Two Events. One Day. The Whole Story.
On March 24, 2026, two things happened at the same time. OpenAI killed Sora. And Anthropic launched Claude computer use.
One company gave up on making pretty videos. The other gave AI the ability to control your computer.
That tells you everything about where AI is going.
It is worth revisiting what we wrote then, because the reasoning still holds. In September 2024, when every AI newsletter was breathlessly covering Sora demos and Midjourney was the most exciting word in tech, we published a piece called Image and Video AI Is Impressive. But Here's Where the Real ROI Lives.. People told us we were wrong. Some told us we were naive. A few told us we just did not understand the technology.
Eighteen months later, Sora is dead and Claude is controlling desktops. We understood the technology just fine. What we understood better was the business model -- or in Sora's case, the total absence of one.
What We Said in September 2024
In September 2024, we wrote that the real value in AI was not in pixels. It was in text, reasoning, and code -- the boring stuff that actually runs businesses. We argued that the AI labs were chasing Hollywood when they should have been chasing Main Street. We said that impressive demos do not equal business value, and that the market would eventually correct the hype.
We followed that up in July 2025 with Where the Real Business Value in AI Lives, where we doubled down. By then, Sora had launched as a standalone app and the initial excitement was already fading. The downloads were declining. The content moderation problems were piling up. The revenue was embarrassing. We pointed all of this out and said the trajectory was obvious.
Some people agreed with us. More people told us we were underestimating OpenAI. That Sora would find its market. That consumer video AI was a billion-dollar opportunity waiting to be captured.
It was not. It was a content moderation nightmare that generated $2.1 million in in-app purchases. For a company that is burning billions. That is not a rounding error -- it is a rounding error of a rounding error.
Why Sora Failed
The key point: Sora did not fail because the technology was bad. The demos were genuinely stunning. The first time we saw a Sora-generated video of a woman walking through a Tokyo street, we were as impressed as everyone else. The physics simulation, the lighting, the coherence across frames -- it was a remarkable technical achievement.
Sora failed because there was no business model.
The numbers tell the story with painful clarity. According to app analytics data, downloads peaked at 3.3 million in November 2025, right after the standalone app launched with a wave of media coverage and influencer hype. By February 2026, downloads had cratered to 1.1 million. That is a 67% decline in three months. For a product backed by the most recognized name in AI.
The revenue was even worse. Total in-app purchases over the product's life: $2.1 million. To put that in perspective, OpenAI spends more than that on compute in a matter of hours. The unit economics were not just bad -- they were non-existent. Every video generated was a money-losing transaction at a scale that would make even the most aggressive growth-stage startup blush.
Then there was the content problem. Deepfakes. Copyrighted characters. What the media started calling "video slop" -- low-quality, AI-generated content flooding social platforms and diluting the information ecosystem. OpenAI built a machine that was technically impressive and practically uncontrollable. The moderation costs alone were likely exceeding the revenue by orders of magnitude.
And the partnerships that were supposed to legitimize the whole thing? Disney, which had been exploring a $1 billion investment in OpenAI, pulled back. When Disney walks away from your creative technology, the market is sending a signal.
Sora's death is not a surprise. It is the inevitable conclusion of building impressive technology without a viable path to revenue. We said this was coming. Not because we are smarter than OpenAI's engineers -- we are not -- but because we talk to businesses every day and we know what they will and will not pay for.
Enterprise AI Platform Share (Q1 2026)
Distribution of enterprise AI platform adoption across major providers
OpenAI / ChatGPT
Anthropic / Claude
Google / Gemini
Other
Source: Market analysis estimates, Q1 2026
The Real Scorecard: Anthropic vs OpenAI
While OpenAI was pouring resources into video generation and watching it collapse, Anthropic was doing something less glamorous and far more consequential: winning the enterprise market.
The numbers as of March 2026 are stark.
Anthropic is winning 70% of new enterprise deals. Their revenue run-rate has reached $14 billion as of February 2026. Month-over-month business adoption is growing at 4.9%, according to Ramp's AI Index. A quarter of all businesses using AI -- 24.4%, according to Ramp -- are now on Anthropic. Deloitte rolled out Claude to 470,000 employees. That is not a pilot program. That is a full-scale enterprise deployment at one of the largest professional services firms on the planet.
OpenAI, by contrast, is projecting $14 billion in losses for 2026. One company is building a business. The other is burning one.
OpenAI saw its largest single-month decline in business adoption share in recent months, while Anthropic posted its strongest growth. The market is making its choice, and the market is choosing the company that focused on making AI useful for work over the company that focused on making AI generate spectacle.
And here is the part that really matters: OpenAI is now pivoting to enterprise products and preparing for an IPO. They are moving toward the market that Anthropic already owns. They are not pivoting from a position of strength. They are pivoting because Sora failed, because consumer AI alone cannot support their cost structure, and because the enterprise revenue that Anthropic has been building for two years is the only viable path to the numbers they need for a public offering.
