Snowflake CEO Interview: DeepSeek Shines as a Model, ChatGPT Wins as a Product

In the enterprise data space, Databricks and Snowflake are currently the main competitors.

Databricks is currently one of the most highly valued startups in the world, while Snowflake is a $60 billion public company with $3.5 billion in annual revenue and 30% annual growth. Sridhar Ramaswamy is the current CEO of Snowflake. He joined Snowflake in 2019 after the AI ​​search company Neeva, which he founded, was acquired by Snowflake for $150 million.

Prior to starting Neeva, Ramaswamy spent 15 years at Google, growing AdWords and Google’s advertising business from $1.5 billion in revenue to over $100 billion.

After joining Snowflake, Ramaswamy led the company’s AI strategy. He quickly promoted Cortex AI, a managed service for building large language model applications in the data cloud, and AI intelligence frameworks such as Snowflake Intelligence.

Sridhar Ramaswamy was recently interviewed by the famous podcast 20VC. The host Harry Stebbings talked with Ramaswamy about whether the recently popular DeepSeek will pose a threat to OpenAI, what are the key elements for transforming models from commercial products to sustainable product suites, and how Snowflake innovates in the face of competition from NVIDIA and Databricks.

01 

DeepSeek leads the way in model commercialization.

But ChatGPT is a good product that is more attractive

20VC: As an investor looking at the market right now, I’ve been thinking about where sustainable value will be created. It looks like models are becoming commoditized faster than expected, and I think DeepSeek has proven this over the past few weeks. A lot of the application layer stuff seems to be just lightweight products built on top of commoditized models. How do you think about sustainable value in the AI ​​market today?

Ramaswamy: I think part of it is that — I was having a conversation with some people at OpenAI and an application company, like a coding platform — the lines between infrastructure providers and application providers are very blurred today , which is weird. If there are brand new applications that need to be created, basically OpenAI, Anthropic, Microsoft, or Google, who are doing big models, will go and create them directly.

They think, “Coding assistants are taking off, so I’m going to make a coding assistant or something like that.”

“AI is generating better legal documents, maybe we should release a ChatGPT-like product for the legal industry.”

To me, this is one of the reasons why creating value in today’s environment has become so hard to pin down. I think where the value is absolutely is in companies that have customer relationships, provide clear value, and are willing to embrace AI quickly to prevent being replaced by disruptors. This is why I’m optimistic about Snowflake, because we are a data platform that helps people collect data, understand data, analyze data, and run predictive tools (like machine learning) on ​​top of data. I think AI is a huge accelerator of the data lifecycle and data access.

Will anyone start with these new foundational capabilities and end up doing better than Snowflake? We bet they can’t. So I think AI is clearly creating value. For example, Salesforce is doing some really cool work with Agentforce, and they’re very clear that their advantage is in relationships, and they’re self-disrupting and proactively attacking. But all the new value creation, in the long run, still seems very fuzzy.

The product itself is valuable. It is important to understand that OpenAI is successful not just because they can always create the base model at the best price, but because they have product experience. According to statistics, they have about 500 million loyal users, which is very difficult to replicate.

20VC: Do you think loyal users will allow a product to top the charts in one day?

Ramaswamy: You don’t switch from ChatGPT to DeepSeek. And OpenAI is not stupid, if they could power ChatGPT by hosting DeepSeek, they would do it without hesitation.

20VC: Why don’t you switch to DeepSeek? It’s free.

Ramaswamy: There are a bunch of additional features behind ChatGPT. It’s a product, not a model, and that’s a big difference. There’s a reason why companies like Anthropic haven’t done as well, because it’s primarily operating at the model level. The ability to generate images, the ability to upload files, the ability to run a small piece of code – that makes ChatGPT a complete product, and I think that combination has lasting appeal.

