How Amazon, Google, and Microsoft are dominating the future of AI

AI For Business

But cloud computing and AI Hyperscale“Data centers. Approximately 65% ​​of the world's data center capacity is Owned by just three companies: Amazon, Google, Microsoft. Like the railroad barons of old, they are competing to dominate the market because they understand something we don't: data centers are not just giant digital warehouses. Critical Infrastructure Technologies Almost every business in the world should operate on this basis.

Today, businesses need a full range of computing services, including network, security, data processing, and platforms. just Rent The more data centers from companies like Amazon Web Services, Google Cloud, and Microsoft Azure, the more services they can offer and the more storage and compute they can provide. They can only provide so much capacity, and by aiming to dominate the data center market, they're not just trying to become a big warehouse for data, but a one-stop shop for all the technology companies need.

That's even more true for AI startups. If innovative newcomers need access to the large language models needed to train and run generative AI, they have to go through the big tech companies to get it. And now the tech giants are making venture capital investments in these startups by offering “credits” to use their clouds. That's why Microsoft is The majority of investments in OpenAIFor example, they could give startups access to their own data centers, creating an attractive incentive for them to join their own ecosystem.

“This is real business,” says economist and author New reports It's what he calls “the dynamics of corporate governance beyond ownership in AI.” “Increased AI consumption leads to increased cloud consumption, which not only increases revenue for these companies, but also more digital technologies becoming intertwined within their infrastructure.”

And it is this entanglement that troubles many economists and legal scholars. Regulators call the problem “lock-in.” Moving from one data ecosystem to another isn’t like moving your offices to a new building. For example, the programming interfaces between Microsoft Azure are not simply portable to Amazon Web Services. It’s easy to get into an ecosystem, but like the Hotel California, it’s hard to get out of it. When a tech giant gives a startup access to its cloud services and large language models, it almost guarantees control over that startup, which may one day grow into a competitor. “Market leaders benefit from first-mover advantages that, combined with network effects and high switching costs, lock in customers,” a congressional subcommittee warned. 450-page report Flash forward to 2020 and the data center construction boom, driven in no small part by Big Tech companies' efforts to secure the keys to the coming AI kingdom.

In the short term, the rise of the data center was actually a good thing for startups. “Until recently, the consensus among academics was that the rise of cloud computing was a great thing for startups and innovation,” says Matthew Wansley, a law professor at Yeshiva University who studies competition and regulation. “It used to be that if you were a startup, you had to build your own servers, which was a huge fixed upfront cost.”

That's no longer true: Prices for cloud computing services have fallen every year since Amazon launched its cloud in 2006, then completely crashed in 2014. The team of economistsAWS database prices fell 11% between 2010 and 2014, when Microsoft and Google began advertising more competitive pricing, before plummeting 22% in the two years since.

Cloud computing has also made it easier for startups to raise capital. Venture capitalists The “shoot and pray” investment approachThat meant investing in more companies but putting less money into each one, and taking less direct control over company management, trusting the market to pick winners and losers.

This situation is especially favorable for AI startups. “Small companies like us have access to the computing power and scalability that the big service providers offer,” says Jonas Jacobi, CEO and co-founder of fintech company ValidMind. “There are a few big players dominating the AI ​​space, but there are startups trying to compete with them, and they can only do so because of the cloud vendors.”

The trick, Jacobi says, is to write code that works with any of the three providers, so you're not locked into one company. “You have to be tech stack neutral,” he says. Sure, one of the tech giants could always swoop in and make its own version of your software. It's not like Amazon is “swallow“They buy products from smaller open source competitors and repackage them as part of their own suite of services. Resilient search engine“But that's part of the journey as a startup,” Jacobi said. “It's up to us as a company to move faster and be more agile.”

But over time, economists warn, agility alone will not be enough. In the battle to create the foundational technologies,Key complementary assets“Business” AI startups will inevitably lose out to the tech giants who control the data centers. “AI is a general-purpose technology,” says Licap. “It's being applied to everything. But what type of AI you get, what type of AI you have, will be dictated by the power of just three companies. It's an intellectual monopoly. What they control is the data and the knowledge.” By locking startups into their systems, Google, Amazon, and Microsoft effectively get preferential treatment, giving companies better deals and cheaper services. They have the greatest stake.

Economists warn that over time, AI startups will inevitably lose out to the tech giants that control data centers.

Licapu also found that their increasing control over data centers creates incentives for big tech to share information and cooperate to protect common interests. In a paper with economist Bengt Åke Lundvall of Aalborg University in Denmark, Licapu found that technical and academic journal articles by researchers from Microsoft, Google, and Amazon consistently show that Co-authors employed by competitorsYes, the world of computer science is small, but co-authorship is “a genuine way of showing that they're collaborating and that they know what each other is doing,” and is a prime example of anti-competitive behavior, Licap said.

For now, it's still possible that innovation will triumph over monopolies. Amazon, Google, and Microsoft still compete on price and features, which is good for everyone. And In Europe, regulators are taking a more aggressive approach to technology in general. Especially cloud computingThe Big Three are blaming each other. A Google Cloud executive recently said: He accused Microsoft of being a “monopoly.” and the “walled gardens” and industry groups that include Amazon. filed a lawsuit alleging violation of antitrust laws The two companies are vying for Microsoft cloud-computing licenses but are not yet aligned over market share, creating an opening, however slight, for more agile and nimble competitors.

Also, mature technology companies tend over time to shift from trying to innovate themselves to charging others to innovate. Among economists, this is known asPursuit of interests” This is very similar to what Amazon, Google and Microsoft are doing with cloud computing and data centers.

So what's the best way to ensure big tech companies don't use their data centers to stifle innovation? The researchers point to Google, which offers friendly partnerships to startups: “Its Google Cloud division partners with promising database startups, contributes to open source projects, and works with open source foundations.” Two Scholars Recently ObservedThis is an “architecture of participation”, they say, that allows Google to make profits while encouraging the growth of new companies and ideas.

More importantly, the FTC has recognized the threat to data centers and is forcing big tech companies to Handing over information about AI investmentsJust as new laws caught up with railroad pricing practices in the 1880s, regulators today may catch up with the futuristic technological entanglements of cloud computing. One reason to think so is that the lead author of a 450-page House subcommittee report on Big Tech's anticompetitive behavior was a lawyer. Lina KhanNow she's the hothead of the FTC.

Adam Rogers I'm a senior correspondent for Business Insider.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *