Top Voice AI Built on Free Credits, McKinsey 2025 AI Report, Google Putting Data Centers in Space

Captions as YouTube publishes them, usually machine-generated, so product and people names can be misheard.

1. 0:00

   Good morning Legion. Today is the day the next mission set is live, and the first Legionnaires are receiving their Ascendant SBT. Congratulations to all of those who successfully made it this far and made the odds be ever in your favor with these next missions. It only gets more difficult from here.

2. 0:17

   Our second piece of good news is that Zashi has officially integrated shielded transactions with NEAR Intents. That means that shielded ZEC can now be used on Intents for swaps, but also for purchases. And eventually with any DApp that has also integrated NEAR Intents in the future.

3. 0:32

   There are three AI stories I want to cover today, all of them relevant to how NEAR’s positioning for the long term. First up, following our narrative of low cost, high quality models, we have another case study of this happening from India Two 23 year old Indians.

4. 0:47

   Drop the number two open weight AI voice model in the world, trained purely on free credits. Maya one is the number 20 globally ranked model, better than even Google's best? It was trained with 3 billion parameters, runs on one GPU.

5. 1:00

   and does 20 plus emotions with less than 100 millisecond latency. Now imagine creating a product for anyone to do this on NEAR, whereby the top tier model of its kind could then be monetized, tokenized, or further financialized as it finds product market fit.

6. 1:16

   We're only at the beginning, friends, and the future is bright. Next up, McKinsey dropped their 2025 state of AI report. It has a lot of good nuggets in it for better understanding exactly where the puck is moving, especially when it comes to topics like business, adoption of AI, layoffs, AI training, and much more.

7. 1:33

   Here's what we find as summarized by Greg Isenberg. 90% of companies claim to use AI, but 67% are still stuck in pilot mode. Corporate AI theater is alive and well. 62% of organizations are experimenting with AI agents. 23% are scaling AI agents.

8. 1:52

   Most of these scaled agents are in tech and health care. The impact gap is massive. 64% say AI helps innovation, but only 39% see real EBIT gains. That would mean efficiency gains. The high performers, the top 6% of users think bigger.

9. 2:10

   They rebuild workflows, set growth goals, and invest real budgets, not just building proof of concepts. Risk management is catching up, with 51% reporting to have already seen AI backfire, mostly from inaccuracy. The workforce impact remains foggy for many. 32% expect cuts. 13% expect growth.

10. 2:30

    Everyone else is guessing and has no idea what's going to happen. And the most important one of all AI adoption is mainstream. but true transformation hasn't started yet. We remain in the early days and early days for everyone is especially great for NEAR’s current build out with two products coming to the market in the coming weeks, both B2B and B2C focus that is our final story today from left field.

11. 2:53

    Google is planning to put data centers in space in 2027, known as Project Suncatcher, Google is launching two satellites in early 2027, each carrying four TPUs to prove that the concept works. Now, if the AI bubble pops before then, they'll be proving a concept that won't be needed.

12. 3:08

    But at least we have a year and some change to figure that out. Thanks for tuning in. We'll do this all again real soon.