20VC: I had Sam Altman of OpenAI on the show, and I was still thinking about branding, and the first thing he said was, “We’re going to crush startups.” I was like, “Oh my god, that’s bad.” That’s not good for my venture capital fund. But my question is, if you were Sam today, what would you do? We had the founder of Grok, and he said he had to open source.

Ramaswamy: Sam’s success is as I said, he is building a consumer product that is close to the scale of Meta and Google.

Part of this is probably because OpenAI has an incredible publicity machine, but what other company can you think of recently that got to 500 million users with almost no advertising? And it wasn’t even a social network. OpenAI has always been a bit mysterious, as is their closed-source model.

It’s becoming increasingly clear that they’re very good at misleading people about certain issues. And companies like DeepSeek are debunking some of these myths, which is a good thing. But I would separate the success of OpenAI as a company from the future of the model. Obviously, they have other goals, like dominating the field of AGI. But Sam, I completely agree with you on one point: it would be terrible to build a startup on the basis of OpenAI.

02 

If a startup finds PMF,

It’s not so easy to copy

20VC: There are two questions that puzzle me. The first one is, if NVIDIA enters your field and becomes a competitor, would you be worried about it?

Ramaswamy：I have to be vigilant about this kind of thing. AWS has Redshift, Microsoft hasFabric, Google has BigQuery, Oracle has its own data warehouse platform. You always worry about being a little mouse in a world of giants. If they sneeze, you might get blown away, and I understand that.

But on the other hand, I tell people that OpenAI and Anthropic have created the best models on the planet for the last three years, not Microsoft, Amazon, or Google. Gemini is OK, but honestly, a fast follower at best.

Harry, you know this better than anyone. When a startup finds product-market fit (PMF), it becomes a living, breathing thing. It is magic. It is not created with a few nice words or ideas.

Snowflake is the embodiment of this magic, and it’s hard to create a data product by copying it, you have to keep innovating. Obviously, Databricks has become a very strong competitor in the data space over the past five years, but Snowflake also has a very strong product and operates very differently from Databricks. We need to continue to innovate and stay ahead.

Money can’t buy you an amazing base model, and it can’t buy you Snowflake.

20VC: One question about NVIDIA is on the one hand, and the other is Databricks, which you mentioned. Let’s say some of your board members and investors say that Databricks may be leading in AI workloads, such as Mosaic, MLflow (machine learning process management tool), and direct model hosting. What do you think about this? Do you think Snowflake is lagging behind in this regard?

Ramaswamy: We should make a distinction between machine learning and AI. I joke with people that all the machine learning people label themselves as classical AI. It seems like they were first to market, that’s their sweet spot, and they’ve been doing that for a long time.

We are definitely in catch-up mode, but AI is a much younger field that barely existed before ChatGPT. I feel very confident that we are not only at the forefront, but ahead of them in many ways. We are thinking systematically and quickly about how to do data transformation and data engineering better.

How do we make sense of unstructured data? How do we reliably turn it into structured data? We’re bringing all of this together into an intelligence framework called Snowflake Intelligence, which is coming soon. I’m very happy with where Snowflake is at in the AI ​​space.

We have a lot of impressive customer stories like Siemens, Elevance, Bayer Pharmaceuticals, and a lot of other companies are using Snowflake AI in production. So I think we’re doing pretty well there. This is a fast-moving space, and I’m proud of the fact that I’m working with the AI ​​team in this fast-moving space, and my attitude is: I can run as fast as anyone. Bring it on!

20VC: I would like to ask, which structure is more conducive to innovation and winning the competition? A super late-stage private company like Databricks or a public company like Snowflake?

Ramaswamy: Innovation is not a choice, it’s something we have to do. Do we face more constraints than a private company like Databricks? Absolutely. Obviously they can acquire businesses without worrying about free cash flow and things like that, and they have twice as many salespeople as we do. But it’s easy to burn money, and it’s easy to spend money out of control, and you should be careful.

