Doordash Inc’s AI Roadmap, Dot Commercialization, and EBITDA Guidance Clash in 2026 Earnings Calls

Date of Call: Aug 5, 2026

Business Commentary:

Strong Growth in Grocery Business:

• DoorDash reported that its grocery business is the fastest-growing part of its marketplace, with very healthy relationships with all partners on the platform.
• Improvements in unit economics across all categories, including grocery, were driven by increased adoption of the DashPass program and accelerating growth in the ads business.

Accelerating Performance in International Markets:

• Deliveroo's volume, MAU, and subscription growth were at their highest in the last couple of years, indicating accelerating performance.
• This growth was attributed to successful integration of lessons learned from building marketplace businesses globally and improvements in core propositions like selection, price, and delivery quality.

Improving Unit Economics and Profitability:

• DoorDash's adjusted EBITDA exceeded guidance, driven by strong performance in the core restaurants business and new verticals, with unit economics improving ahead of expectations.
• The improvements were due to better-than-expected performance in ads and subtotal, as well as Deliveroo's contribution profit positivity.

Advancements in Autonomous Delivery (DoorDash Dot):

• DoorDash achieved meaningful scale in its test market with DoorDash Dot, solving complex operational challenges in autonomous delivery.
• The success was due to the integration of operations and technology, allowing DoorDash to manage all aspects of autonomous delivery effectively.

Increased DashPass Subscription Growth:

• DashPass paid subscriber growth in the U.S. was one of the highest in the last two years, with more subscribers added in the last year than in the two prior years.
• This growth was driven by improvements in product quality, increased selection, and better consumer engagement and retention.

Sentiment Analysis:

Overall Tone: Positive

• Management highlighted 'extremely strong performance' and 'fastest-growing part of our marketplace business' for grocery, accelerating growth in Deliveroo, and 'very strong' restaurant growth. They noted 'improving unit economics across all categories' and 'strong demand trends' with MAUs at all-time highs. The tone was optimistic about long-term opportunities in autonomy, tech stack unification, and merchant services.

Q&A:

• Question from Michael Morton (MoffettNathanson): Concerns about grocers paying very low take rates and opportunity to reprice relationships.
  Response: Management emphasized strong, healthy partner relationships and multiple levers to improve economics, focusing on growth and investment rather than specific repricing.

• Question from Mark Mahaney (Evercore ISI): Asked about Deliveroo's accelerating growth and profitability improvements.
  Response: Growth attributed to successful integration of international operations and lessons from Wolt; unit economics are improving, with profitability targets expected to be met.

• Question from Nikhil Devnani (Bernstein): Asked about quality of international growth versus domestic and importance of market share.
  Response: International business concentrated in top 10 markets where DoorDash is a leader or strong number 2; growth and unit economics are improving, with significant long-term potential.

• Question from Deepak Mathivanan (Cantor Fitzgerald): Two-part question on AI spend attribution and drivers of U.S. restaurant GOV acceleration.
  Response: AI applied to customer experience (e.g., Ask DoorDash), merchant onboarding, and Dasher routing. Restaurant growth driven by increased DashPass subscribers, product improvements, and investment in selection/quality.

• Question from Dominic Ball (Rothschild & Co Redburn): Asked about progress and potential launch of merchant POS/Software products.
  Response: Vision to be the best partner to local businesses with an ecosystem including marketplace, digital tools, and in-store products; early proof points are strong with long runway.

• Question from Jason Helfstein (Oppenheimer): Asked about DoorDash Dot unit economics and AOV trends.
  Response: Dot represents scaled autonomous delivery milestone; cost profile shows exponential progress. AOV increasing as product improves and consumers use more categories.

• Question from Shweta Khajuria (Wolfe Research): Two-part question on DashMart fulfillment learnings and balancing EBITDA growth with investment.
  Response: DashMart fulfillment shows 10x better error rates and incremental demand; scaling based on positive early signs. Philosophy is to reinvest efficiency gains into high-ROI areas like autonomy and tech stack while driving growth and profit dollars.

• Question from Josh Beck (Raymond James): Multiple questions on DoorDash Dot characteristics, AI assistant impact, and ad monetization.
  Response: Dot operates in a representative market with real scale; ceiling depends on mastering operations and technology. AI assistant reduces friction, increases usage and orders. Ad monetization not a focus, as product improvements drive more orders.

