Bookings $1.44B (+51%) Cleared the Street by 16%, DAUs Cross 100M to 111.8M, FY Bookings Guide Raised ~$600M — But the Stock Has Run 37% in a Month and CFO/CPO Transitions Land at Peak Operating Moment

Results vs. Consensus

Q2 2025 Scorecard

Year-over-Year Comparison

Quarter-over-Quarter Comparison

Revenue assessment. Revenue of $1,080.7M (+21% YoY) came in ~2% below the ~$1.10B consensus while bookings beat by 16%. Both facts trace to the deferred-revenue model: bookings convert to revenue over the estimated average paying-user lifetime, which was 27 months in Q2. The vast majority of the Q2 bookings outperformance is therefore deferred and will be recognized over the next several quarters. Two comp effects also matter: the YoY growth rate laps a $58.9M revenue acceleration recorded in Q2 2024 from the paying-user-life estimate change, and Q1 2025 revenue grew +29%, so the +21% print reads as deceleration only because of that comp. The raised FY2025 revenue guide of +22-25% is well-supported by the bookings trajectory.

Margins assessment. Adjusted EBITDA of $18.4M fell 72% YoY (from $66.5M), and the FY guide was re-based from $205-265M to $(5)-55M. The mechanism is timing, not deterioration: Adjusted EBITDA excludes the net increase in deferrals, which swelled to $301.5M in Q2 versus $47.9M a year ago, while the operating costs attached to the incremental bookings (DevEx +52%, cloud-burst infrastructure, personnel) hit the P&L immediately. Management's own framing is that consolidated net loss increased and Adjusted EBITDA decreased even as cash generation accelerated. Cash flow is the cleaner scoreboard: operating cash flow $199.3M (+32% YoY; $229.3M and +51% excluding a $30M developer payout delayed from Q1), and free cash flow $176.7M (+58% YoY; $206.7M and +85% ex-delay). The letter also flags roughly 1,000bp of YoY operating-margin improvement on the key fixed-cost lines, with the bare-metal/cloud-burst architecture (described by Baszucki in Q&A) absorbing 30M+ concurrent users without 24/7 peak provisioning.

EPS assessment. EPS of $(0.41) missed the $(0.36) consensus and widened from $(0.32) a year ago; consolidated net loss was $279.8M versus $207.2M. The composition matters: stock-based compensation of $284.8M is the largest single add-back in the Adjusted EBITDA reconciliation and is non-cash, and the deferral mismatch described above pushes the recognized-revenue P&L below the cash economics. The +32% OCF and +58% FCF growth are the more informative read on operating progress. The GAAP loss should narrow as the bookings-to-revenue lag converts, but on the company's own updated guide the FY2025 consolidated net loss widens to $(1,261)M-$(1,201)M precisely because bookings are being raised.

Segment & Geographic Performance

RBLX reports as a single segment but provides regional and demographic disclosures that are critical to the platform thesis. Q2 2025 was an exceptionally strong regional quarter, with material acceleration in APAC alongside continued double-digit US&C growth.

Regional Growth (YoY, per the call)

US & Canada

US&C bookings grew +43% YoY in Q2 2025, a meaningful acceleration from the +31% disclosed in the Q1 shareholder letter. This is RBLX's most mature region and the comparison base is higher than any other geography, making the acceleration particularly noteworthy. US&C DAU growth of +21% YoY is steady-state, and the letter has US&C ABPDAU up +18% YoY: the region's growth is increasingly monetization-led, with the over-13 cohort contributing disproportionately.

Assessment: US&C is no longer a "mature, slow-growing" region for Roblox. It is reaccelerating with the multi-hit ecosystem and the aging-up demographic mix shift, taking share within an attention-saturated US&C gaming/entertainment market. That validates the long-term TAM expansion thesis.

