SaaS is Dead. Long Live De-SaaS

A recent [post] (https://x.com/JulienBek/status/2029680516568600933) by Julien Bek made an important observation: “The next generation of companies will not sell software. They will sell services powered by AI.”

• Instead of selling accounting software, they close your books.
• Instead of selling legal software, they generate your contracts.
• Instead of selling productivity tools, they do the work for you.

In other words: Services are becoming the new Software.

I agree with this. But I think the bigger shift is still being missed. The real transformation is not just Services replacing Software. The real transformation is Decentralized Services replacing SaaS infrastructure.

I call this De-SaaS—Decentralized Software-as-a-Service. And it changes everything.

---

SaaS Was Just a Phase

For twenty years the dominant model in tech was SaaS. You rented software instead of owning it.

• CRM systems
• Marketing tools
• Accounting platforms
• Design software

Everything became a subscription. The cloud companies became the toll collectors of the digital economy. Every SaaS company quietly depended on three centralized infrastructures:

1. Cloud compute
2. Cloud storage
3. Cloud networking

AWS, Azure, and Google Cloud became the invisible backbone of the entire SaaS economy.

But AI has exposed a fundamental problem with this model. Compute is now the most expensive part of modern software. Training models. Running inference. Processing massive datasets. Cloud costs are exploding, and the cloud providers control the pricing.

Meanwhile the World Is Full of Idle Compute

Look around:

• Phones
• Gaming PCs
• Edge routers
• Desktops
• Enterprise machines

Most of them are idle most of the time. We have already deployed the largest compute network in history. We just never connected it.

Instead, we keep building centralized data centers. That approach cannot scale with the explosion of AI demand. What needs to happen instead is simple: Turn the world’s devices into a distributed compute swarm.

GNUS.ai Is a Swarm Operating System

GNUS.ai was designed as exactly that. Not a blockchain project. Not a cloud competitor. A Swarm Operating System for global compute.

It orchestrates AI workloads across millions of devices:

• GPU inference
• AI agents
• LLM applications
• Rendering
• Simulation
• Games

Workloads are distributed across available compute resources. Verification happens cryptographically. Payments happen automatically through the token system.

The result is something cloud infrastructure cannot provide: A global market for compute. Anyone can supply compute. Anyone can consume compute. Pricing is determined by supply and demand. Costs collapse.

De-SaaS: Software Running on Global Compute Markets

Now combine this with Bek’s thesis. Services replacing software. But those services still need infrastructure.

Today that infrastructure is centralized cloud. Tomorrow that infrastructure will be decentralized compute networks. Which means the next generation of services will run on something different.

Not SaaS. De-SaaS. Applications that run across distributed compute swarms instead of centralized cloud platforms create an entirely new economic structure. Instead of paying AWS billions per year, that value flows to the network participants who provide compute.

The Consulting Industry Is Next

There is another disruption hiding inside this model: Consulting. Today if a company wants to implement AI systems they hire consulting firms like McKinsey, Accenture, or Deloitte. These firms charge millions to design infrastructure and deploy AI solutions.

But what if the infrastructure already existed? What if companies could simply deploy AI services directly onto a global compute network?

This is where our Joint Venture model comes in.

The Joint Venture Model

Instead of selling software licenses, Genius Ventures forms joint ventures with partners:

• Telecoms
• ISPs
• Enterprise groups
• Regional infrastructure owners

Each JV activates compute already deployed across their networks (routers, edge servers, customer devices, enterprise hardware). The hardware already exists. The missing piece is orchestration.

GNUS.ai provides the SwarmOS layer that turns those assets into AI compute infrastructure. The JV then becomes something much more powerful than a software customer. It becomes an AI services company overnight.

Instant AI Service Companies

Once the infrastructure is activated, the JV can launch services immediately:

• AI inference platforms
• LLM agent services
• Data processing pipelines
• Gaming compute
• Rendering services
• Industry-specific AI tools

Instead of buying expensive cloud infrastructure, they deploy services directly on the swarm. Instead of building software companies, they build AI service companies. And those services run on De-SaaS infrastructure.

Why This Disrupts Consulting

Consulting firms sell two things: Expertise and Implementation.

But when the infrastructure is already built and the services are modular, the need for expensive consulting disappears. Companies no longer need a team of consultants designing AI architecture. They deploy services directly onto the swarm.

Infrastructure, compute, verification, and payments are already built into the platform. The JV structure lets regional partners move far faster than traditional consulting models. It also aligns incentives. Partners become owners of the AI infrastructure, not just customers of it.

The Next Technology Platforms

Every major era of computing had a dominant infrastructure layer:

1. Mainframes
2. Personal computers
3. The internet
4. Cloud computing

AI will create the next layer: Distributed compute networks. The companies that build those networks will not look like traditional software companies. They will look more like global economic systems for compute: markets where services run on top of decentralized infrastructure.

SaaS Was the Cloud Era

SaaS defined the cloud era. But the next era will not be centralized. It will be distributed.

Services will replace software. Compute will come from the network. Infrastructure will be owned by participants instead of hyperscalers.

That is the shift from SaaS to De-SaaS. And we are just beginning.