Physical Superintelligence Raises Capital: The AI-Funded Physics Lab Is the New ICO — but Nobody Can Audit the Code

The press release landed like a well-aimed shard of glass: Physical Superintelligence (PSI) has secured a nine-figure round to build an "AI-powered physics research lab." No model card. No whitepaper. No architecture diagram. Just a promise that artificial intelligence would soon rewrite the laws of nature. I've seen this play before. In 2017, I watched whitepapers appear with the same polished vagueness, backed by the same eager capital, and I remember exactly how that ended. The market rewarded narrative velocity over verifiable substance then, and it's happening again — only this time, the bait is physics, not consensus. As a crypto analyst who has spent a decade parsing blockchain signals from noise, I can tell you: the pattern is identical. The funding announcement is the hook. The technical depth is the void. And the crowd is already shouting "revolution" without a single line of code to audit. This is not a crypto article. It's a warning about the intersection of AI hype, scientific ambition, and the same speculative machinery that turned Terra into a textbook example of algorithmic delusion. When I audited the LUNA rebasing mechanism in May 2022, I found a circular dependency masked as innovation. Here, I see the same circularity: PSI claims to use AI to accelerate physics, but offers no evidence of how that AI is trained, what data it consumes, or how it differs from the hundreds of AI-for-science projects already racing across DeepMind, Microsoft, and a dozen well-funded startups. The funding is real. The substance is missing. And in a bull market where every announcement is amplified by a chorus of token-sniffing retail investors, the absence of technical detail is not a bug — it's a feature. Let me start with what we actually know. PSI is described as an "AI-powered physics research lab." The funding will go toward building infrastructure that presumably combines machine learning with experimental physics. That's it. No names of principal investigators, no list of peer-reviewed papers, no public repository, no demo video, no clear differentiation from the existing AI4Science ecosystem. The company's positioning suggests they intend to use large language models, neural operators, or reinforcement learning to generate hypotheses, design experiments, and perhaps even control robotic equipment. But all of this is inference. The original announcement, which I parsed with forensic calm, contains not a single technical specification. There is no mention of compute budget, model architecture, training data provenance, or even the specific physics subfields they intend to target. This is not an oversight. It is a deliberate choice to sell vision rather than veracity. Now, let me place this in the broader context. We are in the middle of an AI funding boom that mirrors the ICO craze of 2017. Back then, projects raised millions with a one-page PDF and a promise of decentralized everything. The smart contracts were unaudited, the tokenomics were often circular, and the technical roadmap was a screenshot of a Gantt chart. I built my reputation by parsing Ethereum blocks in real time, identifying pre-announcement signals, and publishing technical breakdowns within hours of whitepaper drops. That speed gave me an edge, but it also taught me a crucial lesson: the market's appetite for novelty far exceeds its appetite for truth. We are seeing the same pattern now with AI startups. Every week, a new lab raises hundreds of millions to "solve" protein folding, fusion energy, or material science. The press releases are longer than the actual technical contributions. And the investors, desperate to get in on the next DeepMind, are throwing money at any founder who can utter the phrase "AI-powered" with a straight face. PSI's funding is part of this broader wave. But there's a twist that makes it particularly spicy for someone like me who has spent years studying decentralized systems. The lab's stated goal is to accelerate physics discovery. That's noble. But the way they're approaching it — by hoarding proprietary data, keeping algorithms secret, and raising capital from traditional VC funds — stands in stark opposition to the open-science ethos that actually drives progress in physics. The Large Hadron Collider shares its data. The Human Genome Project published its findings openly. Even DeepMind's AlphaFold, despite being commercially valuable, released its predictions to the public scientific community. PSI, by contrast, appears to be building a black box. And in a field where reproducibility is the gold standard, a closed AI physics lab is about as useful as a blockchain with no public block explorer. Here's where my contrarian lens kicks in. The crypto community has a phrase for this: "fake it till you make it." But in our world, we have a mechanism to expose fakes — the public ledger. When a DeFi protocol launches, I can audit its smart contract on-chain. I can verify liquidity. I can trace token flows. The code doesn't lie. Uniswap taught me that liquidity is truth. When I look at Aave's interest rate models, I can see they're arbitrary, but at least they're transparent. With PSI, there is no on-chain audit. There is no public testnet. There's just a press release and a promise. And in the absence of verifiable technical progress, the only thing I can audit is the narrative. And the narrative is dangerously familiar. It's the same story we heard from Terra: a complex mechanism that would stabilize everything, a team with strong credentials (though here, even that is unverified), and a community of believers who would rather dismiss skeptics than demand proof. Let me dig into the technical reality. Even if PSI has a