Tina Hou: [Foreign Language] Thanks management for taking my question. The first one is in terms of the distribution network building. As management mentioned, there will be a lot of changes and optimization as well as deepening into the lower-tier cities into the second half of this year as well as next year. So should we expect to a higher level of sales and marketing expenses accordingly in the second half of this year as well as 2024? Thank you.
Brian Gu: Hey, Tina, this is Brian. I think you’re right. We’re actually making changes to our sales channel and strategy. We actually are envisioning more partners to be our sales investor — store investors and sales agents. And that effect, I think, will lead to a better penetration of lower-tier cities as well as, I think, optimize the sales performance because we actually simultaneously will read out weaker performers on the sort of stores, both in terms of owned as well as our current channels. Interestingly, actually, this actually, we envision will lead to a more efficiency gain and lower sales marketing cost for our operations. Because we think right now, the optimal efficiency has not been achieved with our current model.
And also, we actually have not partnered with enough high-quality and efficient operators. So with renewed focus on partnership and also higher standard setting for our sales partners, I think, will lead to high efficiency. That’s actually we envision more efficient and lower sales and marketing ratios.
Tina Hou: [Foreign Language] So my second question is regarding the operating level breakeven point. Since management mentioned that the gross margin will start turning positive into Q4 this year, just wondering because we’re still continuing to invest in R&D continuously. So what would be the expected operating profit breakeven point? Thanks.
Brian Gu: Yeah. Thanks, Tina. I think for guidance on overall company breakeven as well as more cash flow forecast, we’re actually maintaining our current view that by 2024, we will achieve quarterly free cash flow positive. So that’s something we’re confident next year we can actually hit. And then for the overall breakeven year, we still maintain sometimes 2025 will achieve breakeven for the whole company.
Tina Hou: [Foreign Language]
Operator: Thank you. And the next question comes from Paul Gong with UBS.
Paul Gong: [Foreign Language] So my first question is regarding the autonomous driving without HD map. I can understand it could offer more coverage of the applications in terms of cities, how does the cost compare with previous version with HD Map? And to achieve that, what are the key challenges we have to overcome?
Xiaopeng He: [Foreign Language] Thank you for your question. This is slightly technical. I’ll try to provide a simple and straightforward answer. When we are equipped with high-definition maps, it’s very easy for XNGP to be guided because you will know when you expect to change lanes before turning left or right. However, without the high-definition maps, things will get much trickier because the XNGP system will have to judge by on its own where to change or what kind of situations will change beforehand. For example, it will need to be equipped with the hardware sufficient for it to actually identify dotted lines on the road or solid lines on the road or signs and characters that says that this is a left turn corner or right turn signals.
This is just like when a person visits a new environment, you will get — you will need some time to get acquainted. There is always a possibility for mistakes. However, there are a lot of pros that come from this non-HD map reliant XNGP. First of all, you don’t need high-definition maps, which means that you can go wherever you want as long as you have, for example, the conditions and the hardware equipped that comes with the XNGP and basically, theoretically speaking, as long as your phone can guide you there, our XNGP without the HD map reliance can also guide you there. And the second benefit is that you don’t need to jump through hoops to get policy approval because of this high safety. And the third benefit is that you don’t have to spend extra bucks to purchase the high-definition maps.
And also another big benefit is that you actually lower a lot of the maintenance cost because it is by default of the XNGP to actually recognize any sort of road conditions. For example, there might always be construction going on the road that might not be indicated in high-definition maps, and you need to be able to adapt to that kind of changes. Now obviously, when you mention — what you mentioned in your question are the major challenges. Basically, it comes down to the capability to actually identify the environmental elements, including words and characters, signs and signals, pictures. And basically, all of the traffic, all of the vehicles and other kinds of road participants that are surrounding you and you have to actually make decisions based on that ahead of time.
And basically, that is my simple answer. I hope that helps. Thank you.
Paul Gong: [Foreign Language] So my following second question is still regarding the HD — regarding the autonomous driving, in this case, without HD map. Does that mean it leads to observe more environments and calculate more, think more, so that it requires even more calculation power on the chips? And regarding the cost down of autonomous driving hardware heading towards 2024 you mentioned you’re going to cut your cost by 50%. Does that mean you are going to use smarter software to reduce the requirement of the calculation power? Or should we maintain the calculation power even bigger calculation power?
