The AI Power Shift: Can China’s DeepSeek Disrupt
OpenAI and Google?

Featured Speakers:-

Links and Resources

• The Human Touch 2.0–Leveraging AI for Deeper, More Meaningful Sales
• The OpenAI Shake-Up: Navigating Success with DiGGrowth’s Hybrid Approach
• Breaking Down Barriers: AI Tools and the Democratization of Data Analysis in B2B Marketing

Transcript

0:11:
Hello and welcome to another episode of the revenue Focused marketer where we discuss anything and everything related to marketing as well as data.

0:20:
My name is Hershey and I’m your usual host, and I am today joined with two very leaders in the industry that are going to kind of talk to us about what’s really happening with AI, AI implementation and really touching on the topic of DCKI.

0:37:
So first up, we have Rahul.

0:39:
Welcome, Rahul.

0:41:
ASC, thank you.

0:43:
So Rahul is the VP of technical operations at Growth Natives, and he’s also implementing AI solutions and offerings.

0:51:
He has joined with Arpit today.

0:53:
Hi, Arpit.

0:55:
Hi, Ishi.

0:57:
Thank you for joining us today.

0:59:
So Arpit is the co-founder and the product head of DIGGrowth.

1:02:
, he’s leading a very talented team of data engineers to find innovative AI solutions.

1:08:
So I’m really excited to kind of have these both experts with us today to have this conversation.

1:15:
I think we.

1:18:
So I think we can kind of dive into the big picture, and I, I’d kind of love to start off with you.

1:26:
So China’s AI capabilities have grown massively over the years and some say deep CKI is just another iteration of Chad GPT 4.

1:35:
What is really your take and you know, what do you think are some of the factors that have enabled China to compete at this level in AI and technology?

1:45:
Yeah, definitely, that’s a great question, and let me just set some context and a bit of history.

1:51:
So if you think about any country who actually making such bold big moves, there’s a lot of things that are going behind the scenes, and that’s what has happened.

2:01:
With China.

2:02:
So in order to really come up with this kind of innovation, there are some essential factors that are required.

2:09:
First and foremost is the government’s support and having that ambition of global impact, and I think that’s there in their recipe.

2:17:
This should be powered with.

2:19:
Kind of futuristic research that happens in universities.

2:22:
Generally speaking, universities and their research and their patents are 2 or 3 years ahead of the overall advancement and learning curves.

2:30:
So you need to have that, and then it should be also backed by certain investor ecosystems along with innovation labs and.

2:39:
And that is also sort of developing in China and then it should also have kind of engineering talent that is required and specifically if I talk about AI and some of the innovations there, there are two more things required.

2:53:
One is that you have that right infrastructure in terms of the compute and do you have the right and if Go back somewhere like 20-30 years, China has generally been tagged as more like a copycat, but if you think about it, we always say that they have got their own Facebook, they’ve got their own Google, let’s say Baidu is their search engine.

3:15:
While they’re copying all those things, they’ve also built that infrastructure and ecosystem.

3:21:
The universities and research right now probably China is stopping in terms of the number of patents that has been filed and while they were building and copying these products, there has been an engineering talent that has now emerged, so they can easily innovate now.

3:39:
And at the same time, while these companies, because there’s a good consumption that happened within China, there are like really big companies like Alibaba that have mushroom, right?

3:48:
So now they have access to a good talent of engineers.

3:52:
They’ve got some really great tools that have got a lot of consumptions internally, which means there’s a lot of data that’s already there that can be used to model.

4:02:
And then finally there is, there’s a difference in terms of the whole mindset.

4:07:
If you try to compare the US mindset, it’s all about abundance.

4:13:
They have been obviously leading the innovation, but their approach has been through money, through resources, and try out things, be innovative and fail fast.

4:24:
China, on the other hand, has developed a lot more frugal approach I would say, and the outcome of this is models like Deep Seek and more, which is more about not just throwing money, but ensuring that there’s an efficiency game as well.

