The Stickiest Software on Earth Just Took a 100 Million Dollar Punch

Here's the deal: if you ask any finance operator which piece of enterprise software never changes, they'll say ERP without blinking. And inside ERP, the general ledger is the most frozen layer of all. Once it's installed, it stays for a decade or more. Ripping it out means explaining yourself to your auditor, migrating years of journal entries, and living with the knowledge that one bad mapping decision can wobble an entire fiscal year of financial statements. So companies complain and stay put. Not because what they have is great, but because switching feels like elective surgery on a beating heart.

Into that market, on August 19, 2026, walked Rillet with a 100 million dollar Series C at a 1 billion dollar valuation. ICONIQ led. The cap table behind it reads like a roll call: Sequoia, Andreessen Horowitz, Sequoia Global Equities, Bain Capital Ventures, Oak HC/FT, Battery Ventures, FirstMark, Scale Venture Partners, and Creandum. Rillet came out of stealth in 2024. Two years later it's a unicorn.

The pace is the part that should make you sit up. Rillet raised a 25 million dollar Series A led by Sequoia in May 2025. Ten weeks later, on August 6, 2025, it raised a 70 million dollar Series B co-led by a16z and ICONIQ. One year after that, this Series C. By the company's own count, that's three rounds in 14 months and more than 200 million dollars raised in total. Venture capital has been generous to anything with AI in the deck, but accounting software raising at this clip is genuinely unusual.

The label Rillet has slapped on the plan is "accounting superintelligence," which sounds like a lot until you unpack it. What they mean: a real-time general ledger, AI agents doing finance work on top of it, every action wrapped in human approval and a full audit trail. The destination is killing the month-end close as a concept. Sequoia's phrase from the Series A post was "zero-day close" — not closing faster on close day, but never being open in the first place.

Four Parties — The Ones Rebuilding the Ledger, The Ones Funding It, The Ones Defending It

Start with Rillet itself. Co-founder and CEO Nicolas Kopp used to run N26's US business. His co-founder, Stelios Modes, architected N26's payments infrastructure. Neither came out of the accounting software industry, and that's the origin story. As Kopp put it around the Series B, simple requests took weeks because the systems were stuck in the past. They built the product for a pain they personally ate while scaling a bank.

The money arrived in three layers. Sequoia went first, leading the 25 million dollar Series A in May 2025 and publishing a post arguing that a decade of fintech had unbundled every part of the ERP stack except the one at the center — the general ledger. Sequoia framed the choice facing scaling companies as a bad binary: stay on tools you've outgrown, or graduate to what it called arcane, inefficient systems that take ten or more specialists to run. Land in the second bucket and you're spending 15 to 20 days a month closing the books.

Layer two was a16z and ICONIQ, co-leading the Series B in August 2025 and putting a16z general partner Alex Rampell and ICONIQ general partner Seth Pierrepont on Rillet's board. a16z sized the opportunity at 500 billion dollars, deliberately counting software licenses, services, and the manual labor being replaced. It described incumbent systems as brittle, clunky, and deeply manual, with finance workflows duct-taped across Excel, NetSuite, and point solutions. The line that stuck: why does the financial nervous system of a company still run on software built for Windows 95?

Layer three is ICONIQ leading again in this round. A firm that was already inside, with a board seat, writing the biggest check yet, is a signal worth reading — that's an investor who has seen the internal numbers doubling down. Pierrepont called Rillet the clear market leader in AI-native accounting infrastructure, and said what stands out isn't the demo but how customers operate: multi-billion-dollar businesses running finance teams a tenth the traditional size, with books closing continuously.

Then there are the defenders. Oracle NetSuite and Sage Intacct. NetSuite was founded in 1998, went public in 2007, and was bought by Oracle for 9.3 billion dollars in 2016; third-party trackers put its footprint at more than 40,000 customers across 200-plus countries. Sage Intacct was founded in 1999 and sold to Sage for 850 million dollars in 2017, with customer counts commonly cited north of 10,000. Between them they've split mid-market accounting for close to two decades. Against those numbers, Rillet's 600 customers is a rounding error. Whether that rounding error is a beachhead or a ceiling is the actual question in this story.

