Free NewsletterPro Login

Free Live Investors Workshop

Seats limited

Tue, Sep 29.

The dollar is losing value.

Here’s how investors can still profit.

Hosted By

Jaspreet Singh

Founder, Briefs Finance

X
Free Live Investor Workshop

AI token costs hit a new low as cheaper rivals crowd the field

Published Sep 1, 2026
Share:
Summary:
  • Silicon Data's LLM Token Expenditure Index slipped to 97 cents on Monday, its weakest reading since it launched late last year.
  • The index is now less than half its summer peak, which can lower the tab to query ChatGPT, Claude and Gemini while eroding model makers' pricing power.
  • Several forces are pushing prices lower: open-source Chinese contenders such as Moonshot's Kimi K3, OpenAI's late July reductions on two GPT-5.6 models, and other labs' dynamic pricing, all squeezing margins for IPO candidates Anthropic and OpenAI.

What fell and why it matters

A closely watched yardstick for what a large language model token fetches in the market sank to 97 cents on Monday, a record low since the gauge debuted late last year and more than 50% below its early-summer high. Silicon Data's gauge measures what buyers are paying per token. Cheaper tokens can mean smaller bills to run prompts on OpenAI's ChatGPT, Anthropic's Claude and Google's Gemini. They can also reset expectations, nudging customers toward lower access fees and weakening providers' ability to hold the line on price.

What is driving the drop

Charles-Henry Monchau, chief investment officer at Syz Group, cited open-source Chinese systems such as Moonshot's Kimi K3, which can price below leading frontier offerings. OpenAI revealed in late July that it had lowered prices on two of its GPT-5.6 models, and Monchau noted that peers have rolled out "dynamic pricing" so access fees fluctuate with demand. He added that falling industrywide costs to produce a token are helping push prices down.

"Foundation model labs are the most directly exposed," Monchau wrote. "Token deflation compresses the revenue line while compute commitments stay fixed. The strategic response is visible: the moat must shift away from raw model capability - where the open-weight gap is now measured in months - toward distribution, memory and context."

When costs and tools shift, steady investing still matters, so get the free Always Be Buying E-Book today

Silicon Data's head of research, Steve Hou, said the recent slide could be a sign that between frontier systems and cheaper rivals there is already enough supply to "provide sufficient capabilities for most tasks."

The market angle

If token prices continue to decline, margins could get tighter for companies such as Anthropic and OpenAI while they consider the timing and viability of going public. Both confidentially filed for IPOs this summer. At the same time, investors might have to reassess expected returns from the AI buildout, even as mega caps like Nvidia and Microsoft have invested billions to expand the infrastructure that powers AI.

Technology stocks led the market lower on Tuesday. The Nasdaq Composite slipped nearly 1%, while the S&P 500 edged down 0.4%. For your wallet, cheaper tokens change the math: using AI can cost less, but the companies selling access may find it harder to grow revenue, which shifts the story you are buying when you own them.

Disclosure

Recent News

1 2 3 78

Get Market Briefs delivered to your inbox every morning for free!

No fluff. No noise. No politics. Just finance news you can read in 5 minutes.
0 Shares
Share via
Copy link