APAC

The APAC story is the structural growth engine of the next 2-3 years. Bookings +75% YoY reflects genuine product-market-fit deepening across the region, not promotional or pricing actions. The country call-outs are extraordinary: Indonesia +150%+, Korea +120%+, Philippines ~100%, India +90%+, Japan +50%+. APAC DAU growth of +76% roughly matches APAC bookings growth, indicating user-base expansion and monetization are compounding together, with international ABPDAU up +11% YoY per the letter.

Assessment: APAC is the highest-confidence component of the multi-year bookings growth bridge. The regional infrastructure investments, the auto-translation quality improvements (which Baszucki specifically called out for Japan), and localized payment options have removed historic friction points. We model APAC contributing 35-40% of incremental global bookings growth in 2026-2027.

Demographic Breakdown

Assessment. The over-13 cohort is growing more than twice as fast as the under-13 cohort on a DAU basis and roughly twice as fast on hours. This aging-up trajectory has been a stated strategic priority for 18+ months and is now showing in the demographic mix: over-13 share of total DAUs is up from ~58% a year ago (derived) to 64% today. Older users carry more discretionary spending power, so the demographic mix shift is a structural bookings tailwind. Baszucki's Dress To Impress and Grow a Garden references (experiences resonating with over-17 users, "played in class in college") are the anecdotal layer on top of the disclosed cohort numbers.

Key Topics & Management Commentary

Overall Management Tone: High conviction in the strategic vision and the operating execution, paired with a notable strain of conservatism on the financial trajectory. Baszucki was triumphant on the platform metrics and the multi-hit ecosystem proof points; Chopra's first quarter as CFO was characterized by explicit, repeated guidance hedging: Q3 "normalization" assumption, Q4 "heightened uncertainty" flag, viral-hit "too early to extrapolate" framing. The CEO/CFO tone divergence is the most informative read of the call: management itself does not fully believe Q2's pace can sustain.

1. The Q2 Print's Three Defining Numbers: $1.44B Bookings, 111.8M DAUs, ~$600M Guide Raise

Q2 2025 broke records on every dimension that materially matters to the investment case. Bookings of $1,437.6M grew +51% YoY against a ~$1,240M Street consensus, a +16% beat. DAUs of 111.8M crossed the 100-million threshold for the first time ever, up +41% YoY, with an absolute Q2-over-Q1 add of ~14M DAUs. Hours engaged of 27.4B grew +58% YoY, materially faster than DAU growth, indicating that per-user engagement is also expanding: the platform is not just adding users but pulling more time from existing users.

"In Q2 2025, we had revenue of $1.1 billion. This was up 21% year-on-year. Our Q2 bookings were $1.4 billion, up 51% year-on-year. And bookings growth was strong across all regions. U.S. and Canada saw bookings growth of 43%, APAC grew 75%…"
— David Baszucki, CEO and Co-Founder

The ~$600M raise to the FY2025 bookings guide midpoint ($5,322.5M at the Q1 guide to $5,920M now), taking the full-year growth range from +21-23% to +34-37%, is the most consequential disclosure of the print. Management notes the new range compares with an initial FY2025 guide of +19-21% growth: the year's expected growth rate has roughly doubled in two quarters.

Assessment. These three numbers individually each justify a multi-quarter operating thesis upgrade; collectively they represent the strongest single-quarter operating print in the platform's history. The challenge is that the stock has already priced much of this strength: from $100.65 on July 1 to the $137.79 reaction close on July 31 is a +37% move in one month. The thesis is intact and arguably strengthened; the entry asymmetry is what has changed.

2. Grow a Garden as Proof-of-Concept for Algorithmic Discovery + Studio Tooling

Grow a Garden is not just a viral hit. It is the operational proof that RBLX's discovery algorithm + studio tooling + creator economy stack can systematically produce viral hits at scale. The experience passed 20M concurrent users in July, meaningfully above what Baszucki described as the prior Guinness World Record peak-concurrent of ~14-15M for any single game. What matters more than the peak is the speed: launched March 25, 2025, it became the #1 experience by spending within one month of its creation, which the shareholder letter calls faster than any other experience in Roblox history, and hit its concurrency records within its first 100 days. Baszucki was explicit that this trajectory is reproducible because it emerged from algorithmic discovery driven by long-term-value optimization, not from one-time promotional pushes or external paid acquisition.