What Actually Happened — Three Rounds in 14 Months, Customers Up 3x

The short version of the numbers: 100 million dollar Series C, 1 billion dollar valuation, more than 200 million dollars raised in total, third round in 14 months. On the business side, Rillet says it's past 600 customers and doubled new ARR in the last three months. Named customers in this announcement include Mercor, Function Health, and Temporal. At the Series B, the company pointed to Postscript — over 100 million dollars in ARR, closing its books in three days — and Windsurf running finance with a two-person team.

One caveat worth flagging before you get carried away. What Rillet disclosed is "new ARR doubled in three months," not an absolute ARR figure. Small bases make big multiples easy. The company said essentially the same thing at the Series B — ARR doubled in 12 weeks — back when it had 200 customers. In other words, Rillet consistently publishes growth rates and withholds levels, which means nobody outside can compute what multiple 1 billion dollars represents. That's not unique to Rillet; most AI infrastructure rounds work this way right now. But it does mean anyone telling you this valuation is cheap or expensive is guessing.

The round history makes the velocity clearer.

Round Announced Amount Lead Customers disclosed at the time
Out of stealth 2024 Undisclosed — —
Series A May 28, 2025 25 million dollars Sequoia Not disclosed
Series B Aug 6, 2025 70 million dollars a16z + ICONIQ (co-led) 200-plus
Series C Aug 19, 2026 100 million dollars ICONIQ 600-plus

On product, Rillet leans on three claims. First, a real-time general ledger: instead of scraping data into a month-end snapshot, transactions from source systems like Salesforce and Brex land in the ledger as they happen. Second, a continuous close architecture, where closing is a state rather than an event. Third, AI agents that operate with human approval and full audit trails. That third one carries the most weight, because automation an auditor won't accept is worthless in accounting. Rillet also markets implementation speed — four weeks versus the twelve months it attributes to legacy rollouts. That's a vendor claim and hasn't been independently verified.

Quietly, distribution got built too. Rillet launched an alliance with EY in April 2026 and says it's now an official partner with more than half of the firms on Accounting Today's top 20 CPA list. Given that ERP is a channel game more than a software game, that line may matter more than the round size. Half the reason NetSuite went unbeaten for twenty years wasn't the product — it was the consultant ecosystem calcified around it.

Who Gets What, and What They're Risking

For Rillet, the prize is time. ERP is a product where the customer takes six to twelve months just to decide, and the vendor has to pre-hire sales and implementation staff to survive that lag. A hundred million dollars is fuel for the gap. The unicorn label is itself a sales asset, too. The thing a CFO fears most when handing over the company's ledger is whether the vendor still exists in three years. A billion-dollar valuation and 200 million in the bank function as an answer. In accounting software, capital is less about performance than about being a proxy for trust.

ICONIQ gets position. Entering at the Series B and leading the Series C means it accumulated ownership at a lower blended cost with a board seat already locked in. For Sequoia and a16z, which led the A and B respectively, this round is largely pro rata defense against dilution. All three are betting the same way: that this is one of the few B2B categories where AI substitutes directly for headcount rather than just assisting it. The fact that a16z's 500 billion dollar market figure explicitly includes manual labor gives away the thesis.

Customers — finance teams — get headcount relief. The US accounting talent pipeline has been short for years. Qualified accountants are hard to hire, and the ones you get burn out during close season and leave. So to a CFO, "cut your close team from three to one" doesn't read as cost savings, it reads as a fix for a recruiting problem. ICONIQ's line about teams a tenth the traditional size is exactly that pitch. The flip side is real, though: fewer people also means fewer eyes reviewing AI-generated journal entries, and how auditors weigh that tradeoff is still unsettled.

Fintechs like Brex and Ramp benefit sideways. In January 2026 Brex shipped an AI-native Accounting API and launched it with Rillet and Campfire as first partners — real-time webhooks and two-way data flow instead of batch syncs. For Brex, that's a direct pipe from its transaction data into the ledger. To these companies, AI-native ERPs aren't competitors, they're distribution. The slower legacy vendors are to open comparable interfaces, the tighter that alliance gets.

And the losers are worth naming. The consulting partners who built careers implementing NetSuite and Intacct made money precisely because implementation was hard. If deployment collapses to four weeks, twelve months of billable work evaporates. Rillet signing EY and half of Accounting Today's top 20 looks like a deliberate move to co-opt that resistance rather than fight it head-on.

This Has Been Tried Before — One Worked, One Quietly Died

Start with the success, because it's ironic: the incumbent defending the hill today was the insurgent yesterday. NetSuite launched in 1998 telling companies to run accounting in a browser instead of buying servers, and the SAP and on-premise Oracle establishment treated it as a toy. It IPO'd in 2007, nine years in. Oracle bought it for 9.3 billion dollars in 2016, eighteen years in. Being right about cloud took two decades to pay off. Sage Intacct is the same shape — founded 1999, sold to Sage for 850 million dollars in 2017. Both eventually won. Both took longer to win than the average venture fund's life.

The failure is Kenandy. Founded in 2010 by Sandra Kurtzig, a genuine Silicon Valley legend who built ASK Computer Systems in the 1970s and pioneered manufacturing MRP software. On paper there was no reason to lose. Kleiner Perkins led the first round, Salesforce invested, the company raised more than 50 million dollars and at its peak carried a 350 million dollar valuation. The ending came in January 2018, when Rootstock — a competitor building on the same Salesforce platform — acquired it. Terms weren't disclosed, and the industry read it as consolidation rather than a win.

Why Kenandy stalled tells you what to watch here. The product wasn't bad and the founder was elite, but ERP has never been a market you win by being better. Winning requires an implementation partner network, industry-specific accounting compliance, reports in formats auditors already recognize, and above all customers in enough pain to abandon what they have. In the mid-2010s, a company on NetSuite was annoyed but not bleeding. So it didn't move. Kenandy never manufactured a reason to switch now.

Is Rillet different? There are four places it could be. One, the nature of the pain changed. It used to be "this UI is ugly." Now it's "closing requires ten people and I can't hire ten people." Two, migration cost itself dropped — chart-of-accounts mapping and historical journal transfers that used to consume months of consultant time are substantially model work now. Three, source data already flows over APIs. Salesforce, Stripe, Brex, and Ramp are all open, so integration doesn't cost what it did. Four, hypergrowth AI companies don't have twelve months to spend on a legacy rollout, so they pick new vendors by default. Even if all four hold, the open question is what share of NetSuite's 40,000-plus accounts actually moves. Six hundred customers is a hypothesis, not yet evidence.

How the Competition Punches Back

The most direct response came from Oracle. At SuiteWorld in October 2025, NetSuite unveiled NetSuite Next and Autonomous Close — names that describe exactly what Rillet sells. Instead of cramming work into period end, it monitors transactions continuously, flags anomalies, auto-posts predefined entries like rent, depreciation, and payroll accruals, and auto-matches bank, AR, and AP activity against the ledger. Oracle has said internal testing handled up to 98 percent of routine transactions automatically, which is a vendor-reported figure and should be treated as such. Early previews went to select customers in late 2025, with broader rollout tracked across 2026 into 2027.

Sage is moving the same direction. Sage Intacct 2026 Release 1 shipped February 13, 2026 with a Finance Intelligence Agent and an Import Agent, and in April 2026 Sage formally announced an expansion of AI agents across finance, HR, and operations. The architecture puts Sage Copilot as a natural-language front end over Close, AP, Time, and Assurance agents underneath. NetSuite and Sage are both running the same play: if you want AI accounting, don't move your ledger, just switch it on where your ledger already lives. In a market with brutal switching costs, that's a strong card.

Same-generation startups are crowding in too. Campfire raised a 35 million dollar Series A led by Accel in July 2025, then a 65 million dollar Series B co-led by Accel and Ribbit on October 15, 2025 — twelve weeks later — pushing total funding past 100 million dollars. It claims a proprietary large accounting model hitting 95 percent accuracy on reconciliations and variance detection, with customers including PostHog, Decagon, and Replit. Rillet and Campfire were named side by side as the first partners on Brex's accounting API, which tells you they're fighting over the same accounts.

