returns. Every asset issuer or intermediary now wants to own their own chain and stack, yet nobody wants to invest in public infrastructure such as existing L1s and L2s.

SYRUP — Perhaps one of the cleaner setups for alts, almost becoming the go-to DeFi bet. At a sweet spot under ~$500M in market cap and consistently collecting revenue (>$10M annualized), it stands out as an exception whose fundamentals continue printing new highs.

WLD — Dilution vs headline wars where supply unlocks are telling most of the story, especially when crypto’s picture is so bleak in terms of mindshare. OpenAI’s hints at a social media with human verification is not proving to be enough to attract meaningful buyers. Crypto is often a game played by a few whales, leaving mass retail on the sidelines. Without narrative, it’s hard to justify a “value buy.”

HASH — Very little volume and illiquidity are driving the chart of this token. It is not listed on CEXs despite FIGR_HELOC positioned in the top 15 by market cap (~$15B market cap yet ~$500k 24h volume). The actionable driver is still a community‑approved update to the Provenance Blockchain foundation structure (Jan 21), with operational execution tied to a Jan 14 community vote; the narrative is “more institutional structure + potential value accrual via fees.” Provenance’s RWA chain seeks to professionalize operations and turn fees into a security budget, meaning that it’s not a token-as-network-cashflow attempt.

The flip side is that Figure holds ~25% HASH and they are trying to reduce governance capture risk by abstaining from voting Foundation-controlled HASH. Today, HASH holders get paid from i) fee token demand source for Provenance transactions, ii) fees distributed to security providers as yield source (network-collected fees to validators, proposers and delegators), and iii) HASH Marketplace auctions + burn where 40% of network fees + 100% of settlement fees are routed into onchain auctions (e.g. users bid with HASH to buy assets; the winning HASH bids are burned (i.e., supply reduction)).

BERA — Price hugging all-time-low while unlocks keep coming in (Feb 6’s unlock was big: 63.75M BERA, ~$30M representing 12% of the total supply). More importantly, circulating supply jumped ~40% overnight. The bullish tell is: it holds ~$0.46 through unlock week with real volume and doesn’t wick new lows—then you can treat it as “supply cleared.” Otherwise, wait for the flush.

ZK — Another one approaching a big cliff, again reminding us how you can be just one gamble away from a 30% move that can move in either direction. Two concrete drivers lifted prices recently: Upbit’s listing, and the 2026 roadmap. They are embracing the institutional angle, with Citi + Deutsche Bank + Mastercard + 30 more institutions engaged around ZK’s enterprise “Prividiums”. Unless it’s trending hard with persistent spot bids, fade strength into Feb 17; if it loses ~$0.019 support, step away. If it holds >$0.03 into the unlock and doesn’t dump on the day, that’s your “supply absorbed” signal to re-rate it.

JUP — Disclosed a $35M strategic investment in JUP from ParaFi, settled entirely in JupUSD at spot with an extended lockup. The goal is to become the “everything app” and position itself as a critical onchain financial infrastructure (prediction APIs, discovery, distribution), and not just a swap router.

NMR — Announced a new $1M NMR buyback to be executed over the next two weeks (second disclosed $1M buyback, following the first strategic $1M buyback announced July 17, 2025). NMR is fixed-supply (max 11M) and already fully minted. Supply can still shrink over time via burns from unsuccessful staking. Their core issue, and reason behind these buybacks, is treasury depletion: they pay data scientists in NMR from the treasury and those reserves are running out (3M were left by the time of the first buyback announcement).

Thus, the buyback itself is explicitly framed as a treasury health initiative for sustaining tournaments, not “returning protocol revenue to holders.” It is a time-boxed spot bid (Numerai as a recurring buyer) to restock treasury and keep incentives running, not an explicit buyback-and-burn or a fee-switch style value-accrual program. Treat it as a repeatable marginal buyer in a market that’s often short on real spot bids, but don’t overprice it as permanent supply reduction unless they explicitly burn or structurally redirect fees.

L2s — The narrative never really picked up steam. If anything, it exacerbated the detrimental “too many chains, not many users.” With Vitalik now stepping up to reinforce a new story, Base may have just found the justification needed to become an L1 and launch its own token. The founder of Ethereum poured cold water on the “every L2 becomes an Ethereum branded shard” pitch. His point: 1) the crawl from L2s to “Stage 2” has been way slower than expected, and 2) L1 itself is scaling into 2026 (he flags a big gas-limit step-up). Net: the original L2 vision “no longer applies” and L2s need a new value prop beyond generic “we scale Ethereum.”

What he’s implicitly asking the ecosystem to do is separate winners from tourists: if you’re an L2 and you touch ETH / core Ethereum assets, you should at least be at “Stage 1” and maximize interoperability. Otherwise you need a real differentiator: privacy, application-specific efficiency, “extreme” scaling, ultra‑low latency, built-in oracles, or something genuinely novel.

Prediction Markets — The hottest market sector in crypto. Most recently Coinbase came out with their own. Robinhood is following a similar strategy. And so is Hyperliquid. Exchanges are all following an explicit strategy to become “the everything exchange” and prediction markets are a core part of the roadmap. Because of how hard it can be to source net new liquidity, most aren’t reinventing the wheel. For instance, the Coinbase app leverages Kalshi-sourced event contracts, tradable alongside crypto/equities and funded with USD or USDC (with Coinbase Custody safeguarding Kalshi’s USDC for event contracts). Jupiter also integrated Polymarket directly into the Jupiter app via a dedicated “Prediction” tab, removing the “bridge + leave your wallet comfort zone” friction and turning it into a native Solana flow. Kalshi is also pushing “tokenized prediction markets” on Solana via Jupiter + dFlow, explicitly leaning into onchain distribution.

DePIN — Escape Velocity just raised ~$62M (their second DePIN fund) and the thesis is “stop launching tokens before products.” They’re explicitly calling out the last cycle’s failure mode: token first, nothing shipped. Their framing is that the category is smaller than people think (~$10B market cap) but has real activity (~$72M onchain revenue in 2025) while tokens are still down 94–99% from ATHs—i.e., price got ahead of adoption, and now you hunt for the subset with actual revenue + deployment velocity.

TradFi — These companies aren’t “buying crypto.” They’re packaging crypto—stablecoin rails, tokenized stock rails, custody rails, compliance rails—then charging tolls on every hop. They don’t want exposure to your bags. They want flow. The product is the middleman, the margin is the spread, and the customer is anyone who needs access. That’s why the market keeps punishing most crypto tokens: the marginal buyer isn’t a bidder, it’s a builder. Capital goes to launching the next platform, the next chain, the next wrapper—because whoever owns distribution and compliance owns the rent. In a market where the default move is “build my own stack,” incumbents get commoditized, and valuations compress until there’s proof of real value capture.
Conclusion
Crypto isn’t dead. Most tokens may be. Reality has hit. Building won’t just cut it this time. Narratives are dead, and the time to face brutal reality has come. At today’s valuations, it’s no longer about who shouts the loudest. It’s about who owns the distribution, and the UX, and the fee surface, and the users. Those are, for the most part, TradFi giants. And they are, indeed, unlikely to launch a “moonshot” token to lift retail spirits. If you want to keep playing, stop paying for commodity rails. The giants will build their own, and they will capture all the value, and keep it all for themselves. If you still want to bet, do it on the few exceptions that accrue and share value, letting you at least capture a fair share of their returns.