4:39:
So I would say that it is following the footprints of OpenAI and some of the top models from US, but then there’s a good substance into it in terms of innovations and efficiency.

4:52:
I guess I have to ask.

4:54:
OpenAI is being accused, , is accusing deep CKI of intellectual property theft, right?

5:00:
But some might even say, isn’t it ironic given OpenAI itself trained on scripts from like, you know, the internet data.

5:09:
So do you think they’re really being hypocritical in that approach?

5:12:
I would love to just get your thoughts on this.

5:15:
What I think is it won’t matter much now to OpenAI, and I was hearing the latest interview from Sam Altman, the CEO of OpenAI, and he clearly mentioned that there are more better things to focus.

5:27:
They are not going after Deep Sea.

5:30:
So and then clearly.

5:31:
There are other open source models available.

5:34:
Mehta is one of them.

5:36:
They offer Lama as the open source, so it is not anymore just China versus US or it’s a whole altogether like an AI arm race that everyone wants to win.

5:49:
Yeah, I think it’s no longer, you know, protecting its own monopoly for Open AI because you like you said, there are other competitors out there.

5:58:
And I guess I like, you know, let’s talk about these meta you kind of mentioned about the AI model like the Llama.

6:06:
What do you think are the strategic implications of these open source models for the industry?

6:11:
Like, do you think, you know, having these different players, how is it really going to disrupt the industry and is it gonna really, you know, impact how people are really adapting to AI now?

6:24:
Yeah, I think definitely open source is welcome.

6:27:
I mean we’ve seen it with Linux and then being part of a startup where we’re also trying to build and make impact in the industry.

6:34:
Open source is the way to go for us across everything that we offer, leveraging that work and if you talk about from the perspective of me or let’s say, they have their own point of view.

6:47:
And they strategic, I would say considerations.

6:50:
Mata in particular has got a lot of undue advantage.

6:54:
Basically they have access to so many of users across their social media platform across WhatsApp, and anyways they are going to use a lot of these AIs.

7:05:
So making this open source, they’re kind of leading it from the front.

7:11:
And if that can be standardized, that will actually be very positive in their own business context.

7:18:
To give you an example, let’s say they open source it and there are a lot of, , startups they’re going to start working on this open source models and there will be a lot of development that has been now outsourced.

7:32:
Freely to a lot of people and there’s a whole ecosystem that is supporting it.

7:36:
Any net new optimization will directly impact their business.

7:41:
So for example, let’s say somehow there there are certain startups or in the community we’re able to achieve 10% reduction in the overall cost and optimization that will have a direct cost implication on how they are utilizing some of those tools and models.

7:56:
So indirectly it will have a.

7:58:
and then they also want to have that brand impact that they are modern, they are up to speed.

8:04:
They want to lead the whole AI game.

8:06:
So that is more on the, I would say on the meta side.

8:09:
Deep seek in general, I would say from the China context they have made it very clear that they are not behind, , they are not just copycat.

8:19:
There’s a lot of innovation and efficiency gain that they can bring into the table.

8:24:
And deepse is just one of the examples, and there are more examples.

8:28:
For Desek, I would say the timing was very good.

8:32:
On one hand, US was announcing.

8:35:
I don’t know $500 or $600 billion of AI investment and they countered it with this whole marketing and PR that doesn’t really need that much of money and that’s how we have got so much buzz around deep.

8:50:
but there are more in fact within China there are more models and companies that are coming up open source that are relatively also.

8:57:
Pretty good.

8:58:
So there is an element of having that AI dominance and challenging and sort of position yourself as a soft power, and that is not like an overnight thing that has happened.

9:09:
It was somewhere back in 2016 and 2017 when there was a model that sort of the real world champion.

9:19:
Go game that was named as AlphaGo and that was a model that Google actually built and that actually triggered that AI can actually challenge some of the human capabilities and from there there was this whole alignment in the China that by 2030 they want to be the leaders in AI and that work in progress is now showing the results.