"Well, the discovery that occurred with Grow a Garden after that was launched with our recommendation system, should be intrinsically repeatable, in that it was algorithmic, and it was based on the things we use to estimate the long-term value of that property for each individual user, not necessarily in the short term, but over the next year."
— David Baszucki, CEO and Co-Founder

The letter credits the surrounding stack: AI-driven discovery plus curated sorts ("Today's Picks") carried the title to #1, and the Cube 3D generative foundation model is now producing 3D assets at scale (nearly one million generated in experiences and Studio since March, with 7-8x faster object generation). This is the platform's compounding advantage: every discovery and tooling improvement compresses the next viral hit's launch timeline.

Assessment. Grow a Garden's significance is not the title itself but what it proves. The combination of (a) algorithmic discovery optimized for 1-year-forward value rather than short-term engagement, (b) studio tooling that compresses creator iteration cycles, and (c) a creator economy that rewards organic-traffic generation: this stack is now empirically validated as a viral-hit production engine. The Street's "could RBLX repeat Adopt Me, Brookhaven, Blox Fruits?" question has been answered: yes, and faster than before.

3. Multi-Hit Ecosystem Disproves the Concentration Risk Bear Narrative

The most consequential structural disclosure of the quarter is that RBLX now has 5 experiences with >10M DAUs simultaneously: Grow a Garden, Steal a Brainrot, Brookhaven, 99 Nights in the Forest, and Ink Game. Four of these five experiences launched within the past 12 months. Historically, the RBLX platform thesis carried a binding "concentration risk" critique: the top 1-2 experiences (Adopt Me, Brookhaven) generated a disproportionate share of bookings, and a fade in their engagement could destabilize the bookings trajectory. The Q2 2025 disclosure dismantles that thesis.

"Our strength in Q2 was broad-based across the platform, but we do want to share the breadth of new viral hits that have showed up on the platform. Into July, we currently have 5 experiences with over 10 million daily active users, including hits such as Grow a Garden, Steal a Brainrot, Brookhaven, 99 Nights in the Forest and Ink Game. Four out of 5 of these experiences have been launched in the last 12 months."
— David Baszucki, CEO and Co-Founder

The data point Chopra emphasized further: in-experience hours for experiences ranked #11 and below grew +47% YoY, and "more than half" of the growth in experience spending came from non-top-10 titles. The letter adds two more cuts: spending among experiences ranked 11-50 grew +90% YoY, and platform-wide in-experience spending still grew +36% YoY even excluding every dollar spent in Grow a Garden. The long tail is healthy, deep, and inflecting.

Assessment. The concentration-risk bear thesis is structurally retired by this disclosure. With 5 experiences over 10M DAUs (4 of them new), the bookings trajectory is now diversified across multiple viral hits and a measurably growing long tail. This thesis disconfirmation is worth 1-2 multiple turns on the structural valuation framework alone: the discount RBLX has historically traded at versus large-platform peers (META, GOOGL, NFLX on certain frameworks) reflected this concentration discount, and it should compress.

4. Non-Top-10 Engagement +47%; Platform Spending +36% Even Ex-Grow a Garden: Long-Tail Health

Naveen Chopra used his first-quarter-as-CFO airtime specifically to surface the non-top-10 data point, which is the single most underappreciated number in the quarter. In-experience hours for experiences ranked #11 and below grew +47% YoY. This means platform engagement is broadening at a structural level, not just deepening within the top hits. The companion spending cuts from the letter: experiences ranked 11-50 grew spending +90% YoY, and platform-wide in-experience spending grew +36% YoY even with every Grow a Garden dollar excluded. Those are the cleanest reads available on the economic engine beneath the viral layer.