• Question from Youssef Squali (Truist Securities): Asked about drivers of Q2 EBITDA beat and margin guidance for 2026.
  Response: Beat driven by unit economic improvements in ads, subtotal, and Deliveroo volume; focus remains on profit dollar production, not take rate optimization. Expect to land within Q3 EBITDA guidance range.

• Question from Ross Sandler (Barclays): Asked about AV expansion plans and examples of gaining subscriber penetration.
  Response: AV expansion to come after nailing execution in Phoenix; securing permits is ongoing. Subscription growth is strong globally, with share gains in U.K. and other markets, and improving order rates and unit economics.

• Question from Brian Nowak (Morgan Stanley): Asked for examples of early benefits from the new unified tech stack.
  Response: Seeing conversion wins from search improvements and automation in customer support; benefits stem from combining best-of-breed features across all markets into a new engine.

• Question from Ron Josey (Citigroup): Asked about strategies to add value to DashPass and reasoning for new delivery radius fees.
  Response: DashPass value increases as product improves (more selection, quality, affordability), driving adoption and engagement. New fees realign consumer payments with Dasher effort; impact not expected to be massive.

• Question from Justin Post (Bank of America): Asked about traffic from agentic partners and marketing cost implications.
  Response: Agentic traffic volume is currently low; focus is on mastering end-to-end physical commerce experience, with willingness to partner with digital assistants in the future.

• Question from Doug Anmuth (JP Morgan): Asked about drivers of net revenue margin pickup in Q2.
  Response: Increase due to lower Q2 Dasher costs; take rate expected to be flattish in Q3 and lower in Q4 due to higher seasonal Dasher costs.

• Question from Tom Champion (Piper Sandler): Asked about big picture view on AI spending ROI and impact on hiring.
  Response: AI used to build better products and drive efficiency across the company; efficiency gains are being reinvested into high-ROI areas like autonomy and tech stack, not primarily for cost cutting.

Contradiction Point 1

Maturity and Business Impact of AI

AI's role shifts from early testing to clear, modeled efficiency gains.

Deepak Mathivanan (Cantor Fitzgerald) - Deepak Mathivanan (Cantor Fitzgerald)

2026Q2: The company uses internal tools (DoorDash Bench) to model AI tool efficiency. - [Ravi Inukonda](CFO)

In which areas is DoorDash clearly attributing AI spending to business outcomes? - Shweta Khajuria (Wolfe Research)

2026Q1: The focus will remain on creating the best end-to-end shopping experience, which includes using agentic tools to improve discovery and search... The key is ensuring these tools enhance the entire customer journey. - [Tony Xu](CEO)

Contradiction Point 2

Autonomous Delivery (DoorDash Dot) Project Status

Progress shifts from early, exploratory phases to commercialization in test markets.

Jason Helfstein (Oppenheimer) - Jason Helfstein (Oppenheimer)

2026Q2: The technology is now commercialized in test markets. - [Tony Xu](CEO)

What are the current and long-term unit economics of DoorDash's autonomous delivery (Dot)? - Michael Morton (MoffettNathanson)

2026Q1: The vision is to build the most efficient autonomous delivery network... The company is early in its journey, scaling operations and manufacturing. - [Ravi Inukonda](CFO)

Contradiction Point 3

EBITDA Margin Outlook and Cost Management

Confidence in full-year EBITDA margin guidance contrasts with earlier uncertainty on cost offsets.

Youssef Squali (Truist Securities) - Youssef Squali (Truist Securities)

2026Q2: The beat flowed to the bottom line. For the rest of 2026, the company expects to land within the provided guidance range. - [Ravi Inukonda](CFO)

What drove the Q2 adjusted EBITDA beat versus guidance, and were investments delayed to later quarters? - Josh Beck (Raymond James)

2026Q1: The $50M impact is expected in both Q1 and Q2. The company is dynamically managing costs, finding offsets... and remains confident in its full-year EBITDA outlook. - [Ravi Inukonda](CFO)

Contradiction Point 4

Competitive Position and Growth in Europe

Assessment of competitive vulnerability changes from acknowledging risk to stating a stronger position.

