Macro is soft growth plus sticky inflation risk; crypto is concentrating into fewer narrow leaders; and AI capability is running ahead of real deployment. The labor market is weakening, inflation is still muddy, and crypto is no longer paying for long-tail beta. Everything is becoming winner-take-most. A few protocols take the fees, a few AI products get adopted, and a few hard assets still hedge the policy mess. The long tail is where narratives go to die.
Key Takeaways
- Macro: Softer growth and weaker U.S. labor versus higher oil, geopolitics, and inflation risk feeding into rates and central-bank expectations.
- Crypto: Fade the long tail. If a category is already consolidating, the third- and fourth-tier token is a bad bet unless it is clearly taking share on execution, pricing, or distribution.
- AI: The gap between what LLMs can theoretically do and what they are actually doing says deployment is the bottleneck.
- TLDR: Crypto is no longer a broad-beta asset class where everything with a token catches a bid; it is shrinking into a narrow survivor set of cash-generative businesses.
Macro: Weak Jobs, Hard Oil
Soft growth and weak labor vs higher oil, geopolitics, and inflation risk. The labor market weakened faster than the Fed wanted: February nonfarm payrolls (NFP) fell by 92,000, unemployment rose to 4.4%, and average hourly earnings still grew 3.8% year over year. That is not a clean recession print and not a clean disinflation print either; it is softer hiring with wages still sticky. Meanwhile, February CPI matched expectations and printed 0.3% m/m and 2.4% y/y; core printed 0.2% m/m and 2.5% y/y. Shelter was still the biggest driver, but rent slowed to 0.1% while energy rose 0.6% and gasoline rose 0.8%.

The Fed is still at 3.50%-3.75% after its January 28 hold, and the next FOMC meeting is March 17-18. This week, Cleveland Fed President Beth Hammack and Boston Fed President Susan Collins both argued there is no urgency to move, while New York Fed President John Williams kept the door open to later cuts if inflation behaves. March is shaping up as a hold, and any easing case later in the year now depends on whether oil and tariff pressure stop muddying the inflation picture.

AI-led productivity can keep GDP decent while hiring lags, and geopolitics can create left-tail shocks on top. Jobs have decoupled from growth because of AI front-running, overhiring unwind, and policy uncertainty. Weak payrolls may not be a simple cyclical wobble, but part of a structurally uglier labor regime. The late-cycle debt message: too many financial claims on real wealth, bad real returns for cash and bonds, and better relative behavior from hard assets when money is being devalued.

Crypto: Winner-Take-Most
Crypto wins on liquidity, speed, and distribution. Stories and ideology don’t cut it anymore. You want to own protocols with hard cash flow and explicit value capture; “cheap” tokens without buybacks, burns, or fee rights are traps.

Most sectors are becoming monopolies or duopolies, with almost all revenue captured by the top 2 market participants. DeFi still sells infinite choices; revenue says the opposite. There is a clear Pareto distribution where the top 10/20 contenders capture 60-80% of fees. As a result, we are in an environment where you are better off betting on category leaders and avoiding subscale copycats. Case in point: Tether and Circle in stablecoins; Aave in lending, etc. This is why so many tokens stay optically cheap and still fail to rerate. The pie may be growing, but the slices are not being shared evenly.

Moats are clearer than crypto likes to admit. Stablecoins win on distribution, compliance, and reserve scale. Perps win on liquidity depth and execution. Lending wins on risk controls, trust, and institutional access. These are real business advantages. That is why the leader keeps taking more of the fee pool. The best businesses, however, are not always the best token exposures. So you have to be selective: own the protocols where fee concentration can plausibly flow through to tokenholder value, and be honest when the winning business sits outside the token.

Take agentic commerce and onchain payments as an example. Phase one is cards because Visa, Mastercard, and Stripe can let agents pay today without breaking merchant acceptance; phase two is stablecoins as the settlement upgrade; phase three is fully agent-native stablecoin flows for APIs and machine-to-machine workflows. Visa already supports USDC settlement for U.S. institutions on Solana with 7-day windows, Mastercard is enabling SoFiUSD settlement, Stripe is supporting agentic payment tokens from Visa and Mastercard, and Cloudflare/Coinbase’s x402 lets agents pay directly in stablecoins. There is no perfect liquid-token winner because the cleanest beneficiaries—Visa, Mastercard, Circle, Bridge, Stripe—sit outside crypto.
AI: At Work, Not At Scale
The latest Anthropic Economic Research piece, “Labor market impacts of AI: A new measure and early evidence,” is not the “AI killed white-collar jobs” bomb people want it to be. The paper’s main point is narrower and more useful: real AI deployment is showing up in work, but actual usage is still well below theoretical capability, and there is still no statistically meaningful unemployment shock for the most exposed workers.

The market will overreact to the scary headline and underreact to the boring truth. Actual coverage remains far below feasible coverage. Computer & Math is 94% theoretically exposed, but current Claude coverage is only 33% of tasks in that category. The most exposed occupations are computer programmers (74.5%), customer service reps (70.1%), data entry keyers (67.1%), medical record specialists (66.7%), market research/marketing specialists (64.8%), and financial/investment analysts (57.2%).

