Beyond Tomorrow: OpenAI Founder Raises $1 Billion, Musk's Supercomputer Cluster And More — Weekly AI Wrap

Welcome back to the weekly edition of Beyond Tomorrow. What a busy week it has been!

Nvidia's stock continues to struggle as artificial intelligence sceptics continue with their grim determination to refuse to see the potential value of the technology.

Here at home, Hyderabad just wrapped the two-day Global Artificial Intelligence summit. The theme at the event was Making AI work for Everyone. Telangana Chief Minister A Revanth Reddy is pushing hard to make the state a hub for AI.

There is a lot of conversation happening around the world about how AI will potentially democratise work. There is also the flipside to it of course: When (not if) AI gets better, will some people still have jobs?

OpenAI co-founder Illya Sutskever's new venture, Safe Superintelligence or SSI, raised $1 billion from venture capital funds NFDG, Andreessen Horowitz, Sequoia, DST Global, and SV Angel, according to the company’s website. Reuters reported, citing sources, that the money comes with an eyewatering $5 billion valuation.

SSI plans to use the money raised to acquire computing power and hire talent, since their team is only 10 people at this time.

Instead of doing what everyone in AI is doing at the moment, SSI wants to make a "straight shot to safe superintelligence and, in particular, to spend a couple of years doing R&D on our product before bringing it to market", co-founder Daniel Gross told Reuters. Gross previously led Apple’s AI initiatives.

SSI has positioned themselves as a for-profit company. Their whole thing is based on AI safety. An unusual choice at a time when market rhetoric is pushing companies to be on the cutting edge or risk shutting shop.

What's more, enthusiasm about AI is waning and investors are looking more and more concerned about its potential. That is why in recent weeks, a bunch of smaller AI startup founders threw up their hands and sell to Big Tech.

AI company, xAI, has come a long way since it popped out from the ramblings of the world’s richest man, Elon Musk. The company has completed a 100,000 cluster of H100 Nividia GPUs in four months. The new cluster, dubbed Colossus by Musk, will be used to train xAI's future offerings.

According to two data centre developers who spoke to The Information, a cluster of this size would take at least a year to complete.

Back in June, Musk had said Colossus was already running. But a utilities company back then publicly said that xAI only had few megawatts of power, enough to power a few thousand chips. They later clarified that xAI would eventually have enough power by August to power 50,000 chips, according to The Information. That’s still 50,000 short.

The utilities company said an electric substation was going to come along eventually, capable of producing 150 gigawatts of power, more than enough for 100,000 chips. But that’s not happening until 2025.

XAI is going after OpenAI's current dominance, a move that Microsoft won't be too happy about. In fact, OpenAI founder Sam Altman has voiced concerns to Microsoft executives that xAI could surpass their computing power, per The Information’s report earlier this week.

That certainly doesn’t bode well for OpenAI or Microsoft. But given the latter’s deep pockets, compared to Musk, they could very well fork over the money needed. But will shareholders be happy to see that happen, especially when we seem to be entering a period of AI scepticism? Depends on how well OpenAI’s project Strawberry does, if it is indeed coming out in a couple of months.

· Meta, Nazara Technologies To Help Telangana Boost Digital Development And AI Innovation

· Yotta, Telangana Government To Set Up High-Performance GPU-Based AI Supercomputer In Hyderabad

· Delhi High Court Blasts Wikipedia For Disobeying Previous Directions

· India, Singapore Sign Pacts For Semiconductors, Healthcare And Talent

· Gen AI, Predictive Analytics Among Technologies That Will Impact Digital Governance: Gartner