CZ Says Liquidity Is Still There, It’s Just Flowing Toward AI Infrastructure
Bitcoin is still under pressure, but Changpeng Zhao says the problem is not a dry market. In his view, plenty of liquidity is still around. It’s just being pushed elsewhere, and right now AI infrastructure is soaking up a lot of the attention.
- CZ says crypto weakness is not a liquidity drought
- Capital is rotating into AI infrastructure
- Land, power, and shell buildings are becoming the new bottlenecks
- Data center opposition is now a real political risk
The contrast is pretty stark. Bitcoin is trading near $63, 037, still down about 45% over the past year and roughly half of its all-time high reached on October 6, while investors keep hunting for the next place where scarcity can be monetized. For some, that means crypto. For others, it means power contracts, land, and the boring metal-and-concrete plumbing behind artificial intelligence.
That’s the real divide here, not money versus no money, but where money wants to go.
Liquidity Isn’t Dead. Conviction Is Just Elsewhere.
CZ’s argument is straightforward. [CZ Says Liquidity Remains Despite Crypto Bear Market as](https://www.tokenpost.com/news/insights/22505) he says crypto’s bear market is not being driven by a shortage of capital, and that significant liquidity remains available. That distinction matters.
A true liquidity crunch means there simply isn’t enough money around to take risk. A capital rotation means the money exists, but investors have found a shinier trade. Right now, that trade is AI infrastructure.
That’s not exactly what Bitcoin bulls want to hear, but it’s more useful than pretending every weak chart is a sign of apocalypse. Markets do not owe any asset permanent attention. They move on. Sometimes they move on because the story got worse. Sometimes they move on because another story looked easier to monetize.
In this case, the new obsession is the physical layer of AI.
Chamath’s Bet: Own the Bottlenecks, Not Just the Hype
Chamath Palihapitiya is leaning hard into what he calls LPS, Land, Power, Shell. That means acquiring land, securing power connections, and owning buildings that can later be converted into AI data centers.
“Shell” here is plain old real estate speak: an existing structure that can be repurposed into compute space. Not glamorous. Not sexy. Very profitable if the demand is there and the power keeps flowing.
It’s a classic pick-and-shovel strategy. Instead of betting on which AI chip startup becomes the next darling, the idea is to own the infrastructure everyone needs regardless of which model, chip, or cloud stack wins the race.
That logic is hard to dismiss. AI workloads are hungry for electricity, cooling, floor space, and grid access. A brilliant chip design is useless if you can’t energize the rack.
Palihapitiya’s interest in this space also fits the wider move from pure chip hype toward the physical reality underneath it. The source notes that he helped launch Groq in 2016, and that Nvidia licensed Groq’s technology in late 2025. Even so, the broader takeaway is not about one startup. It’s about where durable value sits when compute demand keeps rising and the easy narratives get crowded out.
[Binance CEO Changpeng Zhao's Leadership Challenges](https://www.bloomberg.com/news/features/2021-04-07/binance-ceo-changpeng-zhao-says-i-just-want-to-keep-crypto) can be a reminder that even the most influential crypto figures have had to navigate far more than market cycles, from regulation to reputation management to the sheer messiness of running a global platform in a sector that loves chaos almost as much as it loves decentralization.
Chip startups can be amazing businesses, or expensive science projects with a pitch deck. Infrastructure, by contrast, is less glamorous but often easier to underwrite if you control scarce assets. That said, “easier” is doing a lot of work there. Nothing about this space is automatic.
The Catch: Data Centers Are Running Into Political Reality
This is where the AI infrastructure trade stops being a clean spreadsheet story and starts looking like a fight with local politics, utilities, and regulators.
According to Q1 2026: Data Center Watch Report, at least 75 U.S. data center projects were blocked or delayed during early 2026, representing roughly $130 billion in investments. The organization also said community opposition has spread across 49 states, and lawmakers introduced more than 300 state data center bills within six weeks.
That is not background noise. That is backlash.
One of the sharper examples came from Maine, where a proposal narrowly failed by a single House vote after coming close to becoming the first state to ban new data centers. Whether the final legislation is framed as a broad ban or a narrower restriction, the signal is the same: these projects are no longer getting a free pass.
Why are people pushing back? For the same reasons they push back on almost any giant industrial buildout: electricity demand, grid strain, water use, land use, noise, traffic, tax incentives, and the suspicion that ordinary residents get the headaches while the operator gets the upside.
Data Centers and Their Energy Consumption is not exactly cocktail-party reading, but it captures why the issue keeps surfacing: these facilities are power-hungry by design, and once you scale them up, the utility bill becomes a public-policy problem real fast.
That’s a serious problem for the bullish AI infrastructure thesis. Scarcity makes land and power valuable, yes. But scarcity also makes projects harder to permit, slower to build, and more exposed to local resistance. A great power contract means nothing if the project gets stuck in permitting hell.
The resistance is not just local grumbling, either. Opposition to AI data centers has become a broader political and environmental flashpoint, and that’s exactly the kind of friction that can slow the next wave of buildouts if lawmakers decide residents are getting fleeced while hyperscalers and financiers cash in.
Why the Trade Still Makes Sense
Even with the pushback, there is a real investment case here. Physical AI infrastructure is not just hype; it is the hard constraint underneath the hype.
Land near power, grid access, and ready-to-convert buildings can generate faster and more reliable cash flows than speculative chip startups that burn through capital while chasing technical breakthroughs. That doesn’t make the infrastructure trade risk-free. It just means the value proposition is easier to understand.
