CZ Says YZi Labs Crypto Winter Bets May Outperform

Have you ever watched a market sink so far that even seasoned operators start sounding almost cheerful? That is the odd mood around Changpeng Zhao’s latest comment on YZi Labs. He did not claim a victory lap. He claimed a timing advantage. In early September he said investments made over the last few months could become some of the firm’s best performers because capital went out during the depth of the crypto winter. It is a familiar idea in private markets. Buy when the room is quiet. Hold when the noise returns. The hard part is proving it before the next financing round, the next listing, or the next product that actually ships.

Why Crypto Winter Timing Still Matters To Long-Term Capital

Markets have a habit of punishing people who sound certain too early. Zhao’s remark is an opinion, not a performance report. No acquisition prices. No marked-to-market book. No realized multiple. Still, the logic is not exotic. When public token prices fall, private founders often accept tougher terms. That can mean lower entry valuations, larger ownership, or more protective rights. It can also mean you simply own a company that never finds product-market fit. Cheap can still be expensive if the business dies.

I have found that winter narratives travel faster than winter data. People hear “bought the dip” and imagine a clean spreadsheet. Private portfolios rarely work that way. Equity, tokens, warrants, and side letters can sit in the same fund. One robotics check does not behave like a custody investment. A prediction market does not reprice the way a lending protocol does. Mixing those stories under one slogan is convenient. It is also sloppy if you care about how money actually compounds.

The broader backdrop is simple enough. Digital asset prices spent the first half of 2026 under pressure. Bitcoin later traded near the high seventy-thousands, better than the earlier trough, still below prior records. That bounce tells you almost nothing about a private robotics book or an early payments protocol. Public charts are a mood ring. Private marks are a negotiation.

What YZi Labs Actually Is, And What It Is Not

YZi Labs presents itself as an investment vehicle spanning Web3, artificial intelligence, and biotechnology. It grew out of a rebrand of a well-known exchange-linked lab and now says it manages more than $10 billion. The firm operates separately from the exchange business and is associated with Zhao and co-founder Yi He. That independence matters in a regulatory climate that still treats exchange capital and investment capital as different animals.

Zhao has previously suggested Web3 still accounts for roughly 70% to 80% of the book. The rest is stretching toward infrastructure that sits next to models rather than competing with them. Robotics. Agent payments. Institutional custody. Fixed-rate on-chain credit. In May he framed the preference clearly: build the pipes that AI needs, do not try to outspend the model labs. That is a respectable thesis. It is also a thesis that takes years to test.

I strongly believe the YZi Labs investments over the last few months will be some of the best performing because those investments were done during the depth of the crypto winter.

– Changpeng Zhao

Notice what the quote does not say. It does not say the positions are already winning. It does not say every check will work. It says the entry window was attractive. In venture, that is half the story. The other half is selection, follow-on discipline, and whether the company can sell something real when cheap capital disappears.

The Publicly Named 2026 Bets Worth Watching

Several deals from 2026 are now in the open. Robotics company RoboForce. Digital asset custodian BitGo. Prediction platform Predict.fun. AI payments protocol AEON. Fixed-rate lending platform TermMax. Some round sizes are known. Most individual checks are not. That gap is the first thing a careful reader should keep in mind.

Look at that table for a second. Five names. Five different demand curves. You cannot score them with one token chart and call it a day. Perhaps the most interesting aspect is how deliberately the book is leaving “pure token beta” behind without abandoning crypto rails.

RoboForce And The Largest Disclosed Check Of The Year

In March, YZi Labs led a $52 million financing for RoboForce. The company said the oversubscribed round lifted total funding to $67 million. That $52 million is the whole round, not the firm’s exact contribution. No valuation. No percentage. No liquidation preference disclosed to the public. So the headline is large. The economic picture is still foggy.