Are we operating under constraints? Absolutely. But that’s what innovation is all about — how do you push the envelope within constraints? In a way, DeepSeek is another example of how having “rich relatives” isn’t always a good thing.

20VC: How have these restrictions helped you and how have they hurt you?

Ramaswamy: For example, in AI, we decided that we would not make unreasonable demands on the size of the team. I didn’t ask for a blank check, we knew exactly what we needed to do, we needed to have a clear vision of what we were going to build, and we had to execute very well. So with a modest investment, we were able to catch up very quickly in AI. I think those constraints bring clarity. You don’t try to do everything, and you don’t chase businesses that are not scalable.

Limits can be problematic when there are dramatic market reactions. For example, people might read too much into changes in quarter-to-quarter numbers. To some extent, the ups and downs of Snowflake’s stock reflect the increased scrutiny facing public companies.

In an ideal world, I could tell my team and investors, “We’ve got this under control, don’t worry, everything’s going to be fine.” But in reality, you have to worry because people make decisions like mortgages based on the value of the stock.

So you get into the problem of externalities that are difficult to manage, and that raises the bar for responsible operations. That’s something we had to learn very quickly last year. But it’s also an example of how, as a public company, you don’t have to worry about dramatic reactions because they can set off a cascade of second-order consequences.

20VC: If you could, would you choose to take the company private to remove these restrictions?

Ramaswamy: No, I think transparency is a good thing. It provides liquidity and it allows you to evaluate your performance, which I think is an important part. Sometimes it’s easy for a CEO to lull themselves into thinking that everything is going to be fine, when you’re not actually getting any feedback.

Our public markets are very realistic, either you make money or you don’t, either you have free cash flow or you don’t, and I think that sense of reality is helpful.

Note: Second-Order Consequences refer to the subsequent impacts caused by the first-order consequences of the initial action. It goes beyond the direct and superficial results and requires in-depth thinking from the dimensions of time, system, causal chain, etc. to be foreseeable.

03 

AI is definitely in a bubble.

But there is still room for innovation

20VC: I want to talk about enterprise adoption of AI. I was sitting with a bunch of CEOs at a conference the other day, and they were saying, “You know what? We’re still waiting for the new guys in the technology space to really deliver a return on investment.” My question is, do you think we’re going to go through a period of excitement and then plateau like we did with autonomous driving, or do you think we’re going to see exponential acceleration in enterprise adoption of AI?

Ramaswamy: I think the growth will be more modest. On the other hand, I can tell you with great confidence that AI is creating value today and will continue to create lasting value.

Many things that were once extremely difficult for us are now much easier. If a CEO tells you that they don’t see any value from AI, I’ll say, “I’m not sure you understand, so let’s look at some examples, the obvious, simple ones.”

I did 30 meetings in Davos. I used some AI tools like dictation and transcription to write notes for myself. I also wrote a lot of notes by hand, and someone on my team said, “Really, 25 pages? I really don’t want to read this. Can you give me a summary?” So I uploaded all my notes to the cloud and typed in, “I want a one-sentence summary of each meeting.” And it came out with a concise and beautiful summary. Similarly, we have a chat system internally that can access structured data and handle questions that you need to click through the dashboard many times to find the answer, or even questions that require a lot of context. This has brought huge value.

20VC: Davos is a gathering place for global CEOs. What have you learned from the emotions you felt in Davos?

Ramaswamy: First, I took a very pragmatic approach to Davos, rather than focusing on the bigger issues. It was my first time attending, and I was terrified about being away from work for five days, especially with board meetings and earnings season coming up, but it turned out to be well worth it.

I have a lot of meetings here and I have met with a lot of key people in the company. The message I hear more from CEOs is: “Help us create practical AI and show us what can be done.”

When I walked them through how we can leverage unstructured data to quickly create a chatbot based on a document corpus, or how to leverage structured data, they all understood.

People get excited when I tell them that now you can mix and match these capabilities to create an intelligent platform that accesses a wealth of relevant information in one place.