9:41:
Definitely.

9:43:
I think it’s so extremely important for people to be investing in AI, and that’s how we’re seeing sort of the shift being in terms of investment opportunities also.

9:54:
But I know many open source AI projects do struggle with monetization as an aspect.

10:00:
So Rahul, maybe you can help us understand, like, you know, what is the sort of long term sustainability plan when it comes to deep seek AI.

10:09:
Do you think them sort of, you know, having that funding from the government is going to help them stay afloat or like, what about other opportunities and other projects that are happening in China?

10:19:
How do you think their sustainability plan is working?

10:23:
Deep began as a quantitative trading firm, so it is not something like that.

10:28:
They’re not having some revenue on their plate.

10:31:
It is they are working like a typical maybe company you can say like we’re in like some of the revenue is being generated from the services side and that particular revenue is acting as a fuel on the product side.

10:43:
So again, like, so this goes hand in hand.

10:47:
So that’s why like they were able to invest that much because they already have a well running business.

10:54:
In behind and which was acting as a fuel for this particular game changer I would say.

11:00:
And moreover down the line what we can expect from them is like a monetization via low cost APIs as well.

11:09:
See, for now the buzz is like it’s an open source.

11:13:
It’s kind of a free, it’s easy to use.

11:16:
Every new starter or every new developer can have that access, but down the line, what I personally feel is.

11:23:
They are gonna bring some low cost but premium APS services.

11:29:
Because like I believe currently their focus is on the adoption.

11:34:
There is something, if not similar, but yes, competitive enough, already in the market.

11:39:
So they want to have that audience first of all.

11:43:
They want to make the dependency first of all.

11:45:
They want that all the users across the globe should get used to it.

11:50:
So once the adoption is there, then they may come up with some pricing tweaks as little as a few cents per million tokens you can see, but again, if you have volume of users irrespective if you are putting less or if you are generating your Revenue slightly higher than your investment, but when you will do the total of all these things, the accumulative profitability will go huge.

12:17:
And that will give an advantage over the competitors in the market as well.

12:22:
So primarily I strongly believe, which is the right practice as well, that they should.

12:28:
Make this tool being used globally by maximum folks.

12:33:
This will help them to like get a major.

12:38:
, you can say a stake in the market as well.

12:41:
Plus it will be a kind of a maybe initial couple of months are going to be a good proof check of this tool as well because when we start with maybe thousands of users are going to use that, definitely the feedback will also come.

12:53:
This is basically currently evolving, maybe.

12:56:
Like we say like initially, usually the evolution is there, but maybe down the line revolution will come.

13:02:
So to start with, I think they are aiming on the, on like on the market share, I would say.

13:10:
So maybe like it’s currently a mixture of experts that they can bring in along with the optimized and ongoing refining models because these techniques are gonna help them to a greater extent in a way, maybe like it can reduce the number of active parameters during a task.

13:28:
For example, like I would say activating a fraction of a total of 671 billion parameters and allow the model to run on a less advanced hardware.

13:39:
That’s where the edge is because like the infra cost is also there.

13:43:
Then again, like because it’s a new era, I would say an AI era, so the ecosystem and the community engagement.

13:52:
It’s gonna impact as well.

13:55:
So by embracing an open source approach, Deepeek invites a global community of developers.

14:02:
We see like there are like multiple committees.

14:05:
Maybe we talk about HubSpot, maybe we talk about Salesforce, any renowned tool that is being adapted globally that comes with a global community of the developers as well.

14:16:
Who are gonna Suggest their improvements as well.

14:20:
They might build some derivative applications on top of that as well and maybe integrate its model with another ecosystems as well.

14:27:
So this collaborative ecosystem will not only accelerate innovation and adoption, but also will open the door to many other opportunities as well.

14:36:
Maybe down the line we can have some consultants.

14:38:
So once you have that much muscle.

14:42:
Like which is working on your particular community, then definitely that opened the doors to another verticals as well.

