Microsoft has made 57 investments in AI companies since 2019, more than Nvidia, Meta, Amazon or Google, according to new research for Tortoise’s Global AI Index. It’s capex spending this year would rank it among the top five nation states for public AI spending.
So what? Microsoft has the money and influence of a nation state, and AI partnerships represent a key part of its foreign policy. Under CEO Satya Nadella the company has taken a multipolar approach to AI, spreading capital and risk beyond Silicon Valley and OpenAI in a bid to dominate AI talent and infrastructure globally.
Its interests align closely with those of the Biden administration, as demonstrated by investments in
Magnificent (G)7. Microsoft isn’t alone in the scale of its AI ambition. All of the magnificent seven are signing AI deals at speed – often with companies that have yet to demonstrate meaningful return on investment.
Speaking at the launch of Tortoise’s index, Jonas Andrulis, CEO of Aleph Alpha (a German startup that is an outlier in its refusal to take US big tech funding), suggested the tech states aren’t really interested in cash flow. The game is bigger than that.
“The impact that strategic influence can have on the value of a company like Nvidia is so much bigger than the ticket they have to write,” he said. “This is leading to a situation where there’s so many interests beyond just normal entrepreneurial interest.”
Microsoft epitomises this macro approach. It’s been shrewdly spreading its AI bets, while seeking to hedge against:
Boardroom battles. The saga at OpenAI last year was a learning experience. Nadella moved quickly to fix the disruption and reinstate founder Sam Altman, but high profile resignations have rolled on ever since and there’s renewed debate about scrapping OpenAI’s non-profit structure. Nadella, a cricket fanatic, has looked to cover gaps in his field by diversifying Microsoft’s AI holdings and building its own models.
Regulatory reprisals. One of Nadella’s more controversial plays has been his “aqui-hire” of Inflection AI, which positioned itself as a pioneer of “small language models” for mobile apps. Microsoft hired its cofounder, Mustafa Suleyman, then most of his staff, paid the company $650 million and acquired the right to use its models. The deal passed the scrutiny of the UK’s CMA, but US regulators’ investigation is ongoing.
Running out of power. Sources close to Microsoft say there’s rising concern about sourcing enough energy to power its AI ambitions. It’s a concern shared by governments: last week, US officials met tech executives to discuss data centre permitting and energy supply. It turns out AI sovereignty can’t be left solely to the market.
Last year, Politico reported on “the revolving door” between Microsoft and the Biden administration. It found at least 13 people serving as senior or mid-level Biden administration officials – including Antony Blinken and Jake Sullivan – who used to work or consult for Microsoft. Brad Smith, the company’s president, is known in Washington as a “long-standing regulatory whisperer”.
Public-private partnerships are inevitably part of the AI revolution. But as the big players entrench their positions it’s worth bearing in mind they have the power of countries.
Read more: The Global AI Index 2024