Open-Weight AI Handles 72% of Traffic but Only 27% of Spending. Who Gets Paid?

Companies don't send every package by courier, and they don't run every task on the top AI model. As cheaper models take over routine work, price and choice matter as much as raw performance.

Executives reviewing AI model usage and spend dashboard in a meeting room

The 20-second read

1
59% of executives at large companies said their AI spending came in well above what they expected or budgeted.
2
On OpenRouter, open-weight models handled 72% of traffic but accounted for only 27% of spending.
3
The new question in AI investing is not just how good a model is, but who keeps pricing power.

1. Does a company need the best AI every time?

Nobody pays for a courier when standard shipping will do. Fast delivery is for the package that truly can't wait. Companies are starting to choose AI the same way.

They use cheaper models for document sorting and simple summaries. They put top-performing models on complex coding and analysis. That does not mean every company is abandoning premium models. The value a company will pay differs by task.

AI is no longer judged only on how smart it is. It is also judged on what it costs to get the job done.

2. Cost pressure is changing model choice

Behind the shift is an AI bill that has grown faster than many expected.

Capgemini surveyed 500 executives at companies with annual revenue above $1 billion across 10 countries. 59% said their AI spending had significantly exceeded expectations or budget. Among those respondents, the average overrun was 24%.

59% of executives said AI spending significantly exceeded expectations or budget.

Prices differ widely. The open-weight models compared in the survey were 87% cheaper per token on average. For companies, that means a lower-cost option for similar work.

Still, half of respondents said they plan to keep using closed, high-performance models for complex tasks. They are not handing everything to the cheapest model. They plan to spend where performance is needed.

The enemy of the expensive model is not the cheap model itself. It is any task that cannot justify the higher price.

3. Does the most-used AI earn the most?

One set of numbers shows how AI revenue works. On OpenRouter, a model-routing platform, open-weight models handled 72% of total traffic in September 2026. Their share of spending was only 27%.

Same platform, different shares (OpenRouter, September 2026)

  • Open-weight model traffic: 72%
  • Open-weight model spending: 27%

The gap shows that usage and revenue do not necessarily grow at the same pace. If models that process many requests cheaply drive traffic higher, overall AI use can grow while spending per request falls.

Going forward, an AI company's report card may depend as much on how well it turns usage into revenue as on usage itself.

4. The investor question: who sets the price?

Providers of high-performance models, such as OpenAI, Anthropic and Google, face two tasks. They must keep their performance lead on hard tasks. They must also answer growing price pressure on easy ones. Launching a low-cost model or cutting prices can keep customers, but it also changes how the company earns money.

Cloud companies such as Amazon (AMZN), Microsoft (MSFT) and Alphabet's Google (GOOGL) have a different opening. They can act as a "model department store," helping customers compare multiple models and run them reliably. The contest may turn less on any single model's performance and more on how easily customers can pick the right model for each job.

Open-weight camps such as Meta (META) also gain importance. The more cheap alternatives there are, the harder it may be for top-tier model makers to hold prices high. The value of an open-weight strategy lies not only in direct model fees but in changing the competitive terms of the AI market.

AI profits may not flow to one smartest model. They may be split across the layers that build, choose and supply models.
What is an open-weight model?
  • It is an AI model whose weights are made public so anyone can download and use it.
  • In the article, it appears as a cheaper alternative to closed, high-performance models.
What share did open-weight models have on OpenRouter?
  • As of September 2026, they handled 72% of total traffic.
  • Their share of spending was 27%.
Why are companies splitting their work across AI models?
  • Their AI bills grew faster than expected.
  • In a Capgemini survey, 59% of executives said AI spending went well over budget. The average overrun was 24%.
If cheaper models grow, will high-performance models be pushed out?
  • That is hard to say. Half of respondents said they plan to keep using closed, high-performance models for complex tasks.
  • However, price pressure could grow for easier tasks.
Does the AI that gets used the most also earn the most?
  • Not necessarily. The gap between 72% of traffic and 27% of spending is an example.
  • When spending per request falls, revenue may not grow at the same pace as usage.