14:48:
So we need to understand like they are starting very small, but they’re very thoughtful about it.

14:54:
So initially, typically the focus is just to have the capture the market, make everyone dependent on it, or maybe like every part depended on it.

15:03:
And once we are like used to or once we are like dependent on any of such tools, then definitely the use.

15:10:
The adaptability, the acceptance of all such tools keep on increasing day by day, and that’s what they are aiming at Currently they are not focusing primarily or I would say majorly on the revenue generation.

15:22:
Only focus is to get a good amount, goody.

15:26:
Users on this tool and who are using this tool globally, not in any specific market.

15:32:
So I think these all parameters.

15:36:
I definitely wanna help them down the line.

15:40:
I think that’s really helpful, Rahul.

15:42:
From what I can understand is they’re kind of playing the long game instead of just focusing of upfront front profits.

15:50:
They’re kind of building an AI ecosystem that will sustain over time.

15:55:
What I kind of want to ask you, Arpith, is when we talk about adoption, of course costing is, you know, Huge aspect and the lower costing here helps, but there’s also a lot of ethical considerations that are coming up with kind of utilizing, you know, Chinese AI.

16:13:
India has already banned TikTok because of the kind of, you know, claims that they had of the Chinese government having access to people’s data and the similar shift we saw in the US also.

16:24:
, that’s a fight that’s kind of ongoing still.

16:27:
So TikTok is just giving you instances of what people like, dislike behaviors, but a format like Deep CKI will have a lot more private information.

16:40:
People can really have conversations about their personal lives, their personal, like, you know, their work lives.

16:46:
They can share their strategies.

16:48:
They can be sharing a lot of what’s going on in their own brain also, right?

16:53:
So a lot of people are using like such platforms.

16:56:
For more than just write an email.

16:59:
And even if they’re using it for an email, they’re using a lot of like, you know, personal data that kind of goes into these emails that should be private.

17:08:
So how do you think, you know, do you think this will kind of stop people from adopting this Chinese AI software or is it gonna kind of because of the lower costing to move forward?

17:21:
Yeah, what I think is, , as far as adoption is concerned, , if you talk about some of the major US players who are into AI space and are utilizing some of these models, you talk about perplexity, they actually have integrated DeSeek in their platform.

17:40:
They are considered to be number one answering engine right now in the purview of AI.

17:46:
And then you have Grok.

17:47:
We have also integrated deepse.

17:49:
See, the thing is there is an element of privacy and data and the whole narrative that we have for China, which could be also be debated because if you think from China perspective, they will also have something to say why they are doing it and what is their view on the US.

18:06:
It’s also like a lot of narrative that happens that we consume through media that gives you.

18:12:
About a country, about everything else, because irrespective of the origin, all of those models are consuming data and they’re hosting data.

18:19:
Europe is, for example, , a lot more stricter when it comes to data privacy and all, but what I can clearly see that the model is open source, which means you can pretty much download it and sort of host it wherever you want.

18:33:
And when you have that, you pretty much control where data resides, right?

18:37:
So let’s say somewhere, somebody in a startup.

18:40:
, community, they want to use Deepeek.

18:42:
They’re not sort of using their APIs.

18:44:
They can pretty much own everything that they have offered as an open source thing and then decide where they want to host their services.

18:53:
Is it in the US?

18:54:
Is it in Europe or in, is it in APAC?

18:57:
So I don’t think, I don’t see that major challenge there.

19:01:
It’s more of what we’ve been told about China and the whole narrative against them.

19:07:
As long as things are open source, whether it’s on the meta side, whether it’s on the deep sea side, there’s a lot of control that these guys are giving you in terms of how you’re sort of capturing data and where you’re storing it, how long you’re storing it.

19:19:
So I don’t see much of the challenges there.

19:21:
There is also an element of cost that That is being highlighted and that was that like a cornerstone about the whole PR that happened around deep sea.

19:31:
And if you unpack this a bit in order to understand the whole AI life cycle or building that model and utilizing it, there are primarily three cost compartments you can think of.

