In February 2026, roughly $2 trillion in software company market value disappeared in about 30 days. The cause was not a recession, a fraud scandal, or a sudden collapse in corporate spending on technology. It was the fear that artificial intelligence had become capable enough to make entire categories of business software irrelevant. Wall Street coined a word for it: the SaaSpocalypse.
The companies doing the fearing were the ones that sell software subscriptions - think Salesforce, ServiceNow, the whole model of charging businesses a monthly fee per employee seat to access tools for managing customer relationships, workflows, or data. The threat was simple: if an AI agent can do what those tools do, and it can do it without requiring a human to click through a dashboard, why keep paying for the seats?
Rob Goldstein, the chief operating officer of BlackRock - the world's largest asset manager, meaning the company that manages investments on behalf of other people and institutions, currently holding roughly $14 trillion of its clients' money - sat down recently with Bloomberg's Odd Lots podcast to explain why, from where he stands, the SaaSpocalypse looks less like a crisis and more like a vindication.
The background
BlackRock was not always the $14 trillion colossus it is today. When Goldstein joined in 1994, the firm had about 80 employees and $19 billion under management. That is smaller than many regional pension funds today.
What made BlackRock unusual from the start was not its investment philosophy but its relationship with data. The firm's founders were pioneers in structured products - complex financial instruments built around things like mortgage payments - at a time when the big banks were using expensive supercomputers to model them, then literally faxing yield tables to clients around the world. Faxing. Tables. Of numbers.
BlackRock's bet was different. In the late 1980s, you could buy ten Sun workstations for around $10,000 each, link them together, and get computing power that previously only a multi-million-dollar supercomputer could provide. The founders did exactly that, and built risk management models that banks could not offer. That early insight - that asset managementis fundamentally an information-processing business - shaped everything that followed.
The product that came from that insight was Aladdin. The name stands for Asset, Liability, Debt, and Derivative Investment Network, which tells you almost nothing about what it actually does. In plain terms, Aladdin is a platform that tracks risk across investment portfolios. It calculates, in near real time, how exposed a portfolio is to different economic scenarios - rising interest rates, falling currencies, a collapse in a particular sector. It does this not just for BlackRock's own investments but for external clients too: pension funds, insurance companies, sovereign wealth funds, and banks. According to data tracked by business analytics sources, Aladdin now processes risk analytics for approximately $25 trillion in assets - roughly 7 to 8% of the entire global financial system, and nearly triple BlackRock's own holdings.
Over 1,000 organizations use the platform. It has become, in effect, the nervous system of institutional finance.
What is actually happening
The question the podcast posed to Goldstein was a pointed one: if AI can now generate code, automate workflows, and replace the human workers who drive software subscription revenue, what exactly is Aladdin's moat?
A moat, in investment language, refers to whatever structural advantage stops competitors from eating into a company's business. For most software companies, the moat was habit - people kept paying because switching was painful and the products were woven into daily work. AI threatens that kind of moat directly by making the habit replaceable.
Goldstein's answer was precise. He drew a distinction between two kinds of software. The first kind aggregates public information and makes it easy to access - it is essentially a convenience layer. No proprietary data. No deep workflow integration. Just a nicer way to search and display things that exist elsewhere. That kind of software, he argued, is genuinely in trouble. AI tools are better at that job than any subscription product ever was.
The second kind is different. Aladdin - and he drew a parallel to the Bloomberg Terminal, the data and analytics platform used across the financial industry - sits inside the actual workflow of institutions managing real money under regulatory scrutiny. Clients pour their most sensitive data into it. Permissions are granular: one person can see certain portfolios, another can trade, a third can only confirm trades. The compliance obligations are not a bug. They are, in Goldstein's framing, the product.
