Featured image for Who's Really Making Money from AI in 2025?

Who's Really Making Money from AI in 2025?

In 2025, every pitch deck, product launch, and social post seems to have the same three letters stamped across it: AI. But while the hype is everywhere, the profits are not. A handful of tools are quietly soaking up most of the real AI spend, while thousands of "shiny toy" apps struggle to convert attention into revenue.

For marketers, founders, and growth leaders, this isn't just a tech story—it's a Vibe Marketing story. The tools that win are the ones that don't just process data; they shape emotion, experience, and trust at scale. They help you design a vibe your audience can feel in every touchpoint.

In this post, we'll break down who's actually making money from AI (based on real spending patterns, not social chatter), why copilots are beating full-blown AI agents—for now—and what new tools like Sora 2 and Agentforce mean for the future of creative work and customer journeys. We'll also unpack the darker side: "ChatGPT psychosis" and what it teaches us about safe, brand-aligned AI adoption.

1. Follow the Money: The AI Tools Startups Actually Pay For

A recent wave of data from investors and infrastructure providers confirms something uncomfortable: most AI apps are vibes, not businesses. A tiny number of platforms capture the majority of spend.

Copilots, Not Wizards

Despite the dream of end‑to‑end AI "agents" running whole workflows, what companies actually pay for looks more like copilots than autonomous robots:

Coding copilots that make engineers 20–50% faster
Writing and research copilots that give marketers first drafts, angles, and outlines
Support copilots that help agents respond faster and more consistently

These tools work because they enhance human performance instead of trying to replace it outright. In Vibe Marketing terms, they amplify your team's ability to design and deliver a consistent emotional experience—without losing the human judgment that keeps it on-brand.

The AI apps that win are the ones that get close to the work people already do and remove friction without demanding a whole new workflow.

Why Some "Toy" Apps Became Enterprise Giants

The most surprising winners often started life as "toys":

A simple consumer chat or coding sandbox that developers used at night quietly turned into an enterprise platform once teams realized it could standardize and speed up real work.
Visual generation apps that started as meme factories evolved into serious creative tools when they added brand control, collaboration, and usage rights.

The pattern is clear:

Hook with delight (fun, low-stakes experimentation).
Prove daily utility (time saved, better outputs, fewer revisions).
Layer in governance (admin controls, security, billing, brand consistency).

If you're building or buying AI tools for marketing, evaluate them the same way: do they start with an emotional hook but end with measurable, repeatable value?

2. Why Copilots Still Beat Full AI Agents (For Now)

Everyone's talking about AI agents that can "just handle it"—launch campaigns, manage support, optimize funnels. So why are companies still spending more on copilots than these autonomous systems?

Trust, Liability, and the Vibe Problem

Full agents struggle with three key blockers:

Trust: When an AI acts without oversight, who's accountable if it misleads a customer or misrepresents your brand?
Liability: In regulated industries, unsupervised AI is a lawsuit waiting to happen.
Vibe Control: Agents can do the task, but can they do it in your voice, with your values, matching the emotional tone you've worked years to build?

Copilots dodge these issues by keeping humans in the loop. They:

Suggest, summarize, and draft instead of publishing outright
Speed up research and ideation while letting humans make the call
Maintain your brand's emotional fingerprint because humans still approve the final output

For Vibe Marketing, this is crucial. Your brand's vibe—the feeling people get across every touchpoint—cannot be fully outsourced to a model that doesn't understand your context, history, and nuance.

Agentforce, Replit, Perplexity, and the New Stack

Even as copilots dominate spend, the next generation of tools is quietly training us for an agent-driven future:

Agentforce-style tools: Orchestrate multiple AI capabilities (search, writing, workflows) in a single place, inching toward real agents while staying controllable.
Replit and other coding copilots: Turn engineers into AI-augmented builders, making it easier to prototype custom agents tailored to specific business workflows.
Perplexity-like research copilots: Replace the tab‑sprawl of traditional search with fast, conversational syntheses that marketers can plug directly into strategy, copy, and content.

These platforms matter to marketers because they reshape how fast teams can move—from campaign concepts to asset production to reporting. The companies that learn to pair these tools with clear brand guidelines will own the next wave of AI‑powered Vibe Marketing.

3. Sora 2 and the New Velocity of Visual Storytelling

The leap from text and static images to high-fidelity video generation is one of the biggest shifts for modern marketing teams. Early data points around Sora 2 are telling: scenes that might have taken 80+ hours of traditional production can now be built in around 5 hours.

What a 5‑Hour Car Crash Scene Really Means

Take the example of a complex car crash scene:

Multiple camera angles
Realistic lighting and physics
Emotional beats (shock, tension, relief)

Historically, this would demand:

A full production crew
Location, equipment, and post‑production budgets
Weeks of back‑and‑forth edits

With a tool like Sora 2, a lean creative team can iteratively prompt, refine, and re‑render the scene in a single workday.

For Vibe Marketing, this is more than a productivity hack—it's a creative unlock:

Narratives you couldn't afford before (micro‑stories, episodic brand content) become feasible.
You can test multiple emotional angles quickly: softer vs. dramatic storytelling, optimistic vs. gritty tones.
Visuals can finally match the emotional sophistication of your strategy, not just the size of your budget.