A different angle comes from Numeric, which started in close management and raised a 51 million dollar Series B led by IVP in November 2025, bringing total funding to 89 million dollars as it expands into a broader finance platform. Its strategy is the inverse of Rillet's: leave NetSuite in place and layer close automation on top. That's a far lower-risk ask for a buyer, so adoption friction is lower. The cost is that Numeric doesn't own the ledger, which leaves it exposed to being squeezed out of the value chain later.

Fintechs come at it from yet another direction. Ramp launched Ramp Stack on June 3, 2026, an AI operating system for accounting firms targeting a market it sizes at roughly 150 billion dollars. Stack builds and maintains recurring schedules for fixed asset depreciation, prepaid amortization, and deferred revenue, and posts the resulting journal entries each period. QuickBooks Online is the first integration, with NetSuite and Sage Intacct on the roadmap. So Ramp isn't building a ledger — it's replacing the people who work on top of one. Brex chose alliance instead, shipping its accounting API with AI-native ERPs as launch partners, while Mercury stays in the banking and treasury layer as a data supplier. Net-net, the market has split four ways: replace the ledger (Rillet, Campfire), sit on top of the ledger (Numeric), defend the ledger and turn on AI inside it (NetSuite, Sage), and own the data outside the ledger (Ramp, Brex, Mercury).

So What Actually Changes

If you work in finance or accounting operations, this is the most immediate. Over the next two years, when you sit down to pick an ERP, the shortlist likely grows from "NetSuite or Intacct" to "NetSuite or Intacct or Rillet or Campfire." What to evaluate isn't demo polish. It's multi-entity consolidation, revenue recognition treatment, evidence extraction in the format your auditor demands, and the approval-and-reversal flow for AI-generated journal entries. That last one especially: show it to your actual audit firm and get an answer in writing before you sign. Accountability for machine-produced entries is not a settled industry standard yet.

If you're an engineer wiring up internal finance systems, the integration model is shifting. ERP connectivity has lived in a world of nightly batches and CSV uploads. Brex opening a real-time, webhook-driven, two-way API with Rillet and Campfire as launch partners signals this layer moving from batch to event stream. If you're designing an accounting data pipeline now, drop the assumption that syncing happens at month end. But note the corollary: a real-time ledger means real-time errors. When a bad event posts instantly, idempotency and correcting-entry design matter far more than they used to.

If you're investing, treat this round as a test with three specific readouts. First, how many customers are public companies or above roughly 500 million dollars in revenue. Rillet says it has publicly listed customers but hasn't given a count, and moving upmarket is what actually displaces NetSuite. Second, absolute total ARR. As long as only growth rates get published, the billion-dollar mark can't be checked. Third, churn after NetSuite's Autonomous Close reaches general availability across 2026 and 2027. If Oracle bundles equivalent capability into existing contracts at effectively no extra cost, we'll find out how much of Rillet's differentiation survives. Calling the outcome today is premature.

And if you're a normal working person with no connection to accounting, there's still an indirect read. Accounting has long been filed under jobs AI can't touch, because it's regulated, accountability is explicit, and mistakes become legal problems. The fact that an approach built on audit trails and human approval is now pulling in this much capital suggests the same pattern could be applied to other regulated professions. Whether it delivers is unproven. The direction isn't ambiguous.

🥄 Three Things You're Probably Wondering

— So what does this mean for me? If you're not in finance, nothing immediate. But if you touch expense workflows or revenue recognition at your company, there's a decent chance you'll be working inside a system that has no concept of "wait for month end" within a few years. What changes isn't the tool, it's the rhythm of the work.

— Why is this happening now, of all times? ERP is the market that never moves, but three things landed at once: a shortage of accounting talent, AI cutting the cost of migration itself, and a fintech stack already pushing source data over APIs. If Kenandy failed in the 2010s because there was no reason to switch now, the bet here is that the reason finally exists. Whether that bet is right is still open.

— Has Rillet actually beaten NetSuite? No, not close. Rillet has 600-plus customers; NetSuite is tracked at more than 40,000. That's roughly one percent. And Oracle is actively pitching Autonomous Close as a reason to keep your ledger where it is. If Rillet wins, it probably won't be on feature superiority — it'll be from net-new companies that are AI-native from day one choosing it first. Whether that flow reaches public-company scale is too early to call.

Sources

Numbers and criteria are as of announcement and may change. Investment calls are yours to make!