alt Hacker NewsThe US is winning the AI race where it matters most: commercialization
Comments
No, they are not. They are winning because West is forbidden to use Chinese models for anything work-related.
The winner in the long term will be the one that will deliver the best performance and low-memory ratio for local models.
Anthropic, OpenAI and Mistral are just companies that are making money right now (still not profitable), but will lost their tractions and values in the long term.
However, I am more appealing to see how OpenCode Go subscriptions will go in the future: cheaper than big techs, more tokens, and they don't train on our data to (try to) improve...
I feel like the much simpler explanation is that the US is winning because it's dumping the most money into it. By a very large margin.
No, the US is _leading_ the AI race, but the race isn't over.
What's the point of leading the race for 90% of it, if they're gonna slip on their own sweat and fall down by the end? In non metaphorical terms, what's the point of spending billions of dollars rushing to get the best AI tech at all costs, when the competition can distil your progress and catch up in 6-12 months while only spending 1% of what you spent.
Even in the aspect the article cares about, commercialization, the US is starting to lose marketshare, I've seen people move from cc/codex plans to use glm/opencode plans due to the recent squeeze the US companies put on plan usage, the US companies are screwed if that sticks, not everyone needs the bleeding edge models, they just want to pay $20/month and have the models be decently capable.
Considering that you can easily swap out one AI for another and there is zero lock-in potential, does it really matter who is winning now?
Strange. I'm switching from Codex and Claude to Pi with Qwen3.6 27B local and Deepseek V4 Flash which is dirt cheap but powerful.
One interesting thing that Anthropic did was putting their stack on the various cloud providers, I wonder if they'll put it on GCP and Azure next since they've put it into AWS first at a level we have not seen a major AI provider do to date. Your company can have their own Claude stack just like an ELK stack on your cloud, if they can do this for both Azure and GCP then OpenAI has to really catch up.
In my eyes I would rather use the AI I can run on my own paid infrastructure, so if there's an outage its isolated, or I could potentially have a different region / DC to fallback on.
I'm still surprised that neither Microsoft nor Amazon have made their own models available on their cloud offerings. I guess Microsoft probably does have Phi on there, but it's not front and center, especially with something like Copilot for Devs (seriously Microsoft rebrand that damn thing to be clear what you mean by Copilot!) where they could use the cheaper compute by using something like Phi.
The current commercialization isn't economically sustainable.
This para caught my eye
>Frontier cyber models may push states and defense firms toward the opposite logic: security by obscurity, with closed software, closed tooling, closed firmware, and closed chips. If a model cannot train on the code and architecture of a target stack, it will usually have less context and less speed. That does not make systems safe, but it does raise the value of proprietary stacks all the way down to hardware.
Is this really true. Are there any experts who can weigh in on this.
Should we interpret this to mean that in the new world Windows is more resistant to attacks than say Linux.
Article title: “The US is winning the AI Race”
Article content: “The US are capitalizing on AI the best”
A lot of assumptions there that no one can actually verify as true right now. If commercialization into rent-seeking SaaS landscapes is the endgame, then yeah, the US is winning the AI race. If individualization, local LLMs, and consumer hardware are the endgame, China is winning the AI race. If it’s something entirely different - if LLMs are the wall and research is what grants the next breakthrough, or if compute and memory requirements take a dive, or whatever; then we have no idea who’s winning the race because that stuff is mostly happening behind closed doors.
It’s certainly too early to call (if you must view this as some sort of adversarial competition). The US is behind on local models, the future for anyone who cares about privacy. There may be step change innovation yet to come that completely shifts the landscape. There’s basically no switching costs to users to change models. They have no lock-in.
I've been using the deepseek api (not for coding though) and have been getting great results and it's so cheap it may as well be free. Another reason I'm using it is because I like the license and I also have some hope of running it in my own hardware in the future.
But the thing is... I could be using any of the llms for my use - I'm using a middleware that lets me change providers only with a configuration change.
So it's going to be tough for USA ai companies to charge 5x to 20x (depending on what you're doing).
Inclined to disagree.
The winner here will be whoever can move atoms with AI not take notes at the daily standup.
i.e. Think boston dynamics vs unitree
They're both doing well but I'd lean towards China is winning on atoms in light of a huge manufacturing base they can AI-ify.