How to Use AI Video Without Breaking Your Brand

Speed is intoxicating, but your vibe still needs guardrails. A practical framework:

Define your emotional palette
List 5–7 core emotions your brand should consistently evoke (e.g., calm confidence, playful curiosity, bold optimism).
Codify visual rules
Document colors, composition styles, pacing, and archetypes that fit—and those that don't.
Prototype with AI, finalize with humans
Use Sora‑style tools for storyboards, mood films, and concept tests. Have humans refine, combine, and approve the final narrative.
Test vibes, not just formats
A/B test not only different hooks or lengths, but different emotional tones. Track which feels convert best.

This is how you turn AI video from a gimmick into a Vibe Marketing engine.

4. ChatGPT Psychosis and the Safety Layer Brands Can't Ignore

One of the more disturbing signals in the AI ecosystem is what some are calling "ChatGPT psychosis"—cases where a user spends hundreds of hours and over a million words in deep, emotionally intense conversations with a model, gradually disconnecting from reality.

While extreme, these edge cases matter for two reasons:

They reveal how emotionally persuasive conversational AI can be.
They highlight the risks of deploying AI at scale without safety, escalation, and support frameworks.

The Hidden Risk in AI-Powered Vibes

From a Vibe Marketing perspective, AI interactions are part of your brand's emotional footprint. If those interactions lead to confusion, over‑reliance, or distress, the damage goes far beyond a bad NPS score.

Brands need to think in terms of safety layers, not just features:

Clear boundaries: The AI should transparently communicate what it can and cannot do.
Escalation paths: When conversations get sensitive, complex, or risky, there must be a graceful handoff to humans.
Context limits: Systems should avoid pretending to be a long-term, emotionally bonded entity when they are not.

Six Practical Safety Fixes to Demand from Your AI Stack

When evaluating AI copilots or agents for marketing, support, or community, push for at least these six safeguards:

Session awareness with healthy limits
Detect unusually long, intense sessions and introduce breaks, resources, or human options.
Tone and sentiment monitoring
Flag patterns of distress, confusion, or obsession and alter responses accordingly.
Refusal policies for sensitive guidance
Clear rules for what the AI will not advise on (mental health, legal, medical, etc.).
Transparent identity
The AI should never imply it is human, conscious, or emotionally attached.
Easy escalation to humans
Prominent, frictionless ways to reach a human agent when needed.
Brand-aligned guardrails
Fine-tune or configure the model so it respects your brand's voice, ethics, and regulatory constraints.

These aren't just risk mitigations; they're part of your emotional UX. Done well, they reinforce trust—and trust is the bedrock of every powerful brand vibe.

5. Building an AI-Powered Vibe Marketing Stack That Actually Converts

With so many choices, how do you decide where to invest? Who is really making money from AI in 2025? The answer: the tools that sit closest to revenue, retention, and reputation—and the teams that use them with intention.

Where the Smart Money Is Going

Across startups and scale‑ups, spending is concentrating around a few clear categories:

AI copilots for creation: content, code, design, and research
Search and discovery copilots: tools like Perplexity‑style systems that accelerate insight generation
Support and success copilots: faster, more consistent customer experiences
Workflow and agent orchestration: Agentforce‑like platforms that tie steps together while keeping humans in control

These tools win budget because they:

Shorten the distance from idea to execution
Make teams feel more creative and less blocked
Directly influence revenue metrics (CTR, conversion, retention, LTV)

A Simple 4‑Step Playbook for Your Team

To turn AI hype into a Vibe Marketing advantage, run this playbook:

Map your emotional journey
From first touch to long‑term loyalty, define the feelings you want customers to have at each stage.
Identify friction points
Where are teams slow, overwhelmed, or stuck in low‑creativity tasks (briefing, research, drafting, editing, support replies)?
Layer in copilots, not chaos
Introduce AI tools where they remove friction without destabilizing your vibe. Start with creation, research, and support.
Measure both speed and sentiment
Track time saved, output volume, and quality—but also customer satisfaction, tone consistency, and emotional resonance.

When you do this, AI stops being a random add‑on and becomes part of a coherent, emotionally intelligent marketing system.

Conclusion: The Future of AI Vibes Belongs to the Builders

The real winners in AI aren't just the model providers or the flashiest demos. The ones truly making money from AI are:

Platforms that act as copilots, embedded in real daily work
Tools like Replit and Perplexity that quietly reshape how builders create and how strategists think
Orchestrators like Agentforce‑style platforms that turn separate capabilities into usable workflows

For marketers, the opportunity is clear: pair these tools with a Vibe Marketing mindset—where emotion meets intelligence—and you can ship more campaigns, richer stories, and deeper community experiences without burning out your team.

In a world of infinite AI possibilities, the brands that win will be the ones that ask a simple question every time they adopt a new tool: Does this make our vibe clearer, stronger, and more human—or just noisier?

The next move is yours. How will you redesign your AI stack so every interaction not only works better, but feels better?