'The Economics Of AI Are About To Change Completely': Why OpenAI's o3 Could Reshape The Industry — Costing Up To $3000 Per Query

Comments
Loading...

Chinese startup DeepSeek has upended the artificial intelligence industry’s assumptions about costs and computing power, just as OpenAI announced its latest model can cost up to $3,000 per query.

DeepSeek’s R1 model, which rivals top U.S. systems in performance, reportedly cost just $5.6 million to develop. The breakthrough sent shocks through global markets Monday, erasing nearly $1 trillion in U.S. and European tech stock value, including a record $589 billion drop in Nvidia’s NVDA market capitalization.

“It’s a paradigm shift,” Ali Ghodsi, CEO of Databricks Inc., said to Bloomberg. “These models that can reason are so much cheaper to produce that you will see it be democratized. You’ll see innovations from unexpected corners of the world.”

Don't Miss:

Meanwhile, OpenAI’s new o3 model, announced just before Christmas, signals a contrary trend—achieving breakthrough performance but at staggering costs. The model requires up to $3,000 to solve a single query that humans complete in seconds, using 172 times more computing power than previous approaches.

“One very important thing to understand about the future: the economics of AI are about to change completely,” François Chollet, an AI researcher, said on X when o3 went public.

The model’s success came from its “test-time compute” approach, allowing it to think harder about each question. While this produced superior results, including a 91.5% score on Chollet’s challenging ARC test, it shattered previous cost assumptions about AI services.

Trending: Built on the trusted network of Fortune 500 companies, this blockchain company partners with Salesforce to uproot lengthy and expensive B2B transactions, and you can invest with just $100.

OpenAI, valued at $157 billion, may charge up to $2,000 monthly for full access to o3, according to The Economist, which cited New Street Research analyst Pierre Ferragu. Even the company’s current “pro” version of o1 loses money on its $200 monthly subscription tier due to higher-than-expected usage.

The cost structure challenges the traditional tech industry model of providing services at near-zero marginal cost. It also strengthens the position of suppliers like Nvidia and cloud providers like Amazon AMZN, Microsoft MSFT, and Alphabet GOOG GOOGL)), which control the essential computing infrastructure.

DeepSeek’s success suggests an alternative path. Founded by Liang Wenfeng in 2023, the company achieved its results through efficiency innovations and reinforcement learning techniques, despite limited access to advanced U.S. chips.

See Also: ‘Scrolling To UBI' — Deloitte's #1 fastest-growing software company allows users to earn money on their phones. You can invest today for just $0.26/share with a $1000 minimum.

“More investment does not necessarily lead to more innovation,” Liang told Chinese media outlet 36kr. “Otherwise, large companies would take over all innovation.”

The development is catching Silicon Valley’s attention. OpenAI CEO Sam Altman noted the shift, calling DeepSeek’s R1 “an impressive model” and “invigorating.” Meta established an internal team to analyze DeepSeek’s methods, according to Bloomberg, which cited people familiar with the matter.

Instead of a winner-take-all market, experts predict an oligopoly where multiple firms compete based on specialized applications. Companies like Anthropic, valued at $60 billion, and xAI at $45 billion, suggest investors see room for multiple players despite the high costs.

Read Next:

Overview Rating:
Good
62.5%
Technicals Analysis
100
0100
Financials Analysis
40
0100
Overview
Market News and Data brought to you by Benzinga APIs

Posted In:
Benzinga simplifies the market for smarter investing

Trade confidently with insights and alerts from analyst ratings, free reports and breaking news that affects the stocks you care about.

Join Now: Free!