We have been tracking AI—from silicon to deployment—for the past 10 weeks. Finally, we have reached the question that matters to investors: Does this technology generate profit?
Unlike the early days of this decade, AI has stopped being just an infrastructure and has moved on to becoming revenue-generating. It’s not just tokens or GPUs anymore; companies are selling productivity alongside automation and completed work.
So, let’s dive into this week’s discussion: computing meets the cash register.
Missed Tesla? This Pre-IPO Move Could Be Even Bigger
Elon Musk’s early investors saw 1,210,000% gains with Tesla over 2 decades.
But we believe history is about to repeat itself…
There’s a new visionary founder who’s being hailed as “The Next Elon Musk.”
His company is already working with every major branch of the military…
Backed by $26 billion in contracts…
And supported by Peter Thiel’s $1B vote of confidence.
Right now, you have a rare chance to get in through a little-known 4-letter ticker symbol—before it goes public.
You won’t find this opportunity on Robinhood or the evening news.
From Assistant to Employee
The global market is experiencing a massive shift. AI is quickly evolving from helpful tools into agents capable of completing tasks from start to finish.
Salesforce Agentforce is a prime example. As customer-facing agents, it charges $2 per conversation. For internal agent usage, the rate is $125 per employee per month. It also offers flex credits: $500 per 100,000 credits with $0.10 per action.
So, what is happening? AI has been converted into metered digital labor.
Microsoft’s broad adoption play involves charging $30 per user per month for Copilot, while Google plans to embed AI into everyday tools by offering Gemini as a workspace for $2 per user.
Pricing will soon be based on activity rather than headcount. Investors must understand this to make the most of their AI investments.
The $40 Million Case Study That Changed the Game
This layer, backed by hard operating numbers, gives a clear perspective:
Klarna is the proof:
2.3 million – customer conversations in the first month.
66.7% – support chats covered.
700 – approximate number of full-time agents replaced.
Under 2 minutes – resolution time reduced from 11 minutes.
Down 25% – repeat inquiries.
$40 million (2024) – estimated profit improvement.
Down 40% (Q1 2023 → Q1 2025) – customer service cost per transaction.
These numbers significantly improve margins. Productivity data suggest a similar pattern: Microsoft Copilot users save up to 14 minutes per day, and GitHub-controlled experiments indicate a 55% improvement in coding task completion time. Small daily gains turn into significant revenue when multiplied by thousands of workers.
McKinsey anticipates over $2.6 trillion in annual revenue from generative AI adoption in marketing, sales, customer operations, and R&D.
Three Pricing Models, One Winner
Application vendors increasingly adopt one of three models:
Per-Seat Subscription
Microsoft Copilot – $30 per user per month.
GitHub Copilot – $10–$39 per month.
Works best when usage is predictable.
Usage-Based Pricing
OpenAI GPT-5 mini – $0.25/M input tokens, $2/M output tokens.
Focuses on consumption rather than input.
Hybrid Pricing
Salesforce – blends per-user plans with per-conversation and per-action fees.
Combines base subscriptions with usage meters for actions, conversations, or credits.
As long as the meter aligns with business results, investors can profit from hybrid models, allowing revenue expansion even if employee count stays constant.
Where the Real Pricing Power Lives
For margins, focus on vertical AI—agents specialized in specific domains like healthcare, finance, IT, and compliance. Vendors embed AI into systems of records, e.g., Intuit Assist across TurboTax and QuickBooks. Customers get used to effective automation, making pricing durable and switching costs high. Vertical products win by reducing time, errors, and labor hours. Prioritize vendors with strong operational metrics over eye-catching user counts.
The Two New Revenue Pools
Profit is no longer limited to subscriptions and usage. Data licensing offers a fresh profit stream:
Reddit: $203 million in AI data licensing contracts (IPO disclosure).
News organizations, including the Financial Times, have entered content agreements with OpenAI.
Data archives are now a recurring revenue stream that improves the reliability of AI.
Your Investment Cheat Sheet
Who Wins:
Companies controlling workflows and distribution: Microsoft, Salesforce, ServiceNow, Palantir, and vertical AI vendors.
Key Signals to Monitor:
Customers expanding spend (ideally >110% revenue retention).
Faster billings signaling acceleration.
Unit metrics: falling cost per ticket, falling resolution time, rising revenue per employee.
AI tools with <30% churn delivering durable value.
Reduced integration time from pilot to production: 9 months for large enterprises, 3 months for mid-sized.
Winners:
Platform players with existing distribution (Microsoft, Salesforce).
Vertical AI specialists (healthcare, finance, compliance).
Hybrid pricing models aligned with business results.
Organizations with large data archives.
Losers:
Per-seat-only pricing models.
High-churn tools without durable value.
Horizontal tools treated as commodities.
Red Flags:
Net Revenue Retention <100%.
Churn >30%.
Long integration timelines.
Bullish Signals:
NRR >110%.
Faster integration time.
Growing data licensing revenue.
The Bottom Line
AI has evolved from a builder of engines into a seller of outcomes. Klarna’s $40 million profit lift, Salesforce’s $2 per conversation pricing, Microsoft’s $30 Copilot upsell, and Reddit’s $203 million data licensing illustrate the same conclusion: the economic center of gravity is shifting downstream.
The chips enable the boom, but applications decide the winners. Everyday investors should focus here: from raw silicon to revenue generation, the supply chain secrets are now clear.
With this, we wrap up the AI Supply Chain series, offering a full understanding of the value chain from silicon to profit.
Important disclosures: This newsletter is provided for informational purposes only and does not constitute investment advice. All investments involve risk, including possible loss of principal. Please consult with your financial advisor before making investment decisions.
