DEV Community

Tonmoy Chandra Mudi
Tonmoy Chandra Mudi

Posted on

Anthropic’s Road to $2 Trillion

Anthropic, the company behind Claude, is preparing for what could become one of the largest technology IPOs ever. But there’s something particularly interesting about how investors are thinking about its valuation.

According to a recent Reuters report, Anthropic is projecting approximately $190–$200 billion in revenue by 2028.

That number is enormous.

But the more interesting question isn't:

“Can Anthropic make $200 billion?”

It is:

“How do investors decide what Anthropic is worth today based on revenue it might generate two years from now?”

Let's break it down from a technology and business perspective.

🚀 From $9B to $47B+ Revenue Run Rate

Anthropic's growth has been extraordinary.

According to the Reuters report, Anthropic's revenue run rate was approximately:

Period Revenue Run Rate
End of 2025 ~$9B
May 2026 >$47B
Projected 2028 ~$190–$200B

The important concept here is revenue run rate.

A run rate isn't necessarily the same thing as actual annual revenue.

For example, if a company generates $5 billion in a particular quarter, someone might annualize that:

$5B × 4 = $20B annualized run rate

It basically answers:

“If the company continued operating at this current pace, what would its annual revenue look like?”

That distinction matters when evaluating rapidly growing companies.

💰 How Can Investors Value Anthropic?

Traditional companies are often valued using metrics such as:

Revenue
EBITDA
Net income
Free cash flow
Price-to-earnings ratio

But AI companies like Anthropic are unusual.

They're spending enormous amounts on:

GPUs
Data centers
Model training
Model inference
Researchers
Engineers
Cloud infrastructure

So their current profits don't necessarily tell the entire story.

Instead, investors may focus heavily on future revenue.

One common approach is:

Enterprise Value ≈ Revenue × Revenue Multiple

For example, imagine a hypothetical company producing:

$10B revenue

and investors believe companies like it deserve a:

20× revenue multiple

Then:

$10B × 20 = $200B

That gives a rough enterprise value.

Of course, real valuation models are significantly more complicated.

📈 Why 2028 Matters

Here's where Anthropic becomes particularly interesting.

Investors aren't only looking at what Anthropic generates today.

They're asking:

“What could Anthropic look like when the AI infrastructure investment starts producing operating leverage?”

Suppose Anthropic reaches:

$200B revenue

in 2028.

If the market eventually assigns a hypothetical:

10× revenue multiple

the valuation would be:

$200B × 10
= $2 trillion

At:

15×

it becomes:

$200B × 15
= $3 trillion

At:

20×

it becomes:

$4 trillion

These aren't predictions.

They're simply examples showing why future revenue assumptions have such a huge impact on valuation.

🤖 The Real Problem: AI Is Expensive

There's a major difference between building traditional SaaS software and operating frontier AI models.

A normal SaaS application might have relatively predictable infrastructure costs.

AI inference can be dramatically different.

Every time a user sends a request to a large language model, the company potentially needs:

User

API Request

GPU / Accelerator

Model Inference

Generated Tokens

Response

At massive scale, those GPU cycles become a significant operating expense.

And training frontier models requires another enormous investment.

So Anthropic has an important challenge:

Can revenue grow faster than compute and operating costs?

That's one of the biggest questions behind the valuation.

⚙️ The AI Scaling Equation

A simplified way to think about the business is:

Revenue Growth

More Customers

More API Usage

More Inference

Higher Compute Costs

But eventually, investors want something different:

Revenue Growth

More Customers

More AI Usage

Better Hardware + Better Models

Lower Cost per Token

Higher Gross Margin

Operating Leverage

That final transition is extremely important.

If Anthropic can make each unit of AI computation increasingly efficient, revenue could eventually grow much faster than costs.

🏢 Why Cloudflare and Palantir Matter

Reuters reports that investors are considering companies such as Cloudflare, Palantir and SpaceX as reference points.

Why?

Because there isn't a perfect publicly traded company that looks exactly like Anthropic.

Each provides a different comparison.

Cloudflare

Cloudflare represents a high-growth infrastructure/software business.