world-class team — and I have no evidence they do — the challenges they face are astronomical. AI for physics has been a hot topic for years, but we're still nowhere near a general-purpose AI that can derive new physical laws from raw data. The current state of the art involves machine learning force fields for molecular dynamics, physics-informed neural networks for solving PDEs, and automated laboratories that use Bayesian optimization to guide experiments. These tools are impressive, but they are narrow. They require massive amounts of high-quality data, careful feature engineering, and domain-specific inductive biases. A generic "AI-powered physics lab" sounds impressive, but it's like saying you're building a "crypto-powered financial system" — it tells you nothing about the consensus mechanism, the security assumptions, or the governance model. The devil is in the details, and PSI has provided zero details. Now, let's talk about what's really happening here. This funding round is not about physics. It's about market positioning. In a bull market, where every sector from crypto to biotech is inflated, investors are looking for the next big narrative. AI is the narrative du jour. Physics is the ultimate prestige field — it's the domain of Einstein, Newton, and Feynman. By combining AI with physics, PSI is tapping into two of the most powerful narratives in modern science. That's a potent cocktail for fundraising. But does it lead to actual discoveries? History is littered with examples of well-funded labs that produced little more than press releases. IBM's Watson for Oncology was a spectacular failure. Google's DeepMind has had successes, but they've been heavily curated and often take years to materialize into practical tools. The gap between lab-validated results and real-world impact is enormous. And for a startup with a closed research model, the path to commercialization is even more fraught. Let me bring in my own experience. I've been through multiple crypto crashes, and I've learned to spot the difference between signal and noise. In 2017, I made a name for myself by filtering signal from the ICO noise. I would parse smart contracts, look for hidden mechanisms, and publish breakdowns that exposed both the potential and the pitfalls. That experience taught me to ask specific questions: Who are the actual builders? What code have they written? What data have they validated? What experiments have they run? For PSI, I can't answer any of these questions because the information simply isn't public. And that's a red flag. In the blockchain world, we have a saying: "Don't trust, verify." PSI is asking us to trust them without offering any way to verify. That's the opposite of how science works, and it's the opposite of how sound investment works. Now, I want to get to the core of my analysis. The funding round for PSI is not just about one company; it's a symptom of a larger problem in the AI and tech industry. We are seeing a dangerous decoupling between valuation and value. Companies are raising billions based on narratives that are not backed by technical milestones. This is exactly what happened with ICOs — projects raised hundreds of millions without any product, and 90% of them died. The remaining 10% either pivoted or delivered something of marginal utility. The same fate awaits many AI startups. The difference is that AI has more real-world applications than blockchain, so the survival rate might be higher. But the principle remains: if you can't audit the technology, you're investing in a story, not a product. Let me also address the elephant in the room: the intersection of AI and crypto. The blockchain community has been talking about decentralized AI for years. Projects like Bittensor, Fetch.ai, and SingularityNET have attempted to create decentralized marketplaces for AI models and data. These efforts are noble, but they face massive technical and economic challenges. The idea is that decentralization can ensure transparency, prevent monopolies, and align incentives. PSI, however, is going in the opposite direction — centralizing all research under one corporate roof. That's a legitimate business model, but it's not necessarily the best way to accelerate scientific discovery. In fact, it might be the worst way. Science thrives on open collaboration, peer review, and the free exchange of ideas. Proprietary labs, whether they're run by Big Pharma or AI startups, often become silos that hoard knowledge and slow down progress. But here's the contrarian angle that few people are talking about: PSI might actually be building something revolutionary, and the lack of transparency could be a deliberate strategy to maintain a competitive edge. In the crypto world, we've seen projects that were initially opaque but later revealed groundbreaking technology. Ethereum itself started with a whitepaper and a promise, and it delivered. So it's possible that PSI has a genuinely novel approach to AI-driven physics that they're keeping secret to protect their IP. The problem is that I have no evidence to support that optimism. The burden of proof is on the company, and they have provided none. In the absence of evidence, I have to default to skepticism. That's what my forensic calm has taught me. Now, let me connect this to the broader market context. We are in a bull market for crypto, and that means the appetite for risk is high. Investors are chasing yield, chasing narratives, and often ignoring fundamentals. This is the perfect environment for a company like PSI to raise money on vapor. The same thing happened in the last bull run with DeFi protocols. Projects with no audit, no TVL, and no real users were raising millions based on a catchy name and a promise of high yields. I remember the summer of 2020, when I wrote my "Impermanent Loss Trap" series. I dissected the math behind Uniswap v2, and I showed how many LPs were losing money without realizing it. That series generated debate, but it also earned me a reputation as a contrarian who could see through the hype. I'm not saying PSI is a scam. I'm saying that without technical transparency, it's indistinguishable from one. Let me also consider the team. The original analysis noted that no team information is available. In a field where expertise matters more than anything, that's a critical omission. When DeepMind was founded, it had Demis Hassabis, a neuroscientist and chess prodigy. When OpenAI started, it had Sam Altman and Elon Musk backing it, and later a team of top AI researchers. PSI has... no names. Who is leading this lab? Do they have a track record in physics, machine learning, or both? Without that information, I can't assess whether they have any chance of success. It's like investing in a DeFi protocol with no founder publicly known. The crypto community has learned the hard way that anonymous teams are a red flag. There are exceptions, like Zcash, but they're rare. And even Zcash eventually had to reveal their identities to gain institutional trust. PSI's opacity is a huge concern. Now, let's talk about the potential impact if PSI actually succeeds. Suppose they do build an AI that can accelerate physics discovery. That would be a massive breakthrough, on par with the discovery of calculus or the structure of DNA. The implications would be felt across every industry — from energy to medicine to materials science. But here's the catch: if PSI succeeds, the benefits will be proprietary. They'll patent their discoveries, sell their insights to the highest bidder, and the rest of the world will be left behind. That's not how science is supposed to work. In the blockchain community, we often talk about public goods — infrastructure that benefits everyone. PSI is building a private good. That's a legitimate choice, but it's worth noting that the biggest advancements in physics have come from open research. Maxwell's equations, Einstein's relativity, the Standard Model — all of these were published openly. A closed AI physics lab would be a radical departure from that tradition. But let me take a step back. The fact that I'm even analyzing this funding round in such depth reflects the current state of the market. We're starved for new narratives. Crypto is in a bull run, but the fundamentals are shaky. Bitcoin's ordinals injected new life into the network, but they also created a lot of noise. Ethereum is scaling, but the blob data might be saturated in two years, causing gas fees to double again. We're looking for the next big thing, and AI is it. So we latch onto any AI project, no matter how vague, and we project our hopes onto it. That's a human tendency, but it's also a dangerous one. The smart investor is the one who stays calm, does their due diligence, and refuses to be swept up in the mania. I've survived multiple crashes by following that principle, and I'll continue to do so. Let me now offer a concrete takeaway for readers. When you see a press release like PSI's, ask yourself: what can I verify? If the answer is "nothing," then treat it as entertainment, not investment advice. The same applies to any project, whether it's crypto or AI or biotech. Look for public code, open datasets, published papers, and verifiable track records. If those don't exist, the project is a story, not a product. And stories, no matter how compelling, don't compound. They just fade away. I've seen it happen with countless ICOs, and I'll see it happen with countless AI startups. The ones that last are the ones that open up their hood and let you inspect the engine. That's why Uniswap taught me liquidity is truth. The smart contract never lies. It might be flawed, but it's transparent. PSI is not transparent, and that's a problem. Now, let me explore a few more dimensions. The funding round for PSI was likely led by top-tier VCs who have a reputation to protect. That means they've done some due diligence. They must have seen something that convinced them to invest. But in the crypto world, we've seen top-tier VCs back projects that later turned out to be fraudulent. I remember when a16z backed a blockchain project that was later revealed to have major undisclosed risks. So VC backing is not a seal of quality. It's just a signal that the founders are good at pitching. The real question is: what are the technical milestones? And since we don't know, we can't judge. Let me also think about the regulatory angle. In the future, if PSI makes any significant discovery, there will be questions about who owns the IP. If the AI generates a patentable invention, does it belong to the company? To the AI? To the researchers who trained the models? These are uncharted waters. In crypto, we have DAOs and smart contracts to handle governance, but in AI, the legal framework is still being developed. PSI might be a pioneer in this space, but that also means they're navigating without a map. That's risky for investors and for society as a whole. I want to bring in another personal experience. In 2026, I wrote a speculative series on "The Sovereign AI Wallet." The idea was that AI agents would have their own on-chain identities and could transact autonomously. I proposed a new token standard for machine-to-machine value transfer. The concept was theoretically sound, and it sparked vibrant discussion in developer forums. But I never followed through with implementation. I moved on to the next shiny idea. That's my weakness — the ideation-execution gap. PSI might be suffering from a similar problem. They have a grand vision, but they might not have the execution capability to pull it off. The lack of technical details