Will AI infrastructure really cost $7 trillion?
Recent reports have suggested that the cost of building and maintaining the necessary infrastructure for artificial intelligence (AI) technologies could reach $7 trillion in the next decade. This figure may seem alarming, but it is important to consider the potential benefits and advancements that AI can bring to various industries.
AI has the potential to revolutionize industries such as healthcare, finance, transportation, and more. With the power to analyze vast amounts of data and make intelligent decisions, AI could lead to significant improvements in efficiency, productivity, and customer experience.
However, building and maintaining the infrastructure needed to support AI technologies is no small feat. It requires significant investment in hardware, software, training, and cybersecurity measures to ensure that AI systems are reliable, secure, and ethically sound.
While the $7 trillion figure may seem daunting, it is important to consider the long-term benefits of investing in AI infrastructure. The potential for increased revenue, cost savings, and innovation could far outweigh the initial costs of implementation.
Ultimately, the decision to invest in AI infrastructure will depend on the specific needs and goals of each organization. It is critical for businesses to carefully evaluate the potential benefits and risks of AI technologies before making any investment decisions.
In conclusion, while the cost of AI infrastructure may be high, the potential rewards of implementing AI technologies could be even greater. It is important for businesses to weigh the costs and benefits of AI investments carefully to determine the best course of action for their organization.
Why raise billions when you can raise… trillions? *evil laugh *
Patreon: https://www.patreon.com/DrWaku
Discord: https://discord.gg/AgafFBQdsc
The semiconductor industry needs one thing more than anything else and that is healthy competition. Nvidia has a monopoly on AI and on discreet gaming GPUs. TSMC has a monopoly on chip production in general. These monopolies are really, really bad for the consumer and for the free market. Take Nvidia for example. With no competition in both AI and in discreet GPUs, the customer is getting hosed. 100k for a chip that costs 3k to make. Gaming GPUs that are so expensive that it is actually damaging the entire platform. Nvidia is and always has been a bad actor in the industry, with a history of being anti competitive and outright deceptive. That is why i hate the CHIPS act so much. Im for the idea of building more fabs but for gods sake dont give the money to the monopolies. Give it to people who are smart and motivated to start their own fabs and bring in some competition. Bolstering monopolies is not the way of a free market society.
Great video. However, monopolies are a result of statism more than stopped by the same. In anything close to a true free market there wouldn't be huge problematic corporations. True capitalists also want to end IP laws. Which of course will help stop monopolies forming in the first place.
I heavily disagree with your explanation at 7:23 on how Taiwan became the market leader in semiconductor. Your explanation seems to be that U.S companies exported all its IP and to a much poorer country (i.e Taiwan) to cut down on cost. This is vast over simplification to the point of borderline insulting. Semiconductor manufacturing is not a low value add like electronic assembly, it is highly capital intensive and requires deep technology know how. It isn't something you just outsource for cheap gains in profit. Many developing countries have tried to create a semiconductor industry and many have failed, see india, and singapore.
IP transfer was certainly a key factor in how Taiwan was able to get a lead in the semiconductor industry, but this wasn't a result of cost cutting measures by American companies. Rather this was a direct result of industry policy of the Taiwanese government itself in the early 1970s. There is a great video about this by
Asianometry see: https://www.youtube.com/watch?v=9fVrWDdll0g.
What do you mean they're not innovating the ring it does everything the fitbit active does … I see what you mean.
Thanks for your informative videos. Your speaking style is excellent. "Junk words" are endemic to some communities — mostly progressive youtubers and far too many of the tech people, including Altman himself. Thanks for your work and for sounding like an intelligent human being and not a — like, ya know, I mean, sort, of, right monster. 👃👍😄 (I mostly watch youtube outside of my account to avoid ads, but wanted to thank you for what you're doing.)
If that's how much it costs, better tax some millionaires, billionaires, and the soon to be trillionair.
Thank you for the info, albino homeless man
That's a lot to pay for a rhetoric parrot.
A parallel AI infrastructure? The computers need new chips?
It's insane that all our (the world) advanced chip manufacturing is in one place. What a setup for world wars. There should be at least 4 or 6 major producers distributed around the world.
Sam altman is he related to Sam bankman?
the entire GDP of the United States in 2020 was about 21 trillion dollars
I don’t trust any corporations or individuals to have too much control. Some countries shouldn’t be a major player. If outside investors want a financial stake, betting on dividends, but not with influence/control over the algorithm & security relevant technology. AI is built on public sector R&D, universities, and private individuals. A balance of human benefit with national security should be found. Cooperate with legit allies, but strict caution with problematic players.
I’ve read “The Interdependency” series – fabulous books!
The ignorant hype around A.i is far more concerning than A.i itself.
This was really good and will probably need to listen to it multiple times. You’ve got a subscriber. Nice job.
Sounds like there's a dead man walking
Actually, suppose you need a MRI machine to create a scaffold file and then ask the candidate tens of thousands of questions while he's sitting in a comfortable couch and wearing a non-intrusive BCI to pre-train the transformer layer and prepare the stencil of the attention matrix, then a compiler creates the NN image file from all this goodness automatically.. Still, you won't be able to transfer more than 1B people this way for this kind of money. It's just enough for a small real world simulation. As more and more billionaires on deathbed make the transition, and the price comes down, Altman will need a bigger server.. 7T$ is not enough. But it's a good starting point for the uplift of humanity🙂.
We are on pre-era of AI-bubble burst. Does it look much the same with the telecom bubble, or I'm getting old?