I continue to be impressed by our collective willingness to engage with obvious AI slop, as long as it also talks about AI. Sincere question for any of the nearly 300 folks seem to be arguing about the article: why? The author couldn't be bothered to present their case, so they probably don't care about our opinion. They just want traffic and search ranking with the least amount of effort. The community is literally being played for clicks.
Is it just that the subject line alone is a springboard for casual discussion? If so, maybe that's fine, but then, it feels like we'd be better off cultivating these discussions as "ask HN" posts instead of boosting this kind of web content.
> where it matters most: commercialization.
It begs the question because both its premise and assertion are already wrong. Has AI improved the industrial capacity of the US in order to improve the lives of its citizens? No it hasn't. Has AI increased the wealth of its citizens by being able to do laundry or any household task in a generalized way? No it hasn't. The only thing it's really done is make very narrow slices of white-collar work more fungible. In what way has AI been able to address existing shortcomings of the US?
>Trump fits this moment well. He is a salesman at core, and Larry Ellison is too. That helps explain why AI infrastructure is an easy political product. Selling AI today is easier than selling Oracle databases in the 1980s.
I feel like the author (and perhaps many here on HN) are on a different planet than almost everyone I interact with.
I don't think so. From a nation state perspective, AI is a munition. Every advanced nation is going to have their own cyber division with their own AI hosted within its borders. Considering how xenophobic and belligerent the US is, nobody is going to want their national cyber defenses hosted in the US.
On a personal level, I simply do not trust the US anymore. I won't host any of my personal data in a US company. I don't want the US govt invading my personal privacy, and their corporations are constantly leaking and selling private data. I consider US to be rapidly approaching complete autocracy (on par with China) so US-hosted AI is a non-starter. And let's not forget local inference keeps getting more efficient, with higher context and TPS in the same amount of RAM. Within a year even small consumer machines will run local models good enough for basic coding, and in 3 years RAM prices will lower and everyone will be able to afford a decent rig.
Finally, open weight models are now good enough for daily work. They may never be as good as SOTA (SOTA will just keep increasing indefinitely), but that doesn't matter; my car may not be as fast as a Porsche but it still gets me to the grocery store and back. So I use non-US hosted model providers which provide open weights, which are both significantly cheaper than Anthropic/OpenAI, and actually allow me to use my subscriptions without a moat.
But yes, Anthropic/OpenAI are absolutely the new Oracle. They will win for US govt and Enterprise contracts. But that's far from the only users of AI.
Flagged for AI content: I hope this submission dies and the user is penalized (look at their submission and comment history!), because IMO the article does not belong on the front page. Quick polemic:
>The US is winning the AI race where it matters most: commercialization
If you ask me, one could name different criteria for winning, and commercialization would not be the first thing to come to my mind:
https://english.www.gov.cn/news/202604/15/content_WS69df29e6...
https://fortune.com/2026/05/03/chinese-court-layoffs-workers...
https://www.reuters.com/world/china/china-moves-regulate-dig...
> It also owns platforms that generate and organize the data of the AI age. YouTube is a video corpus. Google Drive and Microsoft 365 sit inside daily office work. GitHub sits inside software development.
Yeah, okay. China does not have any platforms nor data.
Same as software in general I guess? Lots of critical software has been developed by Europeans, but it's the US who build hyperscalers with it
Alibaba's cloud is something that the author of this article seems to dismiss. It's being used massively in Asia and they're pretty close in services and offerings to what AWS, GCP and Azure provides. Once they start doing inference on their own custom chips it might be hard to compete with them due to the energy costs
Even when other countries won, the teams would have reloated to US.
> Many people use the wrong scorecard.
Correct. "Revenue" is the wrong scorecard when they're selling 20$ bills for 15$. I too can make a bajillion dollars in revenue with that strategy.
Show me a company not speed running the uber/doordash playbook and we can talk.
US AI models would not survive a free market re: this metric.
“It is not the same as profitable AI leadership”
Where are these profits of which you speak?
That's like saying that Louis Vuitton is monetizing shoes the best. Sure, but it's not winning shoes
"Where it matters most": accuracy and repeatability?
Sorry, nobody's winning that AI race.
Ppl don't understand Commercialization is not incidental to the Western system, it's why we beat out Communism. Commercialization incentivizes ppl to build, bc ownership and control.