When I explain to them how we can automate parts of the underwriting process, like insuring a building, by integrating all the structured and unstructured information, they say, “That’s awesome.” We need to keep focusing on practicality, but I don’t encounter too many enthusiastic skeptics about AI because I just pull out my phone and show them five silly but fun things I did with AI yesterday.

20VC: It feels like we’ve never seen the giants innovate as fast as we are now. For example, Scott Belsky (former Adobe executive), obviously his company used to move slowly, but now everyone is moving faster. In the past, we thought the giants were slow and bad, but now they are not.

Ramaswamy: Yes.

20VC: Have you ever seen giants move as fast as they are now?

Ramaswamy: We all remember history, and no one wants to be the person known for the IBM-Microsoft deal (the early collaboration with IBM led to Microsoft’s success).

People understand what a dramatic shift that was, and how companies like DEC disappeared. The first building Google took over in Mountain View was the purple building of SGI. For five years after that, I would always tell everyone who joined my team, “SGI built these buildings.” And then I would see if they understood that. So we learned from that.

But I would say that mobile was a major platform shift, and the incumbents did pretty well in it. Remember when Mark Zuckerberg initially tried to push Facebook’s mobile web, then gave up and moved to apps? The same thing happened with Google, when they moved search to mobile. I led the team through the five-year “horror period” to increase mobile payments from 10% to 100% of desktop. So I would say that even in the last round of technological change, all tech companies have become pretty smart. Amazon was not disrupted by mobile.

Mobile did give rise to some entirely new things, like Uber and Lyft, but that generation of companies had learned how to deal with the huge challenges of platform disruption. That’s why you see these companies investing like crazy in the future, because they saw something and had the money to do it. That’s why Facebook became Meta.

Mark Zuckerberg didn’t care about the failure of the Metaverse. He was like, “Bring it on! I’m focused on AI now.” I think that’s a sign that the company has really learned from the past and is trying to innovate.

20VC: I saw Mark Zuckerberg invest $65 billion in data centers, and you saw the $500 billion announcement for Stargate. I know there’s equity and debt, but we’ve never seen this much money thrown at something. This is the AI ​​arms race, where is it going?

Ramaswamy: The AI ​​bubble will eventually burst, just like any other bubble. There was a really interesting and insightful conversation between Ben Thompson and Nat Friedman of Stratechery about whether this is a “good bubble” like the 90s telecom bubble, which ultimately laid fiber optic cables around the world, benefiting Google, Facebook, and you and me. Or is this a “stupid bubble” like the early dot-com bubble, which burned tons of money trying to deliver groceries to you, and we had to wait 15 years for Instacart to show up?

Let’s face it, you and I don’t know. If most of these investments go into things like electricity and construction, you can say, “This creates a surplus for the world, and good things will come out of it.” On the other hand, if they go into hardware that depreciates quickly, then the value is gone in an instant, and I think it’s too early to tell.

But back to your question – how do we do that? There is still room for innovation. I don’t think it’s possible for a company like OpenAI to cover all workflows. What you and I need to do is identify those areas that are still ripe for disruption and create value there.

I think Harvey is a great company that will not be disrupted by OpenAI. I think finding niches like these and investing in them is the key formula for our success. Money doesn’t always buy everything.

04 

AI ToB market,

There will be no monopoly entrance

20VC: Speaking of money not always buying things, some in the market say, “Where is the growth for Snowflake?” To what extent can you “buy” growth? What is your view on M&A strategy?

Ramaswamy: First of all, we have $3.5 billion in revenue and are growing. Obviously, we have to continue to maintain this momentum.

What we mentioned in our last earnings call is that we have significantly expanded the footprint of Snowflake. We used to focus on the data layer, and we also did a little bit of machine learning.

Now we’ve expanded from that to: We’ll help you with data ingestion, we’ll help you with data engineering architecture. We want to be a critical part of your data engineering, analytics, machine learning, and enabling user access through AI. Whether it’s the tools we create or platforms like Snowflake Intelligence, our scope has expanded significantly.

