Will a win-win for Nvidia and OpenAI end up as a lose-lose for the rest of us?
AI chip major Nvidia’s pact with ChatGPT’s maker may look like a business masterstroke, but an emerging web of big-money alliances should make the world’s AI users sit up. Nobody should get to dominate the market for artificial intelligence. It’s way too important.
Dystopian visions of artificial intelligence (AI) feature bots breaking free of human control and running amok. But what if the AI industry itself goes wild? While US investors must watch out for an outsized investment bubble, AI users everywhere need to worry about a web of convergent business interests that could plausibly cramp market rivalry.
On Monday, news broke of Nvidia’s deal with OpenAI that could see the world’s top AI chip company invest up to $100 billion in the firm behind ChatGPT, one tenth at a time.
Nvidia is the world’s most valuable company, with a monopolistic hold on its chip market, while OpenAI leads the global chatbot race with a claimed 700 million weekly active users. On the face of it, this is a clever way for Nvidia to boost demand for its chips as OpenAI invests heavily in data centres. For every gigawatt of capacity, the AI world’s dominant chip-seller might pump in $10 billion.
Days earlier, Nvidia said it would put $5 billion into Intel. As for Sam Altman’s OpenAI, it recently struck a mega deal with Oracle to lease AI infrastructure. For a 2015 startup with revenues placed at only around $13 billion, that’s quite some deal-making. It speaks of how explosively today’s ‘smart money’ expects AI to grow.
An early bet-placer was Microsoft, which invested over $10 billion in OpenAI soon after its chatbot stormed our digital gadgets. Since Altman’s effort to carve its business subsidiary apart from its non-profit parent has not yet taken formal shape, its control structure remains hazy. What is visible, all the same, is a web of alliances that AI users have reason to be wary of.
What’s touted as a win-win for businesses need not work out well for customers. While Nvidia’s CEO Jensen Huang has spoken of how both his company and Altman’s will gain from their pact, critics point out the ‘circular dynamics’ of one firm funding another’s purchases of its wares.
If this gives the deal an artificial air, the fact that OpenAI will partly use its closely held shares as a currency to create backup infra adds to its opacity. OpenAI is reportedly valued at some $500 billion, but reports say that the equity acquired by Nvidia would have no voting rights, assuring Altman executive leeway from his megacorp allies.
The fortunes of OpenAI and its allies hinge on leadership of a blistering new age of AI-driven gains. For all the hype, however, the impact of AI on corporate productivity has been modest so far, like the learning curve of its offerings.
In a dream scenario, AI will put economies on sizzling growth paths and generate vast revenues off a global user base to justify the mega-sums being invested in it. But this is yet to be tested and billions could end up wasted in pursuit of a grand fantasy.
Either way, if the ‘groupthink’ of a closely aligned quest for AI market success gets in the way of true competition, AI users could be dealt a bad deal.
In a market that serves users best, various AI visions would compete fiercely for both users and resources—with their aims and rewards free of each other’s. True, Big Tech players like Google have competitive AI offerings (and even allies), but OpenAI’s web of alliances, some of it blessed by the White House, could emerge as a collusive force to reckon with if mass-market AI becomes a winner-takes-all game.
Indian AI is in infancy and Chinese AI untrustworthy. If smaller AI players get squeezed out by capital intensity and network effects, we risk a global failure of ‘intelligence,’ with its artificial version crafted for the rest of us but controlled by a few.
topics
