RIC AI Weekly: Capital Markets Are Becoming Machine-Native
Updated: Sep 19
September 18, 2026
For years, artificial intelligence and digital assets were treated as separate technology stories. This week made that separation harder to defend.
The SEC opened a regulated path for real U.S. stocks to trade on public blockchains. S&P Global agreed to acquire OpenZeppelin, one of the core security companies behind onchain finance, only days after investing in digital-asset data provider Kaiko. Ripple connected XRP and RLUSD to a machine-payments standard built by Stripe and Tempo. A startup is seeking regulatory approval for an exchange where GPU capacity could trade like a commodity. Crusoe, a company that began by monetizing stranded energy through bitcoin mining, raised $3.9 billion at a $30.9 billion valuation to build AI infrastructure.
These are not disconnected headlines.
AI is learning to make decisions and take action at the same time that capital, securities, payments and compute are becoming programmable.
That convergence may be one of the most important business and capital-markets stories of the next decade.
The Signal
1. Wall Street is beginning to move onto blockchain rails
On September 17, the SEC introduced a five-year “innovation exemption” that creates a regulated path for qualifying venues to trade real tokenized U.S. stocks on public blockchains without first becoming a traditional national securities exchange.
The distinction matters: these are not synthetic tokens that merely track a stock price. Qualifying tokenized shares must preserve the underlying shareholder rights, including dividends and voting rights. The SEC is allowing controlled experimentation with smart contracts, automated liquidity pools and public blockchain infrastructure while keeping permissioning and securities-market safeguards in place.
This moves tokenization from a crypto experiment toward actual market structure.
The capital-markets implication is significant. If securities can ultimately be issued, owned, transferred and settled on programmable rails, the infrastructure between investor and asset begins to compress. Settlement, collateral management, transfer agency and market access all become candidates for software-driven redesign.
2. S&P Global is buying the risk layer of onchain finance
The most revealing M&A story of the week may have come from a company most people would not describe as a crypto company. On September 17, S&P Global announced an agreement to acquire OpenZeppelin, whose smart-contract libraries and security work underpin a large portion of tokenized finance. S&P says OpenZeppelin’s contracts have supported more than $37 trillion in value transferred.
Three days earlier, S&P Global led a strategic investment in Kaiko, the institutional digital-asset market-data and analytics company.
Put the two moves together and the strategy becomes visible: one of the most important institutions in global ratings, benchmarks and financial intelligence is building capabilities around onchain data, onchain security and onchain risk assessment.
That is a much stronger signal than another crypto startup raising money. Traditional financial infrastructure is beginning to acquire the tools it needs to underwrite and monitor programmable capital markets.
3. XRP is moving into the machine economy
Ripple gave the AI-capital convergence a particularly concrete form this week.
On September 17, Ripple expanded its XRP Ledger AI Starter Kit to support the Machine Payments Protocol developed by Stripe and Tempo. The integration allows AI agents to pay for services using XRP and ledger-issued assets such as RLUSD. The software also supports controls such as spending limits and approved destinations without giving the agent direct access to private keys.
That matters more than whether XRP moved up or down on a particular trading day.
The interesting question is whether digital assets become native settlement instruments for software that buys data, compute, APIs and services without a human manually initiating every transaction.
Ripple is attacking the same problem from the corporate side. Its Ripple Treasury platform now embeds governed AI into cash forecasting, liquidity, risk, reconciliation and reporting, while keeping human approval over financial actions.
In other words, Ripple is simultaneously working on AI that helps manage money and money that AI can use.
4. Compute itself is starting to look like a financial commodity
A startup called Liquid Compute raised $15 million this week to build what it intends to be a regulated exchange for AI computing capacity.
The concept is straightforward but potentially profound: GPU hours would become a priced, tradeable resource, with the possibility of futures and options that allow companies to hedge compute costs or monetize unused capacity. Liquid Compute has applied to the CFTC to operate an exchange and clearinghouse.
Electricity, oil, interest rates and currencies developed deep financial markets because businesses needed ways to price, finance and hedge critical inputs.
If compute becomes one of the dominant inputs of the AI economy, it is logical that financial markets begin forming around it.
That would turn GPUs from technology equipment into something closer to a financeable, hedgeable production input.
5. Crusoe may be the cleanest example of capital migrating from crypto to AI
Crusoe began by using stranded energy to mine cryptocurrency. It then redirected that energy and infrastructure expertise toward AI.
This week, Crusoe raised $3.9 billion in a Series F round at a $30.9 billion valuation. The company is building both enormous AI campuses and smaller modular “AI factories” that can be deployed where power is available.
The pivot itself is as interesting as the financing.
Crypto mining helped prove that computational workloads could migrate toward cheap or stranded energy. AI has taken that same basic economic insight and multiplied the capital requirement dramatically.
The result is a new infrastructure category sitting at the intersection of energy, real estate, chips, project finance and cloud computing.
6. AI infrastructure is beginning to carry capital-markets risk, not just technology risk
The scale of the buildout is now large enough that the financing architecture deserves as much attention as the technology.
SoftBank-backed SB Energy is preparing for a potential public listing while planning more than $170 billion of capital expenditure. Its growth case is heavily tied to AI customers and infrastructure partners, while much of its contracted future revenue sits years in the future.