We have a team that is not only pushing how to capture more analytics market share, but also going into other very large data segments and disrupting them one by one.

Of course, we will continue to look at some companies, but non-organic acquisition strategies that are not related to our core business are not what I am interested in, and I think that would be a distraction. Snowflake is a product-driven innovation company, not a private equity company. Therefore, most of our growth must be driven by product innovation. Will we make some smart acquisitions? For example, Snowflake spent about $150 million to acquire Neeva, and I think that investment has paid off. These are things we should do.

20VC: Is there a revenue line that is currently small or insignificant for Snowflake but will become a dominant revenue source in 7 to 10 years that we should pay attention to?

Ramaswamy: 100% AI. I think it will be a big thing and it provides an opportunity to disrupt current business intelligence (BI) and be more consumer-oriented.

The key to accelerating Snowflake’s growth is that people are building applications on top of Snowflake. Companies like JPMorgan Chase, BlackRock, and Siemens are building data applications on top of Snowflake. Our platform allows them to do some very clever things, like integrate their own data with their customers’ data to create data applications.

I think it’s still small right now, but I expect it to be a huge overall revenue opportunity for Snowflake. The cool thing about it is that you go from being an expensive line item to being part of these companies’ revenue. As you can imagine, the dynamics of that partnership are much better. I hope we can say, “When you make money, we make money.”

20VC: We can expand on the model space discussion. I wonder what it will look like in 5-7 years. Will we live in a world with many specialized, verticalized models or a world with a few generalized, hyperscale models? I remember reading that historically some people thought that search would eventually be replaced by some kind of verticalized service, but they didn’t anticipate that there would be these huge multi-domain giants that control almost everything.

Ramaswam: That is a very good question. Unfortunately, my answer would be that predicting the future is always difficult.

If we break down Google Search and analyze why it became so dominant, first and foremost, Google made a series of deals to essentially become the default search engine for many portals or browsers, like Yahoo, AOL, Firefox, and all the PC manufacturers. This was a very deliberate strategy to be at the heart of search. Microsoft ignored this opportunity in those early years, and that’s how Google got started. Organic marketing is part of it, but it’s only a small part of it. What really worked was that “gravity effect” that allowed them to knock down other verticals one by one.

Interestingly, Bing’s predecessor, Live.com, actually had a better image search than Google. But Google’s approach was, “You’re looking for images? We’ll put them right on the main search page, and you don’t have to go anywhere else.” That’s how Google conquered almost every vertical, whether it’s shopping, video, or maps. Through this core sticky attribute, Google conquered the consumer world.

If you look at the AI ​​space, on the enterprise side, and especially on the consumer side, I don’t think there’s a single entry point. But ChatGPT is potentially on its way to being that entry point. So if you’re someone who’s betting on the consumer side, then in terms of variants and specialization of models, I think that’s going to favor the incumbents, which are 100% ChatGPT and OpenAI, and the entry points for other models will eventually become fragmented.

I used to worry, “Will there be a single enterprise entry point that rules them all?” I still worry about that. I think that opportunity is still there, but on the enterprise side, I see all kinds of specialization opportunities emerging because there hasn’t been anything like Google search in the last 50 years.

20VC: I also enjoyed the historical details about Google’s distribution strategy through partnerships. I had no idea it went to this extent.

Ramaswamy: We all thought Google was successful because of “perfect consumption.” One day, everyone decided that we should use Google.

20VC: Distribution, which partnership do you think has had the biggest impact on Google’s distribution capabilities?

Ramaswamy: Yahoo and earlier companies, those were portals. AOL didn’t know how to do search, and Yahoo didn’t think search was important.

20VC: Does it really matter if you get paid first?

Ramaswamy: 100%. If I remember correctly, we paid AOL more than we made. The founders of Google were incredible and made a series of extremely smart business decisions. Yes, the product was great, but people underestimated the power of business.