Yet AI is not shrinking work, but intensifying it: task expansion (people taking on work they previously would not have done), blurred boundaries (prompting spills into lunch, evenings, and “just one more prompt” moments), and constant multitasking across parallel AI threads. From a psychological standpoint, AI behaves like a slot machine plus a status threat, mixing intermittent wins with the fear of becoming obsolete, so the user feels both rewarded and pressured into doing more. Oftentimes, genuine productivity gains turn into burnout. If the company captures all of it, workers get squeezed dry; if workers keep all of it by dialing effort way down, firms lose to faster competitors. The problem is not just the tool but the social bargain around it: early adopters normalize insane output, founders and managers start expecting it, and workers get trapped between addiction, acceleration, and fear of falling behind.

Signals Radar
BTC — Strategy adding 17,994 BTC for about $1.3 billion at roughly $70,946 per coin keeps the same script running: corporate balance-sheet leverage is still a live source of mechanical BTC demand. With holdings now at 738,731 BTC, the company remains one of the most important marginal buyers in the market, which matters every time spot liquidity thins out.

ETH — Back above $2,000 in a macro-first, crypto-second week, which is a constructive setup if it manages to hold the line.

SOL — Even under $100, it presents a high-beta setup among majors. Solana is actively onboarding institutional traders through trading.solana.com, and CME now offers SOL options plus TAS on SOL and Micro SOL futures. Keep it as the preferred risk-on alt while it outperforms ETH; if that relative strength breaks, the trade is crowded.

HYPE — Real usage, real revenue, strong user loyalty, and a product that solves an actual market need. But the bull case is already carrying a lot of weight in price. Today’s valuation implies it can keep broker-like profitability and outsized margins even after competition, regulation, and KYC normalization compress the edge. The right lens is not “perps are big” but “how much of the market structure advantage survives when the easy arbitrage and regulatory slack disappear.”

Portfolio margin is now part of Alpha, unifying spot and perps (lets spot and perps PnL offset and supports carry trades). The next step lets users borrow up to 1M USDC or USDH against spot HYPE or BTC. Once caps are removed on mainnet, the doors will open to basis desks, hedgers, and multi-leg traders.

HIP-3’s Oil perps recorded more volume than ETH, hitting $991M in 24-hour volume. When traditional markets are closed, Hyperliquid is the premier venue for 24/7 price discovery on oil, metals, indices, and other essential assets. It surpassed $1B over the past two weeks in both open interest and weekend volume.

HIP-4 is the next product leg: fully collateralized outcome contracts, now rolling out on testnet, which opens prediction-style and fixed-range products like binary options without leverage or liquidation baggage; HIP-6 is the sleeper that could widen asset issuance for permissionless token launches via Continuous Clearing Auctions (CCA).
OKB — NYSE investing in OKX to collaborate on building digital assets infrastructure lifted prices by 30% last week, despite overall bearish conditions as BTC retraced under $70k. OKB moved from about $77.6 on March 5 to $95.9 on March 6 and briefly above $102 on March 8 before cooling back to the mid-$90s. The bull case: Wall Street validation plus distribution plus a permanently capped 21 million OKB supply. There are still headwinds: ICE called the stake financially immaterial, approvals still stand between narrative and product rollout, and OKX’s prior U.S. settlement means every U.S. expansion headline still carries regulatory headline risk.
AAVE — Governance drama aside, Horizon and v4 are the next incremental reasons to care, both contributing to the growth of GHO as a key revenue driver. AAVE remains one of the few majors with both real cash generation and a solid path to recycle earnings into the token. Rather than questioning its staying power, the trade is more related to whether the most recent governance noise offers a cheaper entry before the market starts paying up for quality again.

MORPHO — Chasing Aave’s tail in the quest for becoming systemic onchain credit infrastructure. The prize is not overcollateralized crypto loans; it is becoming the middleware layer for a much larger credit stack where collateral, underwriting, servicing, and distribution get standardized. This spans both variable and fixed-rate loans as well as DeFi and CeFi distribution. DeFi credit design has barely evolved since 2021. However, there is a clear market opportunity to bridge onchain capital efficiency with legal enforceability, KYC, and real-world origination. The DeFi mullet (DeFi on the backend; TradFi on the frontend) is the key to compounding durable volume, rather than just recycling crypto-native leverage.

ETHFI — Cash is already ~55% of annualized protocol revenue versus ~39% from Stake and ~6% from Liquid; spend is running at roughly $663M annualized / ~$1.8M per day, and ~80% of Cash users also use Liquid. Holder revenue is running at ~$19M annualized, with ~$16M in cumulative buybacks, a live $50M discretionary program while ETHFI trades below $3, and no planned change to the ~10% sETHFI target. After the migration from Scroll to OP the next priorities are Trade via Hyperliquid, tokenized stocks, and EUR/SEPA.