It helps investors think about:

Infrastructure
Developer platforms
Recurring revenue
High growth
Scaling technology businesses
Palantir

Palantir provides another interesting comparison because of its strong connection to:

Enterprise AI
Government technology
Data platforms
High-growth software
SpaceX

SpaceX is particularly interesting because its valuation is heavily connected to expectations about its future scale rather than simply today's financial results.

🧠 The Bigger Question: Is AI Creating Real Productivity?

This might be the most important question.

AI companies can generate enormous revenue.

But investors eventually need to understand where that money is coming from.

Are businesses paying billions because AI:

replaces expensive manual work?
makes developers significantly more productive?
automates customer support?
improves research?
accelerates software development?
creates entirely new products?

Or are companies simply experimenting with AI because they don't want to fall behind?

There's a huge difference.

If AI produces substantial economic value, enormous valuations could eventually make sense.

If AI spending doesn't translate into proportional productivity gains, valuations could face serious pressure.

⚠️ The Risk Behind a $2 Trillion+ Valuation

A high valuation isn't automatically a problem.

The problem is what assumptions are already priced into it.

Imagine investors value Anthropic at:

$2T

based partly on expectations of:

$200B revenue

That means investors are effectively saying:

“We believe Anthropic can become an enormous, highly valuable business.”

But what happens if revenue reaches only:

$120B

instead of:

$200B

Or compute costs remain extremely high?

Or competitors become stronger?

Or open-source models become good enough for many use cases?

Or AI pricing falls dramatically?

The valuation could change very quickly.

🔥 Why This Matters to Developers

If you're a software engineer, this isn't just Wall Street news.

It affects the technology ecosystem you're building your career in.

The money flowing into AI is influencing:

GPU demand
Cloud infrastructure
AI startups
Developer tools
APIs
Open-source models
AI agents
Data engineering
MLOps
Backend architecture

For developers, the interesting part isn't simply:

“Anthropic might be worth trillions.”

The interesting part is:

“What technology needs to exist for a company like Anthropic to generate $200B of revenue?”

That question leads directly into some of the most important engineering problems of the next decade.

🛠️ What Engineers Should Learn

If you're preparing for a software engineering or AI engineering career, I'd pay attention to these areas:

  1. Python

Still one of the most important languages for AI engineering.

  1. Backend Engineering

Learn how APIs, authentication, databases, queues and distributed systems work.

  1. Cloud Infrastructure

Understand:

Containers
Docker
Kubernetes
Cloud Computing
Networking
Observability

  1. Machine Learning

Learn the fundamentals before jumping directly into frameworks.

Understand:

Linear Algebra
Probability
Statistics
Optimization
Machine Learning
Deep Learning

  1. LLM Engineering

Then move toward:

Transformers
Embeddings
Vector Databases
RAG
Agents
Fine-tuning
Inference
Evaluation

  1. AI Infrastructure

This is where things get particularly interesting.

Learn about:

GPU computing
Distributed systems
Model serving
Inference optimization
Caching
Batching
Quantization

Because eventually, AI engineering isn't just about building models.

It's about making those models economically useful at enormous scale.

🌍 The AI Economy Is Still Being Built

Anthropic's potential IPO is another sign that the AI industry is moving from an experimental technology phase toward a massive commercial infrastructure phase.

But the $190–$200 billion 2028 revenue projection is exactly that:

a projection.

The future depends on many variables:

AI adoption

  • Enterprise spending
  • Model capabilities
  • Compute efficiency
  • Competition
  • Pricing
  • Infrastructure costs
  • Productivity gains = Future economics of AI

Nobody knows the final answer yet.

And that's what makes this era so interesting.

We're not just watching companies compete.

We're watching an entirely new technology economy being built in real time.

💭 Final Thought

The most interesting number in the Anthropic story isn't actually $200 billion.

It's the assumption behind it.

Investors are effectively betting that AI will become important enough, useful enough, and economically productive enough to support businesses operating at an extraordinary scale.

Whether that bet is correct will depend not only on AI models...

but on engineers building the infrastructure, applications and products that turn those models into real economic value.

And that's exactly why learning AI engineering today is so interesting. 🚀

Source: Reuters, August 15, 2026 — reporting on Anthropic's projected 2028 revenue and IPO valuation framework.
This post is an independent educational analysis and not investment advice.

ai #anthropic #claude #machinelearning #softwareengineering #artificialintelligence #technology #programming #llm #aiengineering

Top comments (0)