The FSF was not an attack on commercialization, it was about giving users more freedom with their own copy.
AI commercialization is why we will always be a few steps ahead in AI.
The Chinese and Russians are free to join us. It's a pickup game.
Matters the most to whom? I certainly will not care about expensive models that do about the same thing cheap non-american models do.
It's like the USA Librem 5 vs PinePhone. About the same HW for $1600 vs $150.
Sure will not pay 10x for "US" thing just because it's a US thing.
The whole "race" narrative is silly. It is all built on the assumption that one country (corporation, actually) somehow creates AGI and thus, essentially, the singularity. Great for raising VC, apparently, but at its core this is magical thinking.
Even if any of the US corporations would eventually end up in a scenario where their revenue is at least as high as their inference cost, what harm would that do to the other contenders? It's not as if there is any kind of network effect here that would exlude them from market participation.
And, yet, the US AI companies are not actually making a profit, right? They're selling at a loss and trying to make it up in volume (or lock in some kind of monopoly position, in a currently non-sticky product). We're all currently enjoying investor-subsidized tokens from the big guys, and that pushes out the reckoning for US AI. But, I think they're beginning to think maybe they need to ring the cash register. Copilot dramatically reducing usage limits and what models are available on its plans, Anthropic playing games with what's included in the Pro plan, etc. I think they're starting to feel the bleeding.
Not only is the investment that keeps US AI companies flying high slowing, I suspect in two or three years, we'll all mostly be using open models and the people making money will be the hardware manufacturers. Even the small models will keep getting more capable. I'd guess a model you can run on a high end, but not outrageously overbuilt, developer desktop or laptop (something like 128GB of unified RAM), will be competitive with the current frontier when it's allowed to search the web and do research and write test code. You can't fit as much knowledge in a small model (80GB of weights can't store the world's knowledge), but I don't have the world's knowledge in my head, either, and yet I can figure out most problems with a little googling and experimentation. The reasoning and tool use abilities of smaller models is where the gap is closing, and that's what will make the huge models obsolete for huge classes of problem.
Already, there are many classes of problem that the easily self-hostable Qwen 3.6 27B can solve that required a frontier model a year ago. When the self-hosted options reach Opus 4.5-ish levels of capability, the argument for paying for tokens for most work begins to look a lot less compelling. And, looking forward, 1.58 bit models are coming. Incredible intelligence density, and still a lot of improvements happening.
Commercialization is not enough. The US is built on financialization.
Cultivating an ecosysyem of strong capital protections, wealth creation through extraction, and tax advantages for AI finance is what we should be looking for. Commercialzation may be a step towards that, but isn't the destination. We have to create a system where those with money can multiply it, not simple add to it.
I feel like the title of this post should have "for now" appended.
Same can be said of healthcare.
The US is running itself, and humanity, into the ground by massively increasing the amount of electricity it uses, instead of reducing electricity and fossil fuel usage. US residents have already started to feel the crunch in terms of water and power in some areas; and the entire world is experiencing the (admittedly less critical) shortages in RAM and SSDs.
And that's not to mention the warping of US economic life by the concentration of capital around this bizarre endeavor, with the circular multi-hundred-billion-dollar deals and such.
Unfortunately, the detrimental effects of global warming arrive gradually, and are spread out over the entire globe, so the "AI barons"/tech magnates will probably suffer the least, while island countries will be completely wiped out, whole regions will become too hot to sustainabily live in, tens of not hundreds of millions will have to migrate, biological diversity will suffer, etc. They will look back on these times in a 100 years and will think of us, or at least of US, as the people boarding the Titanic. Hopefully not as the people who board the Hindenburg.
This sounds like an ad for US than anything else.
Does any of the US companies earn money on LLMs? No, they bleed money. Github Copilot is switching to token based pricing, which will be costlier than hiring juniors.
Anthropic also is switching enterprises to token based pricing from their subscription one.
From the big three only Codex is still in somekind of subscription pricing, but they'll shift eventually (usage limits are a kind of that, but they have them less stricter than Claude ones)
There is one winner in this race - China. Trump with his agendas and wars makes it even more likely that China will lead this new market.
Well we're certainly losing in terms of public sentiment. There's a real anti-technology mindset that has taken hold. Basically the equivalent of people protesting the electrification of the country early last century.