Okay, so the user wants me to take the input "Nikhil Devnani (Bernstein)" and rewrite it into a single concise earnings-call question. Let me start by understanding the requirements here. The output has to be exactly one line, only the final question text, and it must end with a question mark if a rewrite is possible. If it's not a question or too vague, I have to output the input as-is.First, I need to figure out what the original input is. It looks like a person's name followed by a company or firm name in parentheses. Nikhil Devnani is likely an analyst or a representative from Bernstein, which I know is a financial services company. So this input is probably from an earnings call transcript where Nikhil Devnani from Bernstein is asking a question or making a statement. The task is to turn this into a question. Since the original input isn't a question, I need to infer what the question might be. In earnings calls, analysts often ask about financial performance, strategies, challenges, etc. So, given that Nikhil is from Bernstein, a possible question could be about the company's earnings, revenue, growth, or specific metrics discussed in the call. But the problem is the input doesn't provide any specific details about what was discussed. The name alone doesn't indicate the topic. So I have to make an educated guess. Maybe the original context was a question like "What are your thoughts on the recent earnings report?" or "Can you comment on the Q2 results?" However, without more context, it's hard to be precise. The user's instructions say if the input is too vague or lacks enough info, output as-is. Since the input is just a name and firm, it's very vague. But maybe there's a standard way to form a question here. Perhaps the question is asking for the analyst's question or comment. For example, "What was Nikhil Devnani's question for the company during the earnings call?" But that's more of a meta-question about the earnings call rather than the actual question posed by Nikhil.Alternatively, maybe the user expects a question directed at the company, as if Nikhil is asking something. But without knowing the topic, it's impossible to create a specific question. The safest bet, according to the rules, is to output the input as-is if it's too vague. However, the user might expect that since it's from an earnings call, the question is implied. Maybe the original input is part of a transcript where Nikhil is asking a question, and the task is to reconstruct that question. But without the actual question text, I can't accurately rephrase it. The input only gives the person's name and their affiliation. Therefore, following the rules strictly, since the input isn't a question and lacks enough information, the correct output would be the input as-is. However, perhaps the user expects that the presence of a name from a firm in parentheses implies that there was a question, and the task is to create a generic question based on that. Alternatively, maybe the original input is part of a list of participants in an earnings call, and the task is to form a question about their participation. But again, without more context, that's speculative. Given the strict rules, and the fact that the input is not a question and lacks necessary info, I should output it as-is. However, the example provided in the user's message might have a different approach. Let me check the example again.Wait, the user provided an example input and output? No, the user only provided the input here. So I have to rely on the instructions. The input is not a question, so if I can't form a clear question from it, output as-is. Therefore, the correct output is "Nikhil Devnani (Bernstein)" without any modification. But maybe the user expects that the question is implied, like "What was Nikhil Devnani's question?" but that's a meta-question. However, the user's instruction says if a rewrite is possible, end with a question mark. Since the input isn't a question, but maybe the rewrite is possible by forming a question about the input. But the input itself is not a question. The user's instruction says if the input is not a question, too vague, or lacks info, output as-is. Therefore, the answer should be "Nikhil Devnani (Bernstein)" exactly as input. But maybe the user wants to turn it into a question like "What is Nikhil Devnani's role at Bernstein?" but that's speculative. Since there's no info on his role, that's not possible. Therefore, following the rules strictly, the output should be the input as-is. However, maybe the user expects a different approach. Let me think again. In earnings calls, when an analyst is mentioned like this, they usually ask a question. So maybe the - Nikhil Devnani (Bernstein)

2026Q2: The competitive position is stronger, not weaker. Deliveroo is seeing its highest growth in years, and Wolt is gaining share. - [Tony Xu](CEO)

Does management perceive international growth as lower quality? - Youssef Squali (Truist)

2026Q1: Given the re-platforming, is DoorDash more vulnerable to competition in Europe (e.g., from Uber)? - [Youssef Squali](Truist)

Contradiction Point 5

Timeline for Grocery/New Verticals Profitability

Contradiction on when business will become gross profit positive.

Michael Morton (MoffettNathanson) - Michael Morton (MoffettNathanson)

2026Q2: DoorDash is on track to be gross profit positive in new verticals by late 2026. - [Ravi Inukonda](CFO)

"Can repricing relationships with grocers who have low take rates shift affordability burdens back to them?" - Michael Morton (MoffettNathanson)

20260219-2025 Q4: Expect business to be gross profit positive in second half of 2026. - [Ravi Inukonda](CFO)