RoboForce is building physical AI systems for industrial work: solar sites, data centers, manufacturing floors, mines, logistics yards. The pitch is blunt. Labor is scarce in harsh environments. Robots that can do repetitive, dangerous, or simply unloved tasks could matter. The company also cited letters of intent covering more than 11,000 robots. Letters of intent are not purchase orders. They are not revenue. They are a sales narrative with a number attached.

Still, the fit with Zhao’s infrastructure-for-AI comment is obvious. This is not another exchange token. It is hardware, deployment, maintenance, and the messy physics of the real world. If those letters convert, the winter entry could look clever. If they stall in pilot purgatory, the round size will not save anyone.

• Industrial demand is the real test, not the press release.
• Letters of intent can evaporate when budgets tighten.
• Hardware timelines slip more often than software decks admit.
• A lead investor can still be a minority owner with limited control.

In my experience, robotics stories live or die in the second year after the raise, when demos stop impressing visitors and unit economics start arguing back. That is the window to watch, not the week after a funding tweet.

BitGo, Public Markets, And A Cleaner Price Signal

The BitGo investment sits in a different category. YZi Labs put capital to work around the custodian’s New York listing. Amount undisclosed. Terms undisclosed. What is visible is the strategic logic: regulated custody becomes more valuable if institutions keep showing up. A listed vehicle also offers something most of the book does not. Daily price discovery.

That matters. Private marks can stay polite for a long time. Public shares do not. If custody volumes rise with clearer rules, the listing can validate the winter-timing story in a way a seed round never will. If volumes disappoint, the market will say so without waiting for a founder update. I like that honesty, even when it is uncomfortable.

Custody is not glamorous. It is insurance, controls, audits, and the unsexy work of not losing client assets. That is precisely why it can be durable. Winter capital that lands in boring infrastructure often ages better than winter capital that lands in narrative tokens. Not always. Often.

Predict.fun And The Liquidity Question Nobody Can Fake Forever

In April, YZi Labs announced a follow-on in Predict.fun alongside Susquehanna Crypto. The platform had come through the firm’s EASY Residency program. Again, no check size. No valuation. What the companies did share was activity: more than four million orders and more than $1.8 billion in cumulative volume. Later, the product added a self-service developer dashboard for API keys, usage monitoring, and trading limits.

Volume is not profit. Volume is not valuation. Volume is not even a guarantee that users will stay when a hotter market opens next door. But volume is not nothing either. A self-custodial prediction market on BNB Chain that can keep order flow through a weak tape is at least demonstrating a habit. Politics, sports, crypto prices, economic events. The catalog is broad on purpose.

The follow-on itself is a tell. Firms do not usually add to a name they secretly dislike. They also do not always add for the reasons they advertise. Sometimes a follow-on is conviction. Sometimes it is defense of an earlier mark. Without terms, you cannot know which one you are looking at. That is not cynicism. That is basic fund hygiene.

AEON, TermMax, And The Quiet Shift Into Agents And Rates

In May, YZi Labs led an $8 million pre-seed for AEON, with several other funds in the round. AEON wants payment rails that let software agents execute and settle. Think of a settlement layer between autonomous code and both on-chain and off-chain money movement. If agents start buying compute, data, and services without a human clicking “confirm,” somebody has to clear those payments safely. That is the bet.

Pre-seed language can sound bigger than the company. Eight million for a full round is meaningful at that stage and still tiny next to model-training budgets. The firm’s point seems to be exactly that. Do not fight the model labs on spend. Own a piece of the checkout flow they will eventually need. Cute thesis. Execution will be ugly, because payments always are. Compliance, chargebacks, identity, latency, and the small matter of who is liable when an agent spends money badly.

TermMax arrived later, in August, with an undisclosed strategic investment. The protocol had already passed through EASY Residency Season 3 and has raised more than $8 million across rounds. It runs fixed-rate lending markets on several Ethereum-compatible chains, with pools, strategy vaults, and products meant to give both sides more predictable rates than variable DeFi money markets.