What to watch

Model prices and low-cost lineupsWatch whether major model companies cut prices or expand cheaper models. This shows how much the race to win customers affects pricing power.
Cloud companies' comments on multi-model revenueThis is a clue to whether services that help customers choose and run models turn into a real business opportunity.
Enterprise customers' remarks on AI budgetsApart from growing AI use, it is worth watching which tasks customers are moving to cheaper models.
Heavy AI usage does not guarantee heavy revenue, and profits may be shared across the layers that build, choose and supply models.

Insight Times Editorial Desk

More by topic

Starting with this story's topics, then what readers are following now

Investor psychology, decision-making and books on investing

META

This story's topic39 storiesSee all →

Meta Weighed Cutting Some Teams by 60% for AI, Then Backed Off. Why?

Meta tested whether AI doing half the work means half the staff. Employee pushback and weaker-than-expected productivity gains led it to drop further cuts.

Tech

After the Smartphone, the Contest Is Over Delegation, Not Screens

Meta Connect 2026 showed where AI glasses are heading. The question is who will handle users' tasks for them, and how much control users keep.

Tech

A person at a kitchen table looks at a food delivery app on a phone beside a laptop and a DoorDash delivery bag

AI Was Supposed to Do Our Shopping. DoorDash Got Only Several Hundred Orders From It

DoorDash logged 993 million orders in the third quarter, yet outside AI agents placed only several hundred in total. Orders built through its own in-app AI ran about 50% larger.

Tech

OpenAI's Dots Fights Back: Where the 24-Hour AI Worker War Will Be Decided

OpenAI's Dots, Meta's Muse and xAI's Grok Bot all keep working after the user leaves. Each is aiming at a different seat: the control plane of work, personal life context, or the AI org chart.

Tech

Nvidia's $500 Billion Experiment: Wall Street Asks Who Backs the GPU, Not Just the Chip

The real issue in Nvidia's financing platform is not chip sales. It is how far Wall Street will trust GPU resale prices and rental income three to four years out.

Tech

Why Is Meta Planning Up to $95 Billion in Stock Issuance?

Meta holds $82.45 billion in cash and securities, yet it has outlined an equity raise of up to $95 billion. AI data center spending is growing faster than the cash on hand.

Tech

"You Should Have Bought Only the Crashes": The Hindsight Trap in Big Tech Investing

Cherry-picked chart lows make everyone look like a genius. Nvidia and Meta show why waiting for the bottom and building time in the market are different bets.

Investor Mind

Cloud Computing

This story's topic17 storiesSee all →

Why Nebius Is Keeping Some 2027 AI Capacity Off the Shelf

Nebius mixes long-term contracts, short-term premiums and auctions. Holding back capacity pays only if the servers stay busy, and it is not yet clear the pricing will last.

Tech

Why AI Is Moving Back to the Laptop: Anduril and Qualcomm's BlueChip

Anduril and Qualcomm's new BlueChip laptops point to a bigger question: where should AI run? Production and sales are still unproven.

Tech

Dark navy cover. Flow diagram: Nvidia to Nscale (credit guarantee, in talks) to Anthropic (~$45B compute lease signed in August plus a separate ~$45B, ~460MW deal in talks). Title: Why Would Nvidia Back a Customer's Loan?

Why Would Nvidia Guarantee Its Customers' Data Center Loans?

Nvidia is reportedly in talks to back financing for Nscale's West Virginia data center. Nothing is signed, and the terms are not public.

Tech

Same cloud, three different ways to earn the money back

Microsoft, Alphabet and Amazon are all spending enormous sums on AI infrastructure. The route each one takes to recover that capital is not the same, and that is what decides the winner.

Investing

AI Needs to Find $1 Trillion a Year From Somewhere to Pay Back Its Capex

Saying AI capex is big is no longer the point. The harder question is what end users have to pay for the buildout to pencil out, and Goldman Sachs puts that number at roughly $1 trillion a year.

