Gm, and happy Friday! 

If you’re supplying liquidity on Robinhood Chain, where and when you deposit your capital changes everything. Today, we break down why some pools are printing fees while others hand money to arbitrageurs and why the opening bell leaves little margin for passive LPs.

First, markets.

BTC is the laggard again at -0.3% on the week against 1.9% for the NASDAQ 100, 1.2% for the S&P 500, and 0.3% for gold; year to date it is -12.8% while the NASDAQ 100 has run 16.7%.

Flat is a better result than it sounds given what BTC absorbed. The CLARITY Act failed to proceed in the Senate on Tuesday, failing cloture 49-50 against the 60 votes it needed, with every Democrat and four Republicans opposed, which pushes market structure into next year at the earliest. The Fed followed on Wednesday with a 25bp hike to 3.75% to 4.00%, its first since 2023 and a unanimous 12-0 vote, and Warsh spent the press conference arguing conditions are not restrictive. BTC dipped near $76k on the decision and has since worked back toward $78k, holding the lower end of its range through both.

The 2026 leaders did not wait around. Privacy ran 27.3% on the week and Perps 8.8%, both extending while BTC went sideways, with DEXs topping the board at 30.2%. The only real red was the equity complex: the 2025 cohort fell 9.4%, crypto miners 4.9%, and crypto equities 4.7%. Those two leaders carry the year as well, Perps +226.3% and Privacy +103.2% against BTC at -12.8%.

 ZEC did the work in Privacy, +35.6% on the week to $1,464 and printing new highs into the Fed. The bid is a monetary one: ZEC is increasingly treated as an alternative form of non-sovereign money with properties BTC lacks, privacy native to the protocol, quantum resistance, and formal verification of the shielded pool since Ironwood. The tell is that XMR managed just 1.6% over the same week, so this is a rerating of ZEC specifically rather than a privacy sector trade. HYPE is the Perps engine, +8.8% to $85.46 at Thursday's close and above $90 since, tagging a fresh all-time high of $90.92 this morning that clears the $89.57 it set on September 6, with open interest back near $14.3B and fees still routing into buybacks. DEXs were the loudest print of all at +30.2% on the week and +129.0% on the month, almost entirely UNI at +32.5% to $7.88.

Two of the worst headlines crypto could get landed in 48 hours and BTC finished the week unchanged, which says the sellers who were going to leave on Washington already left. As long as that range floor holds, there is little reason for the bid to rotate out of the handful of assets with momentum and back into majors waiting on a bill that is not coming this year.

— AJC

Where LPs Win (and Lose) on Robinhood Chain

Robinhood Chain has been live long enough to ask whether LPs earn enough in fees to cover what arbitrageurs extract from them. For this measure, I looked at tokenized equity LPs and used loss-versus-rebalancing (LVR), an estimate of AMM adverse-selection costs, as the cost metric. 

We see below that in most pools carrying real equity flow, LPs are covering their costs multiple times over, but surprisingly, in the pools built around memecoins, not so much. Equity/ETH pools earned ~3.3x their (estimated) arbitrage cost. Equity/stablecoin pools earned ~2.9x. Equity/equity pools sit at 1.02x (essentially break-even), while equity/memecoin pools recovered 20 cents on the dollar. Moving away from averages, the dispersion data tells the same story. Most stablecoin and ETH pairs were profitable, but roughly half of equity pairs and less than 2% of meme pairs could say the same.

Timing also matters. Even though fees and arbitrage costs both run roughly 60% hotter during the 09:30–16:00 ET cash session, the ratio barely moves. So at first glance, LPs aren’t being harvested after hours.

However, the ratio collapses at the opening bell. The 9:30 ET bucket carries $95.08 of arbitrage cost per pool against a roughly $7 pre-open baseline (13x difference). Meanwhile, fee income in that bucket covers it 1.05x, down from the 3x to 6x running through the pre-open hours. It decays quickly with about $41 in the next bucket, $24 in the one after, back to low-teens within the hour.

So if you’re supplying tokenized equity liquidity on Robinhood Chain, the takeaway is twofold. 

Pair selection is doing most of the work. An equity token against ETH or a stablecoin is a viable position with a wide margin. The same token against a memecoin is a transfer from the LP to the arbitrageur. Equity/equity pair is interesting as it clears on average, but a 57% hit rate means the median pool is not obviously worth the capital. Notably, this analysis only measures the cost of being traded against on stale quotes, and says nothing about the directional exposure of holding the inventory. An equity/memecoin pair can lose on fees but if the memecoin rockets, then the LP is still in profit.

Timing is doing the remainder of the work. A passive LP eats the opening bucket and anyone who can widen or pull around 09:30 ET keeps the 3x and skips the one window where the edge is roughly zero. This is a core argument that this liquidity wants to be managed, not deposited.

— Marc Arjoon

US AI models make far more revenue than China's

• Chinese AI models only make about 10% of the revenue that OpenAI and Anthropic do.

• Valuations relative to revenue appear exorbitant for Moonshot and DeepSeek at present.

• Chinese AI labs' revenue forecasts have been raised, but remain well below that of US rivals.