20VC: I see a lot of arguments about software engineering, saying that these are the first fields to be dominated by AI. Do you disagree with that? Would you say that software engineering still has tremendous value and students should continue to learn it? Or would you say that no, this is indeed the first field to be affected?

Ramaswamy: I think all kinds of knowledge-based professions will be greatly impacted. I have a simple but fairly accurate description of what AI can do: it is a powerful translation layer that can convert between all kinds of structured and unstructured knowledge, understanding them in a way that only humans can. So I think any knowledge worker, including software engineers, should embrace this technology and see where it goes.

It’s too early to conclude that this means software engineering will become a good career or become a narrow field of expertise. After the Internet came along, big media like the New York Times could cover every corner of the world, and maybe you don’t like them, maybe you want to see other perspectives, but it did become centralized. The same thing happened in the music industry, and other fields. So, I think it’s hard to predict how big an impact it will have. But as long as software remains important, I can see a huge opportunity to apply software and AI to more and more areas.

So I don’t think software engineering as a profession is going to be obsolete any more than, say, an analyst or a CEO. I joke with people – I’m an email machine, that’s all I do, I talk, I write, I read. So I think the key is to embrace the future, see where it goes, and be flexible. I don’t think software engineering is going to go away anytime soon.

05 

True Leadership:

Tell people what they don’t want to hear and make them stay

20VC: I want to ask you some questions about early career, starting from when you were in college or your first job. Did you think you would become a CEO at that time? Why?

Ramaswamy: No, I didn’t think about it, especially as an undergraduate, and not even when I was doing my PhD. In fact, when I was doing my PhD, becoming a professor was more of an ideal choice. But at some point, I wasn’t that interested in doing research, so I switched to software engineering. When you are in the right place at the right time, great things happen.

20VC: When you look back on those days and reflect on them, now as the CEO of a $3.5 billion public company, do you think, “What lessons can I learn from this?” What are those lessons?

Ramaswamy: Relentless drive? I still remember how exhausting it was to have a discussion with Larry or Sergey. They would argue with you on every topic, but that’s how good things often came out. You looked at every possible detail and were very hard on yourself. I joke to people that when I became the head of advertising, things like working with the legal team became much easier because they had been pushed to the limit by Larry and Sergey on projects like Google Books or YouTube—YouTube was a legal nightmare in its early days. So that’s my lifelong takeaway: Stay the course and think in first principles.

That’s Google, and I think the energy that they bring to every discussion and their relentless pursuit of the truth and the right business outcomes is definitely something that I still practice today.

20VC: People always assume that you get better over time, especially as a CEO. What are some things you think you’ve gotten worse at?

Ramaswamy: That’s a good question. Physical skills do become harder to maintain as you age, that’s just the way it is. I used to be able to run 20 miles without hesitation, without any preparation, and I can’t do that anymore. But I think the mental side of things is largely manageable, and that’s what I’m happy about. I’m not relenting in pushing myself to learn, measure, and adapt.

Another thing we have to be realistic about is that as you get older, the time you have to invest in something new becomes limited. You have to be humble about that. I jokingly tell my sons, “You’re not going to be a concert pianist in your lifetime. That part is over.” So we have to accept what can change and what can’t change.

But on the spiritual side, I would say that for an agile and driven person, the world is your oyster. There are so many language models now that can help you create custom programs and do amazing things that you simply couldn’t do before. So in that sense, I think there are even more opportunities on the spiritual side than before.

20VC: What advice would you give to CEOs like me and others who work in high-intensity jobs? We need to bring others along, but it’s hard.

Ramaswamy: I think it’s critical to make sure people understand the big picture, the fleeting nature of the opportunity, and how quickly it can disappear.