ENA — If Ethena’s deployed basis capital is near all-time lows while perp funding is slightly negative most of the time and deeply negative in stress, the market is telling you there is less demand for leveraged longs and more demand for outright shorts than in a normal cycle. That matters because Ethena monetizes a short-perp basis; when funding stays negative, one of USDe/sUSDe’s core yield engines gets squeezed. The protocol can cushion that with its reserve fund, and by shifting more backing into liquid stables, and Ethena’s February reserve update said USDe TVL was sitting idle around $6.5B with a roughly $62M reserve fund while perp hedging remained balanced. The read-through: this is a defensive perp environment, not a reflexive risk-on one. Absent catalysts, this is a short-term headwind for ENA narrative beta until funding turns sustainably positive again.

BABY — Babylon’s Ledger integration adds native signer support for BTCVaults, Clear Signing, and access to Ledger’s 8M-user hardware ecosystem. The most recent outperformance started with BABY up 30% while the broader crypto market declined. This came on the back of an inflation reduction from 8% to 5.5% plus BTC-BABY co-staking, giving BTC holders a reason to also hold BABY rather than just farm around it.

MNT — DeFi TVL broke above $750M, up 66% in seven days, and total market size on Aave topped $1B in three weeks. Good asset to own, more so when liquidity and activity keep compounding.

JUP — Introduced the Jupiter Card as a fully integrated mobile digital card inside Jupiter Mobile. Its vertical integration now spans the neobank angle, allowing users to spend their onchain balances anywhere Visa is accepted. To recap, the 2026 proposal takes net new emissions from roughly 1.2B JUP to effectively zero while Litterbox buybacks continue and Jup Lend is already above $1.8B in total supplied.

KMNO — Kamino is institutionalizing faster than the token is rerating, with official RWA market size above $1B, but the March 30 unlock keeps it buy-on-red rather than chase-on-green while the fee switch remains off.

2Z — Flipped the business model from validators paying for bandwidth to validators getting paid for data; the old 5% block-reward fee is gone as of Epoch 939, and Edge monetizes raw Solana shreds instead, i.e. traders subscribe for faster multicast delivery, 10% of revenue is burned, and the rest is shared across validators publishing shreds, network contributors, and protocol client teams; Edge already has 100+ publishers, partners like Jito, Triton, Helius, and Harmonic are live, and DoubleZero now carries ~46% of Solana stake.

STRK — Released the STRK20 framework, moving privacy from app-level tooling to the token standard itself. This is critical for privacy-focused stablecoins and other strategic assets. Any ERC20 can be made private, with shielded balances, recipients, and activity plus viewer keys for regulated disclosure. Starknet is already positioning strkBTC as the first live expression of that stack. From a trading standpoint, this offers bullish optionality for the privacy trade and a cleaner path for stablecoin and BTCFi experimentation on Starknet, but it stays story-first until actual issuers, liquidity, and shielded flow show up onchain.

KAITO — Turning its attention graph into a monetizable product: brands and creators match on audience/topic fit, campaigns get measured on influence, mindshare, and conversion, and the platform handles workflow, quote submission, progress tracking, wallet-linked payouts, and reward distribution. The moat is the data: Kaito already indexes crypto social, governance, and research with LLM-based analytics, so Studio is a proprietary matching and attribution layer, not just another creator marketplace. The actionable implication is that KAITO only rerates if Studio converts mindshare into recurring marketplace/SaaS revenue, but there is near-term supply pressure too.

INK — Kraken partners with Nasdaq to build an “equities transformation gateway” that would let tokenized equities move between regulated market rails and decentralized networks, with Kraken’s xStocks acting as the permissionless infrastructure layer.
KAST — Raising $80M in Series A funding led by QED Investors and Left Lane Capital. KAST already has 1M+ users and nearly $5B in annualized volume, expecting a $100M run-rate in 2026 while expanding across North America, Latin America, and the Middle East. Stablecoins are moving from backend to frontend. There is more to stablecoins beyond the issuer layer. Watch who owns the wallet, card, and user relationship (value accrues there as well).
VIRTUAL — Introducing ERC-8183 as an agentic commerce layer co-developed with the Ethereum Foundation. The official EIP defines a job-based escrow protocol with Open → Funded → Submitted → Terminal states and evaluator-driven settlement, while Virtuals is pushing it as an open standard for trustless agent commerce co-developed with the Ethereum Foundation’s dAI team.

ACX — Across is valuable crosschain infra, powering over ~$20B in volume, 15M+ transactions, and sub-minute fills. Yet ACX trades at ~$40M market cap and has historically struggled surpassing the $100M mark. A new temperature check exposed the “good project, bad token price” disconnect, suggesting the ability to let holders either swap ACX 1:1 into equity of a new U.S. C-corp (“AcrossCo”) or redeem into USDC at a one-month average price during a six-month window, with a no-fee SPV path for smaller holders. This is the first occurrence in crypto where the issuer of a permissionlessly accessible public token proposes to take development private.

Conclusion
The dominant trade is to own the survivors: perps, stablecoins, credit infrastructure, and a small number of protocols with real revenue, real users, and believable longevity. Most investors are crowding those names because there are few alternatives, not because valuation is obviously cheap. Selectivity matters even inside the winner cohort.