Good luck. Anthropic has a two-year head start, a better product for business use cases, and the trust of the companies that are writing the biggest checks. Pivoting into an entrenched competitor's stronghold is not a strategy. It is an admission.
Where Does Google Fit?
We would be doing you a disservice if we did not talk about the quiet third player in this race. Google is not getting the headlines that OpenAI and Anthropic are getting, and that might be exactly the point.
While OpenAI was building Sora and Anthropic was winning enterprise deals, Google was doing something characteristically Google: playing the infrastructure game. They launched Gemini Enterprise for Customer Experience in January 2026. They have been building out multi-agent architectures and autonomous enterprise agent platforms. They are not chasing consumer viral moments. They are building the plumbing.
This is a smart play. It is not as exciting as Claude Code -- which, full disclosure, is the backbone of everything we build. But Google has the distribution, the cloud infrastructure, and the enterprise relationships to be a major force in AI-powered business operations. They do not need to win the model race if they win the integration race.
Our read on Google is this: do not sleep on them. They are not going to win by having the best model or the flashiest demo. They are going to win specific verticals by being deeply embedded in enterprise workflows through Google Cloud, Workspace, and their existing tooling. For certain types of businesses -- especially large enterprises already in the Google ecosystem -- Gemini's enterprise agent platform could be a natural fit.
The three-way dynamic right now is clarifying: Anthropic leads on model quality and enterprise AI adoption. Google leads on infrastructure and integration. OpenAI leads on... brand recognition and pivot announcements.
What This Means for SMBs
If you are running a small or mid-sized business, the Sora shutdown and the broader market shift is actually good news for you. Here is why.
The AI landscape is consolidating around practical business tools, not spectacle. For the past two years, the AI conversation has been dominated by "look what AI can do" demos -- generated videos, generated images, generated music, generated everything. It was exciting and almost entirely irrelevant to how most businesses operate. Now that the spectacle era is ending, the resources and attention are shifting to the tools that actually help businesses get work done.
Claude for reasoning, coding, and document analysis. Google for enterprise infrastructure and workflow integration. And OpenAI... finding its identity.
For SMBs, this clarity simplifies your decisions. You do not need to evaluate twenty different AI tools with overlapping capabilities and unclear differentiation. The market is sorting itself out. The winner is not the company with the best demo reel. It is the company whose tools help you process contracts faster, write code more efficiently, automate workflows that used to eat up your team's time, and make better decisions with the data you already have.
If you have been sitting on the sidelines waiting for the AI market to mature before making a move, this is the maturity you were waiting for. The hype is clearing. The products are getting better. The pricing is getting more competitive. And the winners are becoming obvious.
Our Position -- And Why It Matters for Your Business
We want to be transparent about where we stand, because our position is also our track record.
OneWave bet on Anthropic in mid-2024 when the overwhelming consensus was that OpenAI was the only serious option. We built our entire consulting practice around Claude, MCP servers, and AI agents. When browser-based coding tools were the hot trend, we went the other direction and committed to CLI agents -- Claude Code specifically. When everyone was talking about image and video generation, we were building document processing pipelines and workflow automation systems.
Those early decisions are shaping how we work with clients today. Not because we were lucky, but because we were listening to what businesses actually needed instead of what tech Twitter was excited about.
We are not saying this to brag. We are saying it because it matters for your business who you align with. The AI consultancy that was chasing Sora integrations six months ago is now scrambling to pivot. The one that was building Midjourney workflows for clients is watching those tools become commoditized. The one that defaulted to OpenAI for everything is now watching OpenAI's enterprise pivot play catch-up to the platform their competitor already mastered.
We are none of those. We are the consultancy that bet on the boring, valuable, business- critical applications of AI from day one. We bet on text over pixels. We bet on reasoning over rendering. We bet on agents over art. And the market came to us.
The Lesson
The lesson of Sora's death is not that OpenAI is a bad company. They are not. They have brilliant engineers and they will remain a major player in AI. The lesson is that in AI, as in every other technology wave, the gap between impressive and valuable is enormous. And the companies that close that gap -- by building tools that solve real problems for real businesses -- are the ones that win.
Anthropic understood this. Google, in its own way, understood this. OpenAI is learning it now, at the cost of a failed product and a humbling pivot.
We understood it 18 months ago. And we built everything we do around that understanding.
If you are a business owner trying to figure out where AI goes from here, the answer is simpler than it has ever been. AI is going where the value is. Not where the demos are. Not where the headlines are. Where the value is.
Sora is dead. Claude is controlling computers. Google is building enterprise agents. The future of AI is not generated video. It is generated productivity.
The future of AI is not spectacle. It is leverage. And there is a lot more to build.