The company I run generates about $3.5 billion in revenue a year. If you think about the growth rate of world GDP, adjusted for inflation, it’s growing 1% to 3% a year. At 1%, it would take about 100 years to double it, and we’re talking about doing that in two years.

The first thing I tell people is that we are very fortunate to be in an environment where we can create so much value and grow so fast. It takes extraordinary talent to sustain and accelerate that growth. It’s not for everyone – the expectations are high, the rewards are high, the opportunity is huge, and the opportunity cost is even greater. So I think the question is, who do people want to be at Snowflake?

We absolutely want to be the data engine for every business on the planet, which I think is a really amazing mission. Plus, we can build an iconic company in the process, and that takes special people.

20VC: When you hire someone, you think they will be able to adapt to different stages of the company’s growth, but it turns out they are wrong. Why do you make this mistake?

Ramaswamy: These things are unpredictable, life is long, people change over time, and certain things — like having $10 million or $50 million, just to name a few — have an effect on people.

I joined Google in 2003 in part because some of the people who had joined in 2001 and 2002 were suddenly worth $500 million.

Not everyone can scale their capabilities. I talk to a lot of people about what it takes to double the size of a team. My advice is usually this: every time you double the size of a team, all those things that made you great on your previous team are usually going to be huge barriers to success in your new role. I’m not sure you can really internalize that. It requires a lot of reinvention and redefinition of who you are and how you operate.

Only a few are able to make those transitions, adapt and thrive, and I see it as my responsibility, my job and my duty to give people that opportunity, but at the same time, I can be ruthless and cold because I give people time and opportunity.

One thing I’ve learned as an adult is how not to run away from uncomfortable conversations. I talk to people about what’s working well and what’s not working well. I give them time. If it doesn’t work out, then that’s it. I’ve asked some people to do half the work and told them that was the better option for them. It was one of the hardest things I’ve ever done in my life.

20VC: You mentioned demotions earlier, does that mean you should just fire them? Someone once told me that if you’re willing to take less in return, never make a deal. I’m applying that principle here: if you’re willing to let them stay but give them less responsibility, shouldn’t they just leave?

Ramaswamy: All of this is happening in a rapidly changing environment. An area that you thought you could handle with 20 engineers, all of a sudden you need 100. That person who was managing 20 people is now responsible for 40 and needs to scale to 100. They are struggling, and the business can’t wait. It’s not because they’re bad people or they can’t do the job, but for me, finding the right context and framework to help someone succeed in that particular moment is a conversation that can be had.

I once had a director who accidentally ended up running two different areas. I told him, “I’m going to take one of these areas. By the way, the area you’re in, it’s going to be a $40 billion business, but it’s five, six years away (Google Shopping). This is a huge opportunity that we see.” That person took the job, eventually became a VP, and was wildly successful.

This takes a lot of convincing because whenever you have a conversation like this, people’s first reaction is usually “I failed,” but I didn’t. You were given an impossible task. Let’s reshape it so you can succeed again. That’s real leadership: telling people what they don’t want to hear, painting a vision for the future, and getting them to stay. It’s hard.

20VC: Do you think that richer leaders make better leaders because you’re more determined and less worried about downside risk and not being afraid of losing your job or something?

Ramaswamy: No, I think it can make them apathetic and maybe even overly tolerant of big risks. It’s a balance, and as a leader you’re always dealing with different stakeholder groups, and it’s important to remember that it’s not just about you, it’s about the employees in the company, the shareholders, and the customers. We went through a very tough time as a large company last year, and when it became clear, we realized that this affected a lot of different people. It’s not always about going all-in on high-risk goals. Sometimes you do need to make a 90-degree turn to get somewhere. But I don’t think that being in a position where you’re not personally impacted by the outcome is necessarily a better leader.

Author：FounderPark
Source：https://mp.weixin.qq.com/s/cKwHyNWTeDaJhcaQfP0h_w
The copyright belongs to the author. For commercial reprints, please contact the author for authorization. For non-commercial reprints, please indicate the source.