This is the counterintuitive core of his argument. Finance is a heavily regulated industry, meaning every significant decision is subject to oversight by government bodies and must be documented, auditable, and explainable. The same regulatory burden that makes financial services slower and more expensive than other industries turns out to be a shield against AI disruption - at least the chaotic, fast-moving variety. AI is non-deterministic, as Goldstein noted: put in the same question twice and you may get different answers. Traditional software is not like that. Write a line of code correctly and it produces the same result a million times in a row. For risk management - where being wrong by a small percentage on a trillion-dollar portfolio has catastrophic consequences - that reliability matters enormously.
According to a Fortune analysis published in April 2026, the SaaSpocalypse is not an overstatement as a direction of travel, even if the timeline is uncertain. The forces behind it are structural. But the same analysis noted that enterprise systems embedded in regulated workflows - the exact kind Aladdin represents - face fundamentally different risk than convenience software. Analysts at Forrester have argued that vertical- or domain-specific platforms controlling proprietary data are the most likely survivors of the shift.
Goldstein went further. He described a real example from the previous Friday: a cross-functional team at BlackRock - portfolio managers, risk professionals, engineers, and product managers - sat together for several hours mapping out how a new tool should work. That meeting was recorded. The recording was transcribed into a functional document. The document was fed into an AI coding tool. Within days, rather than months, a working prototype emerged from the process. Not a thin shell, Goldstein stressed - a prototype that held up when he pressed it.
The unit of measurement for software development just changed. That is either terrifying or exciting depending on which side of the moat you are standing on.
The money trail
Follow the money in this story and you find something worth sitting with. BlackRock makes money in two ways: from managing investments and from licensing Aladdin to external institutions. The technology revenue from Aladdin was approaching $2 billion in annual contract value as of 2025, with 24% revenue growth in that year. That is a meaningful line of business, but it is also the line most exposed to the SaaSpocalypse narrative.
Here is the tension. If AI tools make it easier to build custom risk management software, the argument goes, why would a pension fund keep paying Aladdin's licensing fees? Build your own.
Goldstein's response was to describe what "build your own" actually entails. The permission architecture alone - knowing which of thousands of employees can see which portfolios, trade which instruments, access which data - is itself a years-long engineering project. Aladdin has been building and refining that control layer for decades. The switching cost is not just what you pay to leave; it is the cost of replicating everything you would lose.
Then there is the data. Clients are not just using Aladdin's software. They are feeding it their most sensitive position data, risk exposures, and trading history. That data makes the platform more useful for them over time. It also makes leaving much harder - you would need to port that history somewhere, which is, as Goldstein put it, a lot of work.
The iShares software ETF IGV fell more than 21% year-to-date through late March 2026, and an estimated 40% of enterprise IT budgets are being reallocated from traditional software subscriptions to AI platforms and token usage. That reallocation hurts convenience software directly. But it may actually benefit Aladdin - because as more of a client's workflow runs through AI agents, the value of having a central, governed, permissioned platform to anchor those agents to becomes greater, not smaller.
Goldstein articulated this explicitly. He imagined a future where an AI agent - rather than a human employee - becomes the primary user of Aladdin, accessing its capabilities and executing its logic. The agent would still need the control plane. The permissions would still matter. The audit trail would still be required. BlackRock would still be in the middle of it.
That is a very elegant position to be in. The world shifts to AI, and your product becomes the foundation that AI builds on.
What people are doing about it
BlackRock's internal response to AI has been methodical rather than experimental. Goldstein described a "first draft principle" - a policy under which every document produced inside the firm, from client presentations to regulatory prospectuses, begins its life as an AI-generated draft. The draft then passes through 16 layers of human review. The principle is not a productivity hack. It is a controlled introduction of a new technology into a compliance-heavy environment.
That process is now yielding measurable acceleration. The prototype described above - built from a recorded meeting to a working tool in days rather than months - is the kind of result that finance firms are chasing across the board.
Separately, the firm has been expanding Aladdin's openness through what Goldstein called an "open Aladdin" campaign, launched roughly a decade ago: making the platform accessible via APIs - programming interfaces that let external tools plug into Aladdin's data and logic without requiring a human to log into a dashboard. That architectural decision looks prescient now. It is exactly the kind of surface an AI agent needs to access an enterprise platform.
