Three weeks ago I sat in a Nairobi co-working space at 3 a.m., watching a $40 million liquidation cascade rip through a mid-cap altcoin. My main price feed froze. My backup feed lagged eight seconds behind. My third feed โ the one I actually trust โ kept printing every tick like nothing was happening.
That feed was Kaiko.
I didn't know it then. I found out the next morning, when the wire dropped that S&P Global had led a fresh $110 million extension to Kaiko's B-round, joined by BNP Paribas, Bpifrance, Broadridge, and a name I'm still trying to pin down: Canton.
I didn't read it as a funding story. I read it as a confession. The most boring companies in finance just told you where the next war is being fought, and it isn't on a chart. The chart lies. The crowd feels. But the data pipe never blinks โ and somebody just paid nine figures to make sure it never does again.
Kaiko is not a household name, and that is the entire point.
The Paris-based firm has spent the better part of a decade doing unglamorous work: pulling raw trade prints, order-book depth, and on-chain activity from hundreds of venues, cleaning all of it, and piping it through APIs to institutional clients who need numbers they can defend in a compliance memo. CoinGecko gives you a vibe. Kaiko gives you a receipt.
Four years ago, Kaiko raised a $53 million B-round. That was a different market. Bitcoin was clawing out of a crypto winter, "institutional adoption" was a slide in a pitch deck, and the round went quiet almost immediately. The company shipped. Nobody wrote about it, because there was nothing dramatic to write about.
Now the same company has extended that B-round to $110 million, and the cap table reads like a G20 guest list. S&P Global โ the index and data giant that owns the fixed-income benchmark universe and Platts โ took the lead. BNP Paribas, the largest bank in the EU and a systemically important institution, came in. Broadridge, the plumbing behind most American bond settlement and market-making, came in. Bpifrance, the French state investment bank, came in. And Canton โ a name that could be anything from a Hong Kong family office to a tokenization consortium โ came in too.
The stated use of funds is blunt: strengthen the market data business, and build infrastructure for a 24/7 tokenized market.
That second phrase is the entire story. Everything else is punctuation.
Here is what "24/7 tokenized market" actually means once you strip the jargon.
Traditional finance runs on a clock. Equities open at 9:30 and close at 16:00 in New York. Bonds trade in an even narrower window. When the bell rings, price discovery stops. You cannot trade, you cannot hedge, and you sit with your risk until the next session.
Crypto never closes. It runs 168 hours a week across every timezone, and that has produced a strange mutation: a market that never sleeps but also never fully prices anything, because half the participants wake up to a move that already happened.
Now tokenize a Treasury bond. Put it on-chain. Let it trade around the clock with instant settlement. What happens to price discovery?
You have just imported the crypto problem into the most liquid asset class on earth.
This is the hidden war inside the Kaiko raise. Not "crypto goes mainstream." Not the RWA narrative. The real question is: who provides the reference price for an asset that trades continuously, across venues, in instruments that do not share a settlement layer?
I have spent the last six years watching feeds break during exactly the moments they matter most. Building a 24/7 data spine requires three things Kaiko is being funded to deliver โ and each is harder than the last.
Continuous collection and cleansing comes first. A traditional market data vendor gets a nightly batch window. Kaiko does not, because the market never stops producing garbage. When a Korean exchange prints a fat-finger trade at 04:00 UTC, someone has to flag it as an outlier in milliseconds, or that print contaminates every downstream index that references it.
Millisecond distribution is the next gate. Institutional desks do not subscribe to a webpage. They subscribe to a stream. The latency budget between an on-chain event and a defensible price is measured in milliseconds, and every hop you add โ a normalization layer, a signing step, a geographic relay โ costs you the trade.
And a unified quote across multi-asset, multi-timezone inventory is the hardest of the three and the least discussed. A desk holding tokenized Treasuries, tokenized money-market funds, and spot crypto needs a single coherent book. That book does not exist yet. Kaiko is being paid to build it.
And here is the part that should register: none of this is technically novel. The novelty is the contract.
S&P Global does not invest $110 million for an API. S&P Global invests for the methodology. The firm has spent fifty years learning how to construct a benchmark that courts, regulators, and counterparties all agree to trust. That trust is the actual product. Data aggregation is a solved problem. Data legitimacy is not.
I have audited enough on-chain feeds to know the technology is rarely the bottleneck. The bottleneck is the moment a compliance officer asks, "Show me why this number is correct." A CoinGecko widget cannot answer. An S&P-endorsed feed can.
That is the trade. Kaiko brings coverage and latency. S&P brings the paperwork that makes the number admissible.
Now the layer the announcement does not say out loud.
The crypto data market is not consolidated. It is a food court. CoinGecko owns retail eyeballs. Nomics chased audit-grade transparency before fizzling. Messari sold research. Dune and Nansen sold on-chain query access. Each slices a slightly different cut. Meanwhile the venues themselves โ Binance, Coinbase, OKX โ each publish their own "official" feeds, none of which agree with each other.
For a retail trader, this is fine. You pick a feed, you live with its lies.
For an institution, it is fatal. You cannot run a risk system on a number that three vendors disagree about by 40 basis points.
So the real product Kaiko is selling is not data. It is reconciliation. The promise that when your risk engine and your counterparty's risk engine look at the same asset, they see the same number.
I lived through a version of this in 2022, during the Terra collapse. Different exchanges printed different LUNA prices during the death spiral. Arbitrage was impossible to execute because nobody could agree on a reference. That is the world without a reconciliation layer. Not a crash โ a freeze.