suggests they're still in the ideation phase. And that's fine, but it also means they're asking for a lot of money based on a dream. Let me now address the contrarian angle more directly. The mainstream narrative is that AI will solve all our problems, including physics. But what if that's wrong? What if AI is actually a dead end for fundamental physics? The truth is, AI is very good at pattern matching, but physics is about finding the underlying rules that generate the patterns. AI can approximate those rules, but it can't necessarily derive them in a meaningful way. For example, AlphaFold predicted protein structures, but it didn't provide new insights into how proteins fold. It just gave us the answer without the explanation. That's useful, but it's not a revolution in understanding. PSI might be aiming for the same thing — predicting results without explaining them. That's a valid engineering approach, but it's not physics. Physics is about understanding why. AI is about predicting what. They're related, but they're not the same. Another contrarian point: the lack of technical transparency could be a sign that PSI has nothing to hide, but rather that they're too early to share anything. Many AI labs start with a team and a hypothesis. They need funding to build the infrastructure and run experiments. The first year might be spent on data collection and model training, with no publishable results. That's normal. But it also means that an early investment in PSI is a leap of faith. You're betting on the team's ability to execute, not on any concrete technical achievement. That's a very different risk profile than investing in a project that already has a working product. In the crypto world, we call this "investing in a pre-token project." It's high risk, high reward, and it requires a lot of trust. Let me also consider the possibility that PSI is using "Physics" as a buzzword to attract attention. They could be working on something more mundane, like materials science or battery chemistry, which has a clearer path to commercialization. But they're calling it "physics" because that sounds more profound. That's a marketing tactic, not a technical one. In crypto, we've seen projects use terms like "Web3" and "metaverse" to describe things that are just databases with a frontend. PSI might be doing the same with "physics." Without details, I can't tell. Now, let's talk about what this means for the crypto community. If PSI succeeds, it could inspire a wave of tokenized AI research labs. Imagine a DAO that owns a physics lab, where token holders vote on research priorities and share in the profits. That's an intriguing concept, but it's also fraught with challenges. Scientific research is inherently long-term and uncertain. Tokens are typically short-term and speculative. The two don't mix well. We've seen this tension with projects like Molecule Protocol, which tries to tokenize drug research. It has had some success, but it's still early. PSI is going the traditional VC route, which might be more practical. But it also means that the benefits of their research will be concentrated in the hands of a few investors, not distributed among a community. Let me now return to the core of my analysis. The PSI funding round is a perfect example of the "first-mover technical sprint" that I know so well. The announcement came out, and within minutes, the crypto twitterati was buzzing. Some were calling it a "paradigm shift." Others were calling it a "scam." But very few were actually reading the fine print. That's because in a bull market, we want to believe. We want to think that the next big thing is just around the corner. We want to be early. But being early is not the same as being in. The best investors are the ones who can wait for the signal to appear. And with PSI, there is no signal yet. There's just noise. Let me also touch on the role of data. AI models are only as good as their training data. For physics, the data is often scarce, noisy, and expensive to generate. Unlike natural language, where there's an abundance of text on the internet, physics data requires experiments, simulations, or observations. PSI will need to invest heavily in data collection. That's a capital-intensive process. The funding they've raised is probably just the beginning. They'll need more rounds. And each round will require them to show progress. If they can't show progress, the funding will dry up. This is the same pressure that many crypto projects face. They have a runway, but they have to deliver. The clock is ticking. Now, I want to make a bold prediction. In two years, we'll likely see one of two outcomes. Either PSI will have published some impressive results, or they will have pivoted to a more commercial focus, like selling AI tools to pharmaceutical companies or materials manufacturers. The latter is more likely because it's easier to monetize. But if they pivot, they'll lose the "physics" angle, and their narrative will weaken. In the crypto world, we've seen many projects pivot when their original vision failed. Some of them succeeded, like Ethereum's move from a general smart contract platform to a DeFi hub. Others failed, like BitConnect. The difference is often execution. And execution is impossible to assess without seeing the team and their work. Let me also discuss the psychological aspect. The PSI announcement is designed to trigger FOMO. The phrase "physical superintelligence" suggests something that will surpass human intelligence in the physical realm. That's a powerful idea. It taps into our primal fear and excitement about AI. But it's also a hallucination. We're projecting our hopes onto a name. In 2017, I chased alpha through the same kind of hallucination. I believed that blockchain would change everything, and it did, but not