The basic idea is simple: if AI keeps expanding, the owners of the bottlenecks may capture more durable value than the people trying to invent the next expensive chip miracle. That’s a fair thesis. It’s also a thesis that depends heavily on continued scarcity.
And scarcity has a habit of attracting capital until it stops being scarce.
As 22V Research’s Jordi Visser expects, AI returns may normalize. In plain English: the easy money may not last forever. If financing floods into the same power corridors and data-center shells, margins can compress fast. If permitting gets easier, scarcity premiums can shrink. If new supply catches up, today’s fat returns can turn into ordinary ones. That’s not doom. That’s just capitalism ruining a good story by making it crowded.
There’s also the question of whether the AI trade itself is getting a bit too breathless. [Klarna Co-Founder Warns of AI Bubble: Echoes of Crypto](https://adbytes.media/blog/klarna-co-founder-warns-of-ai-bubble-echoes-of-crypto-crashes-haunt-bitcoins-future) is a useful reminder that when everyone starts acting like a permanent winners’ club, reality tends to show up with a baseball bat.
The market already knows this can happen. [Crypto stocks rally thanks to rotation from AI](https://www.cnbc.com/2026/07/27/crypto-stocks-rally-thanks-to-rotation-from-ai-infrastructure-bitcoin-miners-lag.html) shows how quickly capital can lurch from one hot sector to another, leaving latecomers holding the bag and calling it a strategy. That’s not investing. That’s musical chairs with spreadsheets.
Bitcoin Is Competing for Capital, Not Building Permits
Bitcoin is not trying to become a data center business. Its fight is different. It’s competing for investor attention, risk appetite, and speculative capital.
That matters because short-term price action is often driven less by fundamentals than by which narrative is sucking up the room. Right now, AI infrastructure has the better story for a lot of allocators: tangible assets, visible scarcity, and a clear link to one of the market’s biggest secular themes.
Bitcoin’s long-term case is not weakened by that. It remains the hardest monetary asset in this market, with a fixed supply schedule and no CEO, no power bill, and no zoning hearing. But in the short run, capital rarely behaves like a philosophy seminar. It chases what looks scarce, concrete, and likely to get paid.
That is CZ’s point in a nutshell. The money is still there. It just may prefer power contracts over spot exposure, data-center shells over digital scarcity, and AI infrastructure over crypto, at least for now.
And yes, this is where the old Bitcoin chestnut still lands: wealth is often built in fear, not hype. [CZ: Bitcoin Wealth Built in Fear, Not Hype, Says Binance](https://adbytes.media/blog/cz-bitcoin-wealth-built-in-fear-not-hype-says-binance-founder) is basically the anti-shill thesis in one sentence, accumulate when everyone is busy telling themselves the party has ended forever.
The irony, of course, is that the same people who spent years mocking Bitcoin as dead money now act surprised when capital rotates out of frothy AI trades and back into the one asset that doesn’t need a permit, a substation upgrade, or a local council member with a grudge.
Meanwhile, even the AI chip winners are not immune to gravity. Nvidia and Groq Forge Partnership to Advance AI Inference highlights the kind of strategic deal-making that shows just how fast the infrastructure race is consolidating around power, speed, and efficiency, but it also underlines that not every winner gets to stay independent forever.
For a broader market backdrop, Crypto stocks rally thanks to rotation from AI is the kind of headline that makes the capital rotation thesis impossible to ignore. When money moves, it doesn’t ask Bitcoin’s permission.
Key Questions and Takeaways
-
Is crypto short on liquidity right now?
CZ says no. His view is that significant liquidity remains available, but investors are choosing other places to deploy it. -
Why is AI infrastructure attracting capital?
Because it sits on top of the real bottlenecks behind AI: land, electricity, and data-center capacity. Those inputs are scarce, which gives them pricing power when demand is strong. -
What does LPS mean?
It stands for Land, Power, Shell. Palihapitiya uses it to describe the core physical assets needed to build and monetize AI data centers. -
What is the biggest risk to this trade?
Permitting delays, community opposition, and political backlash. Data Center Watch says at least 75 U.S. projects were blocked or delayed in early 2026, which shows the obstacle is real. -
Does this hurt Bitcoin?
Not structurally. It does suggest that speculative capital is favoring AI infrastructure right now, but Bitcoin’s long-term monetary case is separate from the current hot trade. -
Why does this matter to crypto holders?
Because capital is finite and narrative-driven. When money rotates into AI infrastructure, it can leave fewer marginal bids chasing Bitcoin and the broader crypto market.
The bigger lesson is simple: capital is not gone, it is selective. CZ says liquidity is still in the market. Palihapitiya is betting on the physical layer of AI. And Bitcoin, for the moment, is reminding everyone that the hardest asset in the room can still sit out the party while investors chase the next obvious bottleneck.
Of course, if AI infrastructure gets too crowded, too political, or too expensive, the market may rediscover that the cleanest store of value doesn’t need a server farm, a permit hearing, or an electricity empire to function. The irony would be delicious. Markets love irony almost as much as they love leverage.
For readers tracking the broader rotation, Bitcoin Lags Micron as AI Chips Outperform in Major Capital is a blunt example of how hard it can be for BTC to compete when Wall Street decides semiconductors and AI hardware are the new golden child.
And if you want the longer arc on why CZ keeps returning to these themes, the through line is pretty simple: [CZ: Bitcoin Wealth Built in Fear, Not Hype, Says Binance](https://adbytes.media/blog/cz-bitcoin-wealth-built-in-fear-not-hype-says-binance-founder) is not just a slogan, it’s a warning. The loudest narrative is rarely the smartest trade.