Smile while the liquidity drains. That is what a fragmented data layer produces at the worst possible moment. Everybody stares at a different screen, and nobody can trade.
A 24/7 tokenized market makes this problem permanent. You cannot halt a chain. You cannot impose a circuit breaker on a tokenized Treasury. So the reconciliation has to happen continuously, inside the feed itself, before anyone trades. The order book never sleeps. The market makers do. And that gap is where Kaiko wants to live.
That is why Broadridge is on the cap table. Broadridge runs the back office for a huge slice of American bonds. It knows, better than almost anyone, what breaks when settlement speed outruns reference data. If Broadridge is funding a 24/7 data spine, Broadridge has already run the numbers on tokenized bond trading โ and does not like what it sees without one.
I will be honest about one thing: I do not know what Canton is. I have asked around. The name sits in the gray zone between a Hong Kong family office, an Asia-focused tokenization vehicle, and something tied to a specific chain. The ambiguity itself is informative.
Look at the geographic spread of the syndicate: S&P Global (US), BNP Paribas (EU), Broadridge (US), Bpifrance (France), Canton (Asia, maybe). That is not a coincidence. That is a multi-jurisdictional regulatory hedge dressed as a cap table.
A 24/7 tokenized market cannot launch in one country. MiCA in Europe, the SEC's evolving posture in the US, Hong Kong's licensing regime, Singapore's MAS sandbox โ each has its own rulebook for what counts as a security token and who may custody it. A data provider that wants to serve all of them needs a footprint and a blessing in all of them.
BNP Paribas's participation does something specific here. As a systemically important bank, BNP's legal team does not sign off on investments casually. When BNP writes a check, its compliance department has already concluded that Kaiko's model fits inside a foreseeable regulatory framework. That conclusion is worth more than the money.
Nobody announces a quiet regulatory blessing. They just add it to the cap table.
Let me put on the cynical hat, because the market is going to misread this.
The consensus take will be: "Institutional money flows into crypto data โ bullish for infrastructure." True. But it misses the structure.
A $110 million B-round extension, led by a strategic acquirer-type, in a company that has operated quietly for years, usually means one of two things: Kaiko is being groomed for a much larger round or an IPO, or Kaiko is being groomed for acquisition.
S&P Global's existing product line โ Capital IQ, Market Intelligence โ has a glaring hole where crypto should be. Building that in-house is expensive, slow, and risky. Buying it is cheap, fast, and safe. A strategic investment with board rights is the standard first step.
I have watched this movie. IBM and Red Hat. Salesforce and Slack. The strategic round comes first, the "deepen our partnership" press release comes second, and the acquisition announcement comes third โ usually about eighteen months later, usually on a Friday.
That does not make the funding bad news. It makes it asymmetric news. The upside for Kaiko's customers is real whether the outcome is independence or integration. But if you are pricing this as a pure adoption signal, you are pricing it wrong.
There is a second blind spot. S&P leading the round does not only validate Kaiko โ it signals to every other traditional data giant that the crypto data layer is now contested territory. When Bloomberg decides it wants crypto market data, Bloomberg does not invest in a startup. Bloomberg builds.
The same network effects S&P brings โ trust, distribution, methodology โ also give S&P an incentive to eventually own the whole stack. If Kaiko becomes indispensable to S&P's crypto roadmap, S&P has every reason to absorb it. And if S&P absorbs it, the remaining independents get squeezed.
Data infrastructure does not tolerate many winners. It tolerates two or three, and then it consolidates. Kaiko just chose a side.
A word on the market you are actually trading in.
This is a bear market. I do not care what the weekly candle claims. Liquidity is thin, narratives are exhausted, and most protocols are still bleeding LPs. In that environment, the correct question is not "when moon." It is "which infrastructure survives."
My read: the data layer survives. It always does. Exchanges blow up. Tokens go to zero. Protocols fork and die. But the plumbing that tells everyone what an asset is worth keeps getting funded, because without it nothing else functions.
So if you are trying to figure out where attention should go, stop counting which Layer 2 has the most TVL. Look at which data provider your favorite protocol depends on. That dependency is the one that matters, and it just received a $110 million vote of confidence.
Now let me drain some oxygen, because I have also seen a decade of tokenization decks that promised 24/7 nirvana and delivered a pilot program.
The 24/7 tokenized market is, today, a rounding error. The RWA sector holds maybe tens of billions in tokenized value, not trillions. Most of it sits inside permissioned wrappers that trade during banking hours anyway โ because the legal rails underneath still run on a 9-to-5 clock in Delaware or Luxembourg.
The narrative is running ahead of the fundamentals. That is normal at the start of a cycle. It is also exactly the gap where retail gets hurt โ buying the story before the infrastructure is real.
Kaiko's raise does not close that gap. It funds the attempt to close it. Those are different things.
The next twelve months will tell us whether this is a data story or an acquisition story.
Watch three signals. First: does S&P Global integrate Kaiko data into a named product โ Capital IQ, Market Intelligence, anything โ or does the relationship stay at the press-release level? Named products mean commitment. Vague partnership language means the option is still open.
Second: does Kaiko publish a system availability SLA? A genuine 24/7 infrastructure company eventually has to defend its uptime in public. If nobody is publishing numbers, the infrastructure is still a slide deck.
Third: what happens between Broadridge and bond tokenization? If Broadridge starts piping Kaiko data into its settlement network, tokenized Treasuries just moved one step closer to reality. That is the thread worth pulling.
The chart lies. The crowd feels. And the companies that survive a bear market are the ones nobody writes about until after they have already won. Kaiko just got paid to stay invisible. That is not a bug. That is the business.