in the way I expected. The same will happen with AI. It will change things, but not in the way that the press releases suggest. The real change will come from small, incremental improvements, not from a single magical breakthrough. Now, let me talk about what I would do if I were a researcher at PSI. I would demand that the company publish at least a technical blog post or a whitepaper. I would ask for access to the codebase. I would seek out peer review. But I'm not a researcher. I'm a news aggregator. My job is to filter signal from noise. And in this case, the signal is buried under a mountain of hype. So I'll do what I always do: I'll wait. I'll watch for any technical publications. I'll monitor the team's social media. I'll look for leaks. And when I find something concrete, I'll write a follow-up. Until then, I'm not going to be swayed by a press release. Let me also consider the potential for PSI to be a honeypot. In crypto, a honeypot is a smart contract that traps funds. PSI is not a smart contract, but it could be a honeypot for investors. They raise money, build a nice office, hire some PhDs, and then fail to deliver. The investors lose their money, but the founders walk away with a salary. This is a known risk in the startup world. The lack of transparency makes this risk higher. I'm not saying PSI is a honeypot, but I'm saying that the burden of proof is on them to show they're not. Let me also think about the broader ecosystem. If PSI succeeds, it could accelerate the development of AI-driven science, which could have huge implications for crypto. For example, better battery technology could make mining more efficient. Better materials could lead to more durable hardware. But these are indirect effects. The direct effect would be that we have a proof that AI can accelerate fundamental science, which would validate the AI narrative and attract even more capital to the space. That's good for the ecosystem, but it also means more hype and more risk of a bubble. Now, let me bring in some numbers. The AI for science market is expected to grow to billions of dollars in the next decade. But that's a projection, not a reality. The reality is that most AI for science projects are still in the lab. PSI is one of them. They have no product, no revenue, and no published results. They have a vision and a team (maybe). The funding round is a bet on that vision. In the crypto world, we've seen bets on visions that paid off, like Ethereum. But we've also seen bets that failed, like EOS. The difference is often in the community and the communication. Ethereum had a clear roadmap and regular updates. PSI has neither. Let me also talk about the importance of open-source. In the crypto world, open source is a fundamental value. The code is public, and anyone can audit it. This builds trust. PSI is closed source, which means they don't trust the community to see their work. Or they don't want the community to see it because it's not impressive yet. Either way, it's a negative signal. In the AI world, there's a growing movement for open-source AI, led by organizations like Hugging Face and EleutherAI. PSI could join that movement, but they're choosing not to. That's their right, but it makes it harder for me to evaluate them. Let me now address the title of this article: "Physical Superintelligence Raises Capital: The AI-Funded Physics Lab Is the New ICO — but Nobody Can Audit the Code." I'm not saying PSI is an ICO scam. I'm saying that the dynamics are the same. There's a rush of capital, a lack of technical detail, and a community that is all too ready to believe. The term "ICO" has a negative connotation, but the underlying pattern is just a speculative bubble. And bubbles always burst. The question is whether PSI will be one of the survivors or one of the casualties. I don't have a crystal ball, but I have a pattern recognition system. And the pattern doesn't look good. Let me also consider the role of media. As a news aggregator, I know how easy it is to amplify a press release. You just copy-paste, add a few adjectives, and publish. That's what many outlets did with the PSI announcement. They didn't ask questions. They didn't verify. They just reported the funding. That's not journalism; that's PR. My readers expect more. They want analysis. They want to know what this means for the market. And that's what I'm providing here. The funding is a story, but the real story is the absence of substance. Now, let me wrap up with some actionable advice. If you're an investor, don't put money into PSI or any similar project until they show technical proof. If you're a researcher, consider collaborating with open-source alternatives. If you're a developer, look at what open-source AI tools are available and build on them. And if you're just a reader, don't get caught up in the hype. Remember that every bubble has a story, and the story always sounds good until it doesn't. I've survived the 2017 hallucination, the Terra algorithmic trap, and the DeFi winter. I've learned to trust the code, not the narrative. And right now, PSI is just a narrative. Let me end with a forward-looking thought. In the next year, we'll see more AI labs raise money. Some will be like PSI. Others will be more transparent. The ones that survive will be the ones that embrace openness. Because in the long run, trust is the only currency that matters. And trust is built on verifiable actions, not press releases. The smart contract never lies. It's time for AI physics labs to follow suit. So, the next time you see a funding announcement, ask yourself: can I audit this? If you can, great. If you can't, then you're just a spectator in someone else's game. And in a bull market, that's the most dangerous position to be in. I'll be watching. And I'll be waiting for the code.