Chinese culture is quick to embrace the benefits.
Hahaha, but no.
It's like people forget the entire point, perhaps even definition of technology is "doing more with less."
The "brute force" of power and cycles is almost certainly the least important thing, perhaps even a hinderance.
I'm glad we went to space, truly. Racing the USSR might have been the wrong reason but it got us there. We've benefited immensely as a species from exploring the solar system and looking deep into the universe.
I'm not certain that racing China in AI is the right reason but it might get us... somewhere.
… and it’s substantially due to foreign born researchers and engineers. US will win as long as smart and driven people will want to move over.
GLM, Kimi, MiMo, Minimax, Deepseek, Qwen would like to have a word. :)
"Yet like Musk the ouster wounded his ego". So the journalist believes that reacting to rejection with emotions like a biological person makes him like Musk. Err okay.
> Trump fits this moment well. He is a salesman at core, and Larry Ellison is too.
Larry just fired 30% of his people at Oracle because, apparently, he is in an immediate need for cash. Because Oracle's early AI bets aren't paying off.
No, it’s not.
It’s all about adoption and the bigger picture. The US is an untrustworthy, isolated island in the AI future if you vote another idiot into office in a few years. If you’ll still be able to vote at all, that is.
The largest part of the world is not the US. The cutting-edge US models are way too expensive for most parts of the world, and that also shows in adoption.
China is building an ecosystem of open-source models that are both cheap and good enough for most use cases. While most of the US AI sphere will collapse under the pressure of making profits, which means having their models and infrastructure adopted by as many enterprises and individuals across the world, China’s models will have become global standards and hard to displace.
If Beijing’s AI pitch centers on universal access and cost-effectiveness, then Chinese AI firms do not need the latest chips to win the global AI race. They also don’t need the expensive US-run infrastructure. If you watch Chinese AI adoption closely, they already want as many Chinese people as possible to be able to build and try with AI, whereas for most Americans, US models for productive use are already too expensive.
Kimi K2.6 sits within touching distance of Opus 4.7 and GPT-5.5 while costing about $4 per mil output tokens. That is six to eight times cheaper than cutting-edge US models. If you run hundreds of agents, that’s a significant opportunity to get the same work done for a lot less.
Even early adopters like Singapore, ditching US models, the government kicked Zuckerberg in the nuts and went for Qwen instead to build its sovereign AI models.
To understand why the US is at a severe disadvantage in this race, you need to understand China’s Belt and Road Initiative (BRI). BRI entails Chinese firms delivering fully financed infrastructure projects in a bid to lock third countries into China’s economic orbit. They use the same approach for their open source ai models, but this time the infrastructure is both invisible and free.
No need to build power plants or buy /build ports. AI dependency is invisible to both policymakers and the population, limiting pushback. No pesky activists in Germany nagging about China buying parts of ports. No African nutbags questioning why the humble Xi is building hospitals in areas Chinese mining companies take things out of the ground for pennies on the dollar.
China is going for a marathon here while the US tries to push their ai tech by sheer force into the throats of the world. As soon as Chinese ai models have become global standards, it’s game over for us ai companies. And China is way better at this game than the US. They have proven this over and over again in the past 50 years.
I recommend reading the China Standards 2035 strategy to get a better understanding of their approach and how smart this is.
https://www.china-briefing.com/news/china-standards-2035-str...
AI is not as revolutionary as you think in terms of our experiences with previous technological advances in terms of trade and economics.
Western economies are locked into U.S. models, while China runs on Chinese ones. It’s the age-old game. But the real war of the AI race will be fought in the global south.
I will give you three examples.
Can you really imagine, if you look at what AI needs to cost to make a profit, that even at the current prices, US models and infrastructure, which are heavily subsidised already, being used in cost-sensitive countries? I am not talking about coders, think really big here for a second.
Secondly, US ai models are trained on Western data. How do you expect them to grasp local contexts in the Southern Hemisphere? Chinese open-source models, on the other hand, can be downloaded and finetuned with country-specific data.
Want an example? Check out AfriqueQwen-14B, which is adapted to the top twenty African languages.
So I think this author is wrong. The ai race to be won is not hardware or cloud infrastructure, my money is on it will be a contest to decide which models and standards become the default infrastructure in countries that are up for grabs.