19:44:
One is the pre-training cost.

19:46:
And the pre-training cost is more like your operating expense that you will have to really incur.

19:53:
And over there they have made a lot of at least on the paper they have really claimed that it’s almost reduced by 130.

20:01:
So that’s a great innovation there.

20:03:
And then there is a second bit which is the post training cost, which people like myself who are actually working with our team to build some of the utilities on top of these models.

20:15:
You will have to sort of do a bit of wrangling there in terms of fine tuning these models that they can align with the specific use cases coming up with some rag architecture that can add an element of context to the use case that you’re solving for.

20:29:
So there is a cost, and then there’s some reinforcement learning that you want to introduce and make into that model.

20:35:
So there is an element of cost there as well.

20:37:
And then the third bit is essentially your infreencing cost.

20:41:
I think you touched upon it a bit.

20:43:
Somewhere.

20:43:
So while you’re using these models, there’s an ongoing cost there as well.

20:48:
So we’ll have to think collectively.

20:51:
I think the hype that happened was kind of just focusing on the model building cost and what I really feel eventually what will happen is these models are becoming more like a commodity.

21:03:
Everyone has access to it because it’s open source, the running, , and then find dealing of the model, that cost is also coming down.

21:12:
So, basically, , it’s available for everyone, , whether it’s Deepeek, whether it’s any other company, it’s a matter of what kind of value and benefits we are deriving from these models while we are solving the use cases.

21:29:
Definitely.

21:30:
And I guess as a follow up to that, I just want to understand, like, you know, is this sort of efficiency that you kind of mention in the model building cost, is it a signal that there’s gonna be a fundamental shift in how we sees and big tech really invest in AI?

21:48:
Like, do you see an impact happening here?

21:52:
Yeah, , what I would say is the whole episode that happened with Deep Sea, it was a good mix and a good case study of a great product on one hand and a great marketing on, on the other hand.

22:06:
And you have those two things together, , and at the very right time we all got into it and there’s massive attention globally, but at the base level, when we go into details like I mentioned, those three costs, they exist.

22:22:
So the deep see or let’s say Lama from.

22:26:
Even they are open source and they’re available and they will keep on improving and be more efficient.

22:32:
You still need a lot of computer to run those things.

22:36:
When you activate a use case, you will start running on certain infrastructure, and that’s generally speaking a bigger cost and in the current scheme of things it’s big.

22:46:
Than building a base model that was there, so those considerations we need to think of, I think the whole Nvidia stocks that I think got down to 16% or more, that was a high and eventually what will happen is as the cost will go down, there would be an increase in demand as well.

23:06:
And then demand will increase again there would be a lot of costs involved in order to draw those AI inferences, and that’s where when I mentioned why Mehta is open sourcing it because they are running AI inferences worth a billion dollars every day, right?

23:23:
Any improvement there will directly impact their bottom line, their profitability, their ROI.

23:29:
So we just need to keep that in mind, , but clearly, .

23:35:
On one hand, I would say it is a bit of a buzz about the cost, , but in, in the real sense, , there, there, there’s still a lot of costs that will go on when we just cover all those three aspects.

23:53:
What is the post pre-training, what is the post-training, what about the AI influencing?

23:58:
, and, , but with the innovation and the way they have created this new model and new approach, it certainly is a bigger question for other players where they’re always sort of dominate with, , with because they have access to so much of resources, there’s a lot of abundance.

24:19:
Just throw money, through resources, through funding, and let’s figure out without keeping the whole new approach that China has come up, it’s a lot more frugal deriving efficiency.

24:32:
So I would say that’s a mix of, I would say how I observe those things.

24:38:
Definitely.

24:39:
Thank you so much, Arpit.

24:41:
I think Rahul, what I’m also trying to understand here is like, you know, when we deepCKI is something that’s really new right now, but we wanna kind of be able to see is this something.