Fixed rate is an old banking idea wearing new clothes. Borrowers like certainty. Lenders like knowing the coupon. Crypto cycles punish both groups when rates whip around. If TermMax can keep liquidity deep when volatility returns, the winter check could look smart. If utilization stays thin, a strategic investment becomes a press mention and not much else.

Why A Winter Entry Is Not The Same Thing As A Win

Zhao’s argument rests on price. Lower marks can produce higher future multiples if the company survives and grows. That sentence should hang on the wall of every venture office. The second sentence should hang next to it: survival is not guaranteed, and growth is not evenly distributed.

Private performance cannot be read from Bitcoin alone. A fund may hold equity in one name, tokens in another, and a hybrid package in a third. Most of those terms stay private. Until there is a new round, a secondary sale, a distribution, or a public filing, the winter-outperform story is a hypothesis with good marketing.

1. Confirm the instrument: equity, token, warrant, or a mix.
2. Ask what valuation was used at entry, not just the round headline.
3. Separate company metrics from investor marks.
4. Wait for a second price, not a first press release.
5. Treat incubation alumni and lead investments as different risk stacks.

That checklist sounds dry. It is supposed to. Dry is how you avoid confusing a good quote with a good return.

Incubation, Residency, And The Soft Power Of Deal Flow

YZi Labs does not only write checks. It runs EASY Residency for early founders. Predict.fun and TermMax both passed through that funnel. Incubation is a sourcing machine. It is also a bias machine. You meet the team early, you like the team, you fund the team. Sometimes that proximity creates better diligence. Sometimes it creates sunk-cost affection.

I’ve watched similar programs in other cycles. The best ones act like filters. The weaker ones act like content studios. The difference shows up when the market is ugly and the firm still writes the follow-on. Anyone can host a demo day in a bull market. Writing a check when Twitter is quiet is a different temperament.

There is a second angle. Residency alumni can become a map of the firm’s taste. Payments for agents. Predictable on-chain rates. Markets for event contracts. That cluster says the team is hunting for infrastructure that can sit next to both crypto users and institutional workflows. Taste is not performance. Taste is a leading indicator of where the next bruises will appear.

How To Read A $10 Billion Book Without Seeing The Book

More than $10 billion under management sounds decisive. It is also a number that hides composition. How much is reserved? How much is already deployed? How much sits in liquid tokens versus illiquid equity? How much is concentrated in a handful of names? None of that is in the public remarks. So treat the figure as scale, not as proof of skill.

Scale changes behavior. A large book can lead late rounds and still miss early ownership. A large book can also warehouse positions through a winter that would force smaller funds to sell. That patience is a genuine edge if the underlying companies are sound. It is a genuine trap if the book is full of stranded experiments.

Winter-bet scorecard I actually use:
  1. Entry quality  - unknown until terms leak
  2. Business quality - only partly visible
  3. Path to liquidity - uneven across names
  4. Time needed     - longer than a news cycle

If that scorecard feels incomplete, good. Incomplete is honest. Complete-looking models built on missing prices are just fan fiction with columns.

The Five Businesses Do Not Share One Benchmark

RoboForce needs factories and field deployments. BitGo needs regulated flows and institutional trust. Predict.fun needs repeated liquidity. AEON needs agents that actually pay. TermMax needs borrowers and lenders who prefer a known rate over a floating gamble. You could design five different dashboards and still miss the investor’s economics.

That is why the outperform claim is hard to audit. There is no single index for “YZi Labs winter cohort.” Anyone who pretends otherwise is selling simplicity. Markets do not owe us simplicity.

A fair test would include entry valuations, later round prices, any secondaries, and cash or token distributions. Until those show up, the useful work is watching operating proof. Robots delivered, not intended. Assets under custody, not speeches about institutions. Repeat volume after incentives fade. Agents settling real invoices. Loan books that do not seize up when rates move.