China neither needs the best models nor does it need the best cloud infrastructure, it just, like so often, only needs to be affordable and good enough to become the default choice in emerging markets.
The right choice would be for everyone to step off the gas pedal and think about whether we are willing to become China in order to beat China. Our ancestors worked really hard to get us here, our rights, our ways of life, culture, all the blood, sweat, and tears.
AI better be worth it in the long run for all of humanity if we go back to survival of the fittest. Because that is what it will take to beat China at their game.
Back in the dot-com bubble, people started inventing new metrics to "value" dot-com companies that lost money hand over fist. My favorite was "revenue multiples". So instead of a a P/E ratio, it was just a multiple of revenue no matter how much money you lost.
We've invented a new term here too: revenue backlog. OpenAI and Anthropic in particular need to recover probably at least $2 trillion to recoup their capex investments. Now Claude code has had an impact on software engineering but for a lot of AI uses you're just not going to recover $2T on $20/month subscriptions. It reminds me of Twitter trying to dig itself out of a $44B hole and losing half their ad revenue with $8/month blue ticks.
The only commercial product AI sells is labor displacement and the resulting wage suppression. You lay off 10-20% of your staff and nobody is asking for raises. The people left are happyt o still have jobs (and thus a house). They'll work even harder doing unpaid labor of the displaced workers to keep those jobs. That's what OpenAI and Anthropic are selling.
The problem is that if these companies get their way, 10-20% of the population is going to be out-of-work and society is going to fall apart. Data centers are going to be the targets of increased societal desperation and anger as this gets worse.
There was a report this week that roughly 50% of singles in the US aren't dating because they can't afford to [1]. This goes well beyond the well-understood problems of not being able to afford a house let alone start a family. This is a birth rate death spiral in the making.
So, back to OpenAI and Anthropic, the only way they justify their valuations and can make up the "revenue backlog" is if they have a moat. And I don't think that's going to happen. Hardware will get cheaper. Nobody is talking about how the generation of AI hardware will write of trillions in investments for some reason. I don't know why.
But the dark horse here is China. DeepSeek when it was first released (early last year?) was a shot across the bow. We have it and toher models (eg Qwen) that will close the gap with whatever OpenAI and Anthropic produce such that no company will "own" AI in the way that OpenAI and Anthropic need to. In the coming years, China's chipmaking is rapidly closing the EUV gap and Western companies have zero penetration into this market. China doesn't want to be dependent on foreign tech that can be withheld at any moment.
Don't believe me? Just listen to the NVidia CEO say the exact same thing [2][3]. Huang realizes this is such a problem that he's gone on Air Force One to this week's Trump summit in China to try and convince the Chinese to buy NVidia chips.
[1]: https://parade.com/living/nearly-50-of-single-americans-not-...
> The US is winning the AI race where it matters most: commercialization.
puke
Yeah, go ahead and run your country into the ground because of hypercapitalism and hypercommercialization, you're almost at the end game now! While the rest of us try to figure out how to actually build societies worthwhile to live in and experience, with healthcare and not waging war on our neighbors.
I don't know how people can seriously publish stuff like this and not feel like they're actively trying to make the world worse. Is money really the single thing y'all can focus on? Is there nothing better in life you can chase, even if it's also a number? So sad to see stuff like this.
How about the obesity and fall of democracy race?
By this article, "the US" is not winning. Sam, Dario, Elon, Zuck, etc. are.
It remains what benefit, if any, Americans will see from all this...
This is patently absurd. US AI companies are investing non-existent money on huge infrastructure with negligible income. This cannot be sustained. And if/when it fails it will take down the economy of the US and probably any other country touching us business.
Anthropic, OpenAI and Google are the standouts, but the main question for me is, why is this a war? In their own context China has greatly benefitted from this. They shored up their gpu design and manufacturing expertise.
If this really is a war, trump is kneecapping the country with his lawlessness and eroding America’s good will. If the world cannot trust China with their data and they cannot trust the U.S. to provide good reliable service and not turn it into a mafia style negotiation, then winning the AI war is not helping the U.S. countries as much as it potentially can. It’s probably a good thing for more capable areas like Europe which may develop their own tech stack.
In a weird way because the AI stack is so expensive, China helps the world much more than the U.S. with their really capable open source model.