24:54:
I think that’s going to be able to fit in and like stay in the market and because of that, I just love to hear your thoughts on like, you know, where do you kind of see the future of AI and like what do you think will happen to Deep seek in the next 3 to 5 years to come.

25:10:
So, , it’s gonna be a trial of many such tools this year.

25:17:
Maybe the work or the output that we get from these kind of tools and the global adoption of these AI tools that’s gonna set the cornerstone for the success for 2026 for many companies, so maybe.

25:35:
I would say towards the end of this year, things will take shape.

25:41:
Because like it has just come recently in the marketing deep sea, so people will start using it, people will use it in their different technologies, different ecosystems, different from the operations perspectives.

25:54:
Multiple teams are going to get their hands dirty on it, so.

25:59:
2025, I would say it’s going to be a kind of testing the waters basically, and finally 2026 we come up with some final conclusion that in particular direction such kind of technologies are taking us and primarily or maybe deep AI.

26:16:
And But one thing that is for sure right now that the launch of this deep see AIR one model.

26:26:
Has made a significant disruption in the broader AI market by making that much advanced models at really affordable prices and much more than that easily accessible as well.

26:41:
So it’s an open source.

26:43:
It’s, it’s like it’s a cost efficient approach as well.

26:47:
So this could democratize the AI world as well, right?

26:51:
So this can help the various companies like whether they are at the startup level or whether they are like big enterprise level or maybe like their Fortune 100 or Fortune 500 companies.

27:02:
But one thing is clear, like they have set the bars high.

27:07:
So they can compete with the established players as well, why?

27:14:
Big giants will have to reexamine their pricing and their product strategies as well because it’s no longer, it will be like they have the monopoly and they can go like as they want, but I think introduction of such.

27:29:
AI models have made them rethink of that there is something other than them that also exists on the earth.

27:39:
So it’s one disruption that they have created and it’s a great work nowadays.

27:45:
And with strong backing from China’s strategic ambitions, I would say may also influence global AI geopolitics potentially they may challenge the dominance of US tech giants.

27:59:
So it’s going to be a game changer, I would say the next 3 to 5 years definitely like whatsoever tools we are using because there are like.

28:08:
Infinite number of tools running in multiple companies of different sizes but different number of employees, but different techies, but down the line after 2 or 3 years, the bar that is currently being set.

28:22:
Many of such tools.

28:24:
Will Shape their kind of working environment or working ecosystem or their working methodhologies.

28:35:
In alignment with the AI models.

28:38:
I believe It’s gonna impact hugely and vastly from a small startup to a big giant at multiple levels.

28:51:
On a different and variable scale.

28:54:
Maybe you can add a couple of cents to it.

28:58:
Yeah, definitely.

29:00:
I was also actually thinking around the other players, the big players.

29:05:
I’m actually very close to the whole Google world, and they are also in a kind of a dilemma right now because there’s a huge chunk of money that they get from the ad.

29:17:
, platform and things are changing very fast right now.

29:23:
I mean, you and I can easily confirm that for any general purpose query and no more going to Google rather than we feel a better experience and let’s say Chat GPT or Gemini or you name other platforms, even Deepsek is not that bad, although it’s a little slow.

29:40:
So, , I would say things will change dramatically down the line, .

29:46:
Google has a couple of choices.

29:49:
One is, can they really pivot their search engine game more from this whole new AI era that is emerging at the cost of cannibalizing their market share that we have seen in terms of the advertising.

30:03:
, because obviously they are public listed companies, they are answerable to the investors and then the whole ecosystem out there.

30:13:
So it is, it is going to be tough, and it’s going to be tough for a lot of other players in a similar situation.

30:20:
, the general pattern has been, , if you can’t just buy it.

30:27:
So a lot of the bigger players will try to consolidate, will try to acquire emerging smaller players.

30:35:
For example, there could be a likelihood that Google might acquire publicity.

30:39:
You never know because they are clearly sort of they’re saying that they don’t compete, but they’re clearly competing with Google and somewhere in the initial mission of Google they wanted to.

