“A handful of remarkable things that recently happened in the world of artificial intelligence all point in one direction: AI is becoming a widely available commodity.
First, the price of AI good enough to accomplish most everyday tasks has dropped precipitously. This is due to lightweight models that run in the cloud and on our devices, including new ones from Google, Apple and Chinese AI companies.
On Thursday, Chinese leader Xi Jinping gave a speech in Shanghai advocating for the ongoing release of "open-weights" AI models -- the kind anyone can modify and use freely -- as a counterbalance to U.S. dominance in AI.
And now Chinese models including GLM 5.2 and Kimi K3 are catching up to the best that U.S. tech companies have to offer.
Second, Meta Platforms showed the world it could potentially compete with the two leading AI labs, OpenAI and Anthropic, on their own turf, delivering high-performing models for the lucrative coding market.
And third, the current computing-power bottleneck appears set to ease as more data centers come online and engineers figure out how to deliver AI more efficiently.
For some applications, the supply of tokens -- the basic unit of AI use -- is catching up with demand.
These developments are great for the world. OpenAI Chief Executive Sam Altman hailed intelligence "too cheap to meter" as a goal just a year ago. And rather than taking all the jobs, AI might actually boost productivity of many workers and potentially reduce digital friction in our modern lives.
But is this good news for OpenAI and Anthropic? Poised for IPOs, both depend on maintaining a competitive edge over incumbent tech companies for future profitability. If AI models turn out to be a general-purpose technology like the automobile or electricity, what can they uniquely offer?
As of March, global consumer market share of OpenAI's ChatGPT, measured by unique users across mobile and web, fell below 50%, according to the market-intelligence firm Sensor Tower. That's mostly due to competition from Google Gemini and Anthropic's Claude.
As for enterprise customers, Chinese AI models can now match the leading U.S. models by some measures, at far lower cost. On the OpenRouter leaderboard, which tracks business consumption of AI tokens on its platform, the top five models are all Chinese, and approximately 45% of all tracked tokens now flow through Chinese models.
Thinking Machines Lab, led by Mira Murati, former OpenAI chief technology officer, just released a free-to-use open-weights model it says will balance power and running cost. Translation: Who needs an AI Ferrari to get to work when the AI Honda Civic is right there?
As competitors nip at their heels, OpenAI and Anthropic have been spurred to increase investment in both engineers and data-center access at the expense of profitability. They are committing hundreds of billions of dollars to keep ahead on model progress and pipelines.
Meanwhile, customers of the AI companies are finally getting their heads around this tech, causing them to reassess how much they spend on AI and which of the premium models they actually need. OpenAI and Anthropic are contemplating a price war.
Meanwhile, SpaceXAI, Meta and the Chinese companies proliferating open-source models -- free for anyone to run on their own hardware -- have already kicked off that price war.
Two related trends drive this AI commoditization. The first is the diffusion of knowledge about how to build cutting-edge AIs.
Tech companies still zealously guard their secrets -- Apple just announced a lawsuit against OpenAI over alleged theft of AI-related intellectual property. And yet AI engineers, even those inside companies, continue to publish papers about their discoveries.
Open-source models, from Chinese labs and "neolabs" like Thinking Machines Lab, are regularly released along with extensive notes about how they were built.
Then there's distillation, a fancy term for training one model using another. OpenAI and Anthropic have accused Chinese companies of siphoning proprietary data to create their own AIs by this process. But distillation is also used for legitimate purposes: Within AI companies, a big, expensive model might be used to train a cheaper, faster one. The new AI model powering Apple's refreshed Siri AI was distilled from Google's models under an agreement between the companies.
In a recent essay, Microsoft CEO Satya Nadella defended distillation. Given how AI companies train their models on information they scrape from the internet -- and even their own customers -- he finds it "ironic" that they would try to restrict others from distillation. "If learning flows in only one direction," he writes, "economic value converges toward the owners of the learning infrastructure rather than the creators of the knowledge itself."
Despite AI companies' protests, it appears AI's blueprints are about as hard to contain as those from our previous industrial revolutions.
A truism I've learned from years of prognosticating about the fate of tech companies: Once they have a certain amount of talent and capital, startups in particular can be quite nimble. The fires of fierce competition can temper young companies, and they will pivot to whatever business model and competitive advantage they can.
Up until now, the primary moat for OpenAI and Anthropic was simple: By comparison, their competitors' products weren't very good. Sometimes they were barely usable.
Now that "good enough" -- and sometimes even "pretty amazing" -- AI is available to iPhone users and Fortune 500 chief information officers alike, these companies will have to find new moats.
Google has its massive search business; Meta has its social-media empire; Microsoft and Amazon dominate the hosting of enterprise software; and Apple sells the most popular hardware for consumers to run AI on.
In the future, the moat that will matter most for AI companies will be access to power, says Eric Zhao, a professor at the University of Oxford who is co-author of a recent paper on the subject. As electricity becomes ever more difficult to access and local communities block data-center construction, efficient use of that power will also be essential. "Frontier labs will need to compete on intelligence per watt," he adds.
OpenAI is rapidly diversifying its income streams. The company has said it has over three million business customers. OpenAI is also working on its own hardware, to create a direct relationship with consumers.
Anthropic recently achieved its first profitable quarter and has filed to go public as early as this fall, which would give the company a large cash infusion to spend on acquiring more customers, pursuing new revenue streams or just leasing more data centers.
It's possible that the market for AI is so big that there will be enough future revenue to go around.
It's even possible OpenAI or Anthropic could become the next tech giant in its own right.
But the scale of AI investment is now bigger than any nonwartime boom in history, including railroads and the dot-com bubble, according to the Bank for International Settlements. A growing chorus of critics assert that AI companies participating in that spending will struggle to justify today's investments, much less the even bigger outlays planned for the coming years.
Without a competitive moat, the darlings of today's AI boom could be headed for a serious contraction.” [1]
1. EXCHANGE --- Keywords: AI's Availability Is Good for China, Not Great for OpenAI and Anthropic --- Intelligence 'too cheap to meter' could threaten the leading AI labs, while creating opportunity for their competitors. Mims, Christopher. Wall Street Journal, Eastern edition; New York, N.Y.. 18 July 2026: B4.