What Bitcoin Near $80,000 Does And Does Not Prove

By early September, Bitcoin was changing hands near $79,945 in spot quotes circulating with the story. Recovered from lower 2026 prints. Still under the old high. Useful context. Weak evidence for a private book. A robotics company does not reprice because a coin bounced. A pre-seed payments protocol does not suddenly become a fund-returner because risk appetite improved for a week.

Public beta can help later financing. Founders raise more easily when the tape is green. That can mark up a winter position. It can also lure a firm into adding capital at a worse price than the original check. Follow-on discipline is where winter genius quietly dies.

So yes, the tape matters. It matters as weather. It does not matter as a substitute for company accounting.

A Practical Way For Readers To Track The Story From Here

You do not need a seat on the investment committee to stay oriented. Watch the next financing for each name. Watch whether customers, not just partners, start showing up in hard numbers. Watch whether BitGo’s public market treats custody as a growth business or a utility. Watch whether Predict.fun volume stays above the $1.8 billion cumulative marker in a way that implies ongoing flow rather than a one-time spike.

• New round prices are the first real mark after winter.
• Customer adoption beats logo slides.
• Public filings beat private adjectives.
• Liquidity events beat narrative events.

Will the firm publish a portfolio review? It has not said when. Until then, the measurable path runs through individual companies. That is slower than a quote. It is also how returns are actually made.

The Human Habit Behind Every Winter Speech

There is a psychological layer here that rarely makes the market notes. After a drawdown, people want the pain to have meant something. Buying in the slump becomes a moral story: we were brave, therefore we will be paid. Sometimes that is true. Sometimes bravery was just average capital with a calendar attached.

I do not doubt that some winter vintages become legendary. I doubt that every name in a winter vintage deserves the legend in advance. The honest stance is narrower. The last few months offered better asking prices in parts of crypto venture. YZi Labs used that window across a wider surface than tokens alone. Whether those specific surfaces compound is still an open file.

Lower entry prices can improve future returns when selected companies survive the downturn and grow. They do not guarantee that every investment made near a market low will succeed.

That is the whole essay in two sentences. The rest is detail, caution, and a reminder that charisma is not a mark-to-market policy.

Where The Strategy Could Age Well, And Where It Could Crack

The bull case is coherent. Institutions still need custody. Industry still needs labor substitutes. Agents will need to pay. Traders will keep betting on outcomes. Borrowers will keep wanting a rate they can plan around. If even two of those five lanes work at scale, a $10 billion platform can look patient rather than scattered.

The bear case is also coherent. Robotics stays stuck in pilot programs. Custody margins compress. Prediction flow migrates. Agent payments drown in compliance. Fixed-rate pools thin out when a shinier yield appears. Diversification across themes is not the same as diversification across outcomes. Five stories can fail for five local reasons in the same year.

Which case is more likely? I would not pretend to know from the outside. I would watch capital efficiency more than vision statements. Vision is cheap in this industry. Operating leverage is not.

A Closing Read On Confidence Without Receipts

Zhao sounded confident because winter buyers like to sound confident. That is part of the job. Confidence recruits founders. Confidence keeps limited partners from blinking. Confidence is not a crime. It becomes a problem only when readers treat it as an audited result.

The useful takeaway is smaller and sharper. A large crypto-native platform used a weak tape to lean into infrastructure around AI, custody, event markets, and credit. The disclosed rounds are real. The outperform claim is forward-looking. If you want a story about courage in a slump, you already have one. If you want a story about returns, you will have to wait for prices that have not been published yet.

That wait is not a flaw in the coverage. It is the point. Private markets reward people who can live with unfinished sentences. The next chapter will not arrive as a slogan. It will arrive as a round, a filing, a delivery number, or a quiet write-down. Until then, the winter bets are a portfolio of possibilities with better entry optics than the last boom. Optics can become outcomes. They are not outcomes yet.