30:49:
Really create a search engine that can give you the most perfect answer.

30:53:
But what ended up was the list of those 10 different links that we’ll have to research and really identify what’s most relevant to what we are looking for, and that that has changed now and there’s still a lot of money pouring into the advertising that Google sort of control, but the whole model from being ad driven to subscription driven.

31:19:
It’s a new game altogether.

31:22:
, and then we see a similar situation in other bigger players.

31:27:
So it’s an amazing time, I think, , we all will have to pivot, , unlearn, relearn.

31:35:
, it, it is for our at the individual level also, there’s a lot of things that’s required at the company level, at the organizational level, and the country level as well.

31:44:
I mean like, clearly if China can do it, why cannot India, for example, right?

31:48:
And I’m sure there are people who are working, given, and there’s a lot of inspiration.

31:52:
I say I would say that Deep Sea has given to India when I speak to some of our people in in in the similar space.

32:00:
We are, we all are discussing about this.

32:02:
Why not India?

32:03:
So let’s see.

32:05:
No, I think that’s really insightful, and I think this has been such a lovely discussion.

32:10:
I think we have had the opportunity to dive really deep into some of the aspects of deep CKI but also AI in general.

32:18:
So before we start kind of wrapping up this session, any closing thoughts, any thoughts about the future ahead that you’d like to share, , Rahul, , maybe you can start us off and then you can add on to that.

32:34:
Cool, I would say .

32:37:
We are entering into a democratized AI landscape, I would say.

32:45:
Where with the introduction of deep 6R1 maybe we can consider that as a mainstream, a wider range of companies from startup to established firms will be able to leverage advanced AI without massive investments.

33:01:
So this shift could democratize AI innovation and challenge the dominance of the traditional tech giants.

33:10:
So maybe like we all have seen like just within the one week of the launch like as Arpith mentioned, shares of NVD are plunged by close to 17%.

33:19:
Then Google Patern Alphabet, which has heavily invested in AI development, lost roughly I would say $99 of billions in market value.

33:30:
Then Microsoft as well poured billions of dollars into OpenAI, lost almost $71 billion in their value.

33:39:
So.

33:41:
We are into the era of a drastic transformation.

33:44:
We used to talk about the digital transformation initially and nowadays as well, but now I would say digital as well as a drastic transformation as well.

33:54:
So let’s see what the future.

33:57:
Has for us, but for now.

34:01:
AI seems to be the need of our need of every business as well, so who can take the advantage of it are definitely.

34:10:
Gonna make some good amount of revenue or maybe I would say the operational efficiency as well.

34:18:
Raoul, I think you made a great point around cost and the kind of money that that’s really going on and my focus area has always been the art of understanding whether it’s making sense from business context.

34:32:
I think that would be a vital point.

34:34:
In fact, we were discussing this yesterday with our CEO.

34:37:
Let’s say.

34:38:
With all this AI chatbot and agenttic AI where we are sort of really saying that we will cut the jobs and then there will be an agent that will do everything.

34:49:
I think what is really required is to understand the cost benefit analysis.

34:53:
So let’s say you have an agent or an ecosystem of agents, what’s the annual CDC for those agents?

35:01:
Is it?

35:02:
Less or more than a regular person who’s working in your office, can we get into those kind of numbers and understand what’s the real ROI of AI?

35:11:
I think that that is something that probably will unpack this year because the base models are ready, there is some level of work already done whether you pick any big player, be it sales force, they have some offering.

35:25:
Around AI agents and layer, but everyone is trying to understand what could be that pricing strategy, what could be that pricing model, what could be that combo that can really work well for their own business and also work for their end customers.

35:42:
So that will happen and I think it’s a very interesting, exciting time given now we all can leverage these open source models on one side.

35:52:
, we have the clouds, whether it’s a your or AWS or GCP, which means the infrastructure is ready, right?

36:01:
You spin a server today and run a model and come up with a great use case that can solve customer problem.

36:08:
Great.