At the same time, analysis of the wider AI ecosystem is increasingly highlighting the circular relationships among model companies, chipmakers, cloud providers, infrastructure developers and investors.
That does not mean the AI buildout is a bubble. It means underwriting has to mature.
When the same institutions invest in a company, finance its suppliers, guarantee equipment, lease its capacity and depend on its future growth, counterparty concentration and circular capital become real credit questions.
What We Learned
The RIC AI lens on this convergence is becoming clearer. Capital has always been an information problem before it becomes a money problem.
A transaction moves through a chain: information is collected, risk is interpreted, a capital source is matched, diligence is exchanged, terms are evaluated, authority is obtained, funds move and the result is monitored.
AI compresses the time between those steps.
Tokenization and digital settlement compress the movement of the asset itself.
Machine payments compress the movement of money.
And programmable infrastructure begins connecting all three.
That suggests the next generation of financial systems will not simply use AI as an assistant sitting beside the transaction. AI will increasingly live inside the transaction architecture.
An agent may identify a financing need, retrieve the relevant records, evaluate alternatives, purchase the data required for analysis, request additional diligence, monitor covenant conditions and prepare a decision package. The money, collateral or security it is analyzing may itself exist on programmable rails.
The important constraint is authority.
The most credible systems emerging this week — from Ripple’s treasury tools to regulated tokenized securities — are not eliminating controls. They are embedding controls directly into the architecture.
That is likely to be the winning model: more automation, but also more explicit permissioning, auditability and human authority at the points where capital actually moves.
Why It Matters to Business & Capital
The boundary between fintech, AI and crypto is dissolving
A ratings company buying smart-contract security infrastructure, Stripe-linked payment standards supporting XRP, and the SEC authorizing tokenized stock venues all point in the same direction. These categories are becoming one financial-technology stack.
Capital formation could become faster and more programmable
Tokenized securities do not automatically make a bad investment good. What they can change is the machinery around ownership, settlement, distribution and collateral.
That can reduce friction and eventually create financial products that operate continuously rather than around legacy processing windows.
AI agents will need financial identities and spending authority
If agents are going to buy data, reserve compute, pay vendors or manage treasury operations, they need more than a model API. They need wallets or payment credentials, limits, approved counterparties, transaction logs and revocation controls.
Financial identity for machines could become a major infrastructure category.
Compute is becoming something capital markets can finance and hedge
If GPU capacity develops transparent spot and derivatives markets, compute costs become easier to price and potentially easier to finance.
That could change underwriting for data centers, AI companies and any business whose margins depend materially on inference or training costs.
Crypto’s most important contribution may be infrastructure, not speculation
The most consequential crypto story may ultimately be less about token prices and more about what blockchain technology contributed: programmable ownership, programmable settlement, digital collateral and always-on markets.
XRP’s relevance in that environment should be judged by whether those rails attract real institutional and machine-driven economic activity — not simply by price momentum.
The financing opportunity is expanding — but so is the need for discipline
AI infrastructure is producing enormous financing requirements across power, land, equipment, data centers and networks. That creates opportunities for private credit, structured finance, infrastructure capital and real estate.
But large contracted revenue numbers and marquee counterparties cannot replace underwriting. Capital providers still need to understand who ultimately pays, where power comes from, what happens if technology changes, how concentrated the customer base is and whether supposedly independent counterparties are economically tied to one another.
The RIC Takeaway
AI and capital are no longer running on parallel tracks.
They are beginning to merge into a single machine-native financial infrastructure.
AI is becoming capable of interpreting information and taking economic action. Blockchain is making assets and settlement programmable. Payment networks are creating credentials for autonomous software. Compute is beginning to develop financial markets of its own. Traditional institutions are acquiring the security and data infrastructure required to participate.
The question for businesses is no longer simply, “How do we use AI?”
The larger question is:
What happens when the systems making decisions can also interact directly with the systems that move capital?
That is where AI stops being a productivity tool and starts becoming part of financial infrastructure.
For Real Innovative Capital, that convergence is the story worth following.
Because when intelligence, capital and execution begin operating on the same rails, the opportunity is not only to automate existing finance.
It is to redesign how capital moves.
— Real Innovative Capital
Sources & Further Reading
SEC tokenized-securities framework / CoinDesk, September 17, 2026Real stocks are finally coming on blockchainS&P Global — OpenZeppelin acquisition, September 17, 2026S&P Global Announces Agreement to Acquire OpenZeppelinS&P Global — Strategic investment in Kaiko, September 14, 2026S&P Global investment in KaikoRipple / XRP machine payments, September 17, 2026Ripple adds XRP payments to Stripe and Tempo’s AI standardRipple Treasury — Governed AI for enterprise treasury, September 10, 2026Ripple Treasury GSmartLiquid Compute / WSJ, September 16, 2026Startup building an exchange for AI compute powerCrusoe financing / TechCrunch, September 17, 2026Crusoe raises $3.9BSB Energy / Financial Times, September 18, 2026OpenAI’s listing delay raises stakes for SoftBank’s data-centre IPO
Comments