Tech

Anthropic Talked Slowdown. Its Actions Say Otherwise

A next-gen Claude model, an IPO push, and an actual biology lab. Put the three together and Anthropic's message isn't "stop AI." It's build stronger models while scaling up verification and control at the same pace.

Tech

AI is no longer competing on intelligence. It is competing on leftover cash

The race for better models continues, but the industry's center of gravity has moved. The question now is who can turn enormous data center and power spending into revenue and cash flow.

Investing

A building that houses servers and networking gear and supplies them with powe…

AI ran out of memory before it ran out of GPUs

The HBM bottleneck is spreading into DRAM prices, server costs and the price of the next laptop or phone.

Tech

Is Space Cooler for AI? Why Musk Wants Data Centers in Orbit

The next limit on AI may be power and cooling, not GPUs. Elon Musk is looking to Earth orbit for an answer, but the test is cost per hour of compute.

Tech

White House Swaps AI for "SI." The Real Investment Signal Is Infrastructure, Not the Name

An executive order renames artificial intelligence "Super Intelligence" and six frontier labs sign a voluntary safety accord. What investors should watch is the data center, power, nuclear and chip policy now gathered under one national-competitiveness frame.

Tech

Musk Sees the Next Phase of AI Infrastructure: Who Can Switch GPUs On, Fast

If neoclouds struggle under financial pressure, AI demand may not shrink. It may simply favor the few operators with the capital and execution to serve it, a shift that could matter for SpaceX and Tesla.

Tech

Weekly Check: The AI Bull Market Isn't Over, It's Just Gotten Harder

A 5% 10-year yield, 470 gigawatts stacked up in Texas, Chinese DRAM, and whether data centers actually get switched on are starting to decide who wins the AI trade through 2027.

Tech

Is Memory Peaking? In the AI Era, Watch the Bottleneck, Not the Demand

Demand still outruns supply. But memory stocks can peak before demand does. The real question is how long HBM4, server DRAM, CXL and Big Tech capex keep the cash flowing.

Tech

Marvell's $115 Billion Lumentum Deal: Is AI's Next Bottleneck Light, Not Chips?

Marvell's purchase of Lumentum points to a contest beyond GPU speed, in data transfer and power efficiency. Whether the price is right is not yet confirmed.

Tech

NVDA

101 storiesSee all →

Who Hits $1 Trillion Next: Reading AMD and ASML by Different Rules

One company is chasing Nvidia's GPU empire. The other sells the machine that draws the chips for every advanced chipmaker, including Nvidia. Both get floated as 2028 trillion-dollar candidates, but the numbers investors need to watch are entirely different.

Investing

Box Shows AI Its Manuals, and the AI Starts Redesigning the Workflow

Box Architect, built with NVIDIA, moves AI from reading documents to designing how work gets done. It has not yet reached beta, so the real-world payoff remains to be seen.

Tech

Tech Selloff: Is the Bull Market Over? Wall Street's Four-Year Chart Says Not Yet

The uptrend has survived five double-digit pullbacks in four years. This time, AI companies' debt and cash flow are the new variables.

Investing

Nvidia and Palantir Put AI Inside the Supply Chain - The Real Shift Is That It's Learning Judgment

A single Vera Rubin rack holds 1.3 million parts. The real breakthrough in Nvidia and Palantir's new supply chain AI isn't automating that complexity, it's capturing the judgment calls that used to live only in planners' heads.

Tech

If You Only Watch Tesla, You May Be Missing Half the AI Story

Nvidia builds the AI brain. Who supplies the electricity to run it? Six power companies Tesla investors should know, and why Tesla's storage business is just one link in the chain.

Investing

Nvidia Moves From GPUs to CPUs as AI Agents Reshape a $200 Billion Chip Market

A customer's early test found Nvidia's Vera CPU handled about twice the AI agent workload of its existing chips. The bigger story is where the money in semiconductors is heading as AI shifts from answering questions to doing work.

Tech

China's AI Labs Went Open Source Because They Had to, Not Because They Wanted To

A GPU shortfall pushed Chinese AI labs toward efficiency and open weights, and that constraint is now reshaping the global competition over who controls AI's basic infrastructure.

Tech