36:09:
, but at the same time, the question is that is it delivering a tangible ROI and business value that is sustainable.

36:20:
Mhm.

36:22:
Definitely.

36:23:
I think, , thank you again for, you know, kind of giving your final thoughts on this.

36:28:
I’d like to take a minute to really kind of summarize the biggest takeaways I had from this session today.

36:35:
I think something that kind of came up front was the future of search being changed.

36:40:
So I think Arpit you kind of mentioned about that and like a search engine that understands exactly what you mean and gives you the perfect answer.

36:47:
No endless scrolling ads cluttering your screen.

36:50:
That’s the kind of future we’re looking for.

36:53:
But I think to kind of get there, Google kind of has to rethink their entire revenue model.

36:58:
They have to shift from the ad-driven empire that they’re on.

37:01:
And it’s a massive shift that’s not going to happen overnight, right?

37:04:
So that’s something that’s really evolving.

37:07:
Another aspect that really touched me in today’s conversation was like we started off thinking US is the king of breakthrough ideas and you know, they have good universities, research labs, and culture that really avoids risk taking, but China is also insanely fast at, you know, kind of scaling and executing these things.

37:25:
And while American companies.

37:27:
Spent years in launching and protecting, you know, their tech.

37:32:
Chinese firms are kind of like, you know, doing this launch iteration and doing it in a very agile manner if we might say doing it in a very fast pace.

37:40:
And there are other countries also like, you know, India can potentially also get into this kind of atmosphere and contribute.

37:46:
So AI is no longer going to be democratized or like have a monopoly of just, you know, one particular.

37:53:
There are going to be many within from one country like we kind of talked about, you know, with meta and Google also having their AI sort of ecosystems, but another aspect that we kind of touched in today that really stands out to me is just talking about how investments are going to be rewritten.

38:16:
We kind of talked about how for VCs and even big tech how they’re investing.

38:20:
AI is going to change.

38:22:
I think they built rivals of GPT with like just 5-6 million by focusing on efficiency rather than the size of the data that it has.

38:32:
So I think like, you know, there are like the future what what’s kind of OK to say now is that AI is evolving at a very lightning speed and deep seek is just, you know, one.

38:46:
Of the players that’s kind of changed the game, the next few years, who decides who will adapt, who will fall behind and you know what the future will look like is something we’ll need multiple discussions to have.

38:59:
, and I think from this conversation it’s kind of safe to say that like, you know, this is just the beginning.

39:06:
So thank you so much.

39:07:
Is there anything I missed in the summary that you guys might want to add before we just wrap up this session?

39:13:
I think great summary.

39:14:
I would just add that do we really need GPT to summarize things because you’ve done a great job on the lighter note.

39:22:
Thank you so much.

39:25:
I think just.

39:28:
No, I think, , again, thank you so much to both of you for, you know, being a part of this session and kind of really sharing what’s to come and what’s really happening right now.

39:39:
I think I learned a lot from this session and I hope our listeners did too.

39:44:
So as last thoughts, I guess Arpit, you can start us off, like, you know, if people, our listeners want to learn more about you and your work, where can they really reach out to you?

39:53:
And then Raul maybe you can share after that.

39:56:
Yeah, I’m an open network.

39:58:
Reach out to me on LinkedIn and if you want to learn about the growth or AI analytics platform, just mail me at info@digro.com.

40:08:
Amazing.

40:09:
And Daul, for you.

40:11:
Yeah, like I’m available on the LinkedIn by the name of Rahul Sinjwani and like proud member of Growth Natives team, so maybe under the Growth Natives website you can find me there as well.

40:22:
Amazing.

40:23:
Thank you so much, Rahul and Arpith.

40:25:
Today was a great session and like, you know, I think this is a podcast where we’re here to kind of keep you updated on what’s really happening.

40:32:
So if you like this episode, please share your thoughts with us.

40:35:
And until next time, stay curious and keep innovating.

40:38:
Thank you so much.

See more webinar