Meanwhile, the broader financial industry is grappling with a genuine compute problem. Goldstein acknowledged that BlackRock's token consumption - tokens being the units of text that AI models process, and which companies pay for by volume - has risen by multiples over the past year, without anyone yet knowing how high it will go or how to optimize it. He described the current moment as equivalent to before the national anthem - the actual game of enterprise AI implementation has not started yet.
According to Sumitomo Mitsui Trust Asset Management, which adopted Aladdin in February 2025 to manage $620 billion in assets, the appeal is precisely that consolidation: one platform, one workflow, one permission architecture, across a full range of asset classes. That is the whole-portfolio logic Goldstein kept returning to. The asset management industry spent decades organizing itself by asset type - separate firms for fixed income, equities, private credit, and so on - and then handed clients the puzzle and told them to assemble it themselves. The firms that can offer the whole picture are winning mandates.
The bottom line
The SaaSpocalypse is real for software that was always just a convenience wrapper around public information. For platforms embedded deep in regulated financial workflows - with proprietary data, compliance architecture, and years of client history baked in - the threat looks different. AI may actually increase the value of those platforms by making the agents that sit on top of them more capable, while leaving the underlying control infrastructure as indispensable as ever. BlackRock built its entire history around the idea that information processing is the core of asset management. It has spent 30 years building the infrastructure. The irony is that the technology wave most threatening to the software industry may be the one that finally justifies the scale of what BlackRock built.
Timeline
- Late 1980s - BlackRock's founders identify that linking commercial workstations can replicate the computing power of million-dollar supercomputers; they use this to build risk models for structured financial products
- 1994 - Rob Goldstein joins BlackRock; the firm has approximately 80 employees and $19 billion in assets under management
- 2013 - Aladdin manages approximately $11 trillion in assets, roughly 7% of global financial assets, across around 30,000 portfolios
- 2018 - BlackRock launches its internal AI lab
- ~2015 - BlackRock begins "open Aladdin" initiative, opening the platform to external developers via API access
- 2020 - Aladdin grows to approximately $21.6 trillion in assets on the platform
- December 2024 - Microsoft CEO Satya Nadella argues on a public podcast that SaaS applications will "collapse" in the AI agent era
- February 2025 - Sumitomo Mitsui Trust Asset Management adopts Aladdin to manage $620 billion in assets
- Late 2025 - Salesforce reports a rare revenue miss; software sector growth begins visibly slowing
- January 29, 2026 - OpenAI releases Project Operator; ServiceNow beats earnings for the ninth consecutive quarter and its stock falls 11% in the same session - markets reprice the model, not the results
- February 2026 - The term SaaSpocalypse is coined by a Jefferies trader; approximately $2 trillion in software market capitalization is wiped out in 30 days; the iShares software ETF IGV drops more than 21% year-to-date
- December 2025 - Aladdin reaches approximately $25 trillion in assets on the platform, with technology annual contract value approaching $2 billion and 24% revenue growth for the year
- April 30, 2026 - BlackRock COO Rob Goldstein appears on Bloomberg's Odd Lots podcast to discuss AI, Aladdin, and the future of enterprise software in finance
Summary
Who: Rob Goldstein, COO of BlackRock, the world's largest asset manager
What: In a wide-ranging interview on Bloomberg's Odd Lots podcast, Goldstein argued that the SaaSpocalypse - the AI-driven collapse in software company valuations - will not destroy platforms like Aladdin, and may in fact strengthen them by making their underlying compliance and data infrastructure more valuable as AI agents proliferate
When: The interview published April 30, 2026, against the backdrop of a software sector that had lost more than $2 trillion in market value since February
Where: The broader story spans BlackRock's origins in late-1980s New York, its three decades building Aladdin, and the global financial industry's current reckoning with AI
Why: The rise of AI agents threatens the per-seat subscription model that powered the software industry for two decades; however, platforms embedded in regulated, workflow-critical financial infrastructure face a different threat profile - and may benefit from the same wave that is punishing their competitors