"For experiences ranked #11 and below, growth of in-experience hours was 47% year-on-year. And that engagement is also powering broad-based spending growth. In fact, more than half the growth in experience spend came from non-top 10 titles."
— Naveen Chopra, CFO

Roblox's economic data point: the top 1,000 creators averaged roughly $1M each in trailing 12-month earnings; the top 10,000 averaged more than $110K; and 18 creators earned over $10M in the trailing 12 months, on DevEx of $316.4M in Q2 alone (+52% YoY). The creator economy is now a genuine professional career path at meaningful scale.

Assessment. The long-tail health is the structural defense against any individual viral hit fade. If Grow a Garden tapers in Q3-Q4 (which Chopra's guidance conservatism explicitly assumes), the +36% ex-hit spending growth alone is sufficient to deliver strong double-digit aggregate bookings growth. This is the underlying-platform-growth-rate data point that should anchor 2026 modeling assumptions.

5. APAC as the Structural Growth Engine: Indonesia +150%, Korea +120%

APAC bookings growth of +75% YoY is the standout regional read, but the country-level granularity is more informative: Indonesia +150%+, Korea +120%+, Philippines ~100%, India +90%+, Japan +50%+. These are not purely small-base-effect figures: APAC consolidated DAU growth of +76% YoY indicates the user base is expanding aggressively, with bookings growth tracking alongside it. Baszucki framed the drivers as several things working together, with translation product quality specifically called out as the unlock in Japan, on top of the regional infrastructure and payments work.

"In Japan, especially, we had a bit of an unlock as we got to ultimately product quality on translation that I think supports the growth in that region. So even in APAC, my belief would be it's really a bunch of things working together."
— David Baszucki, CEO and Co-Founder

The mix shift should continue as the country-level user bases compound. The letter's differential-pricing and regional-pricing initiatives (Regional Pricing for game passes launched in April, avatar items in June) are specifically designed to convert the APAC engagement wave into payers at locally-optimized price points, and the letter credits them with supporting creator economics and notes strong early adoption.

Assessment. APAC is the structurally most-defensible growth pillar of the next 3-year trajectory. The investments to drive APAC are largely sunk (regional infrastructure, auto-translation engineering, localized payment rails); the compounding effect is now harvestable. Modeling APAC at 35-40% of incremental global bookings growth through 2027 is the appropriate base case.

6. Aging-Up: Over-13 DAUs +54%, Now 64% of Total

Aging-up has been a stated strategic priority for 18+ months. Q2 2025 confirms the trajectory is real and accelerating. Over-13 DAUs grew +54% YoY, and over-13 hours grew +72% YoY, both materially faster than total DAU (+41%) and total hours (+58%) growth. The over-13 cohort now represents 64% of total DAUs (up from ~58% a year ago) and 66% of total hours.

"Yes. I would feel there would be a nuance, in that our vision is that we're creating a platform for all ages around the world… And I feel we're not just trying to age up. We're seeing it in the numbers as an increasing percentage of our users are over 13, and we continue to show solid growth amongst that. Starting even a year ago with Dress To Impress, we've started to see experiences that are really big for users over 13. Even if over 17, Dress To Impress last summer was being played in class in college. And I think the same thing has happened in Grow a Garden, in that it's a -- it highlights a universal experience that's appealing to all ages."
— David Baszucki, CEO and Co-Founder

The structural relevance: over-13 users have meaningfully higher discretionary spending power and the over-17 cohort within over-13 monetizes at the highest rate on the platform. As the over-13 share of total DAUs continues to expand (we model 68-70% by year-end 2026 at current trajectory), the platform-wide average bookings per user has a structural tailwind from mix shift alone, separate from organic engagement deepening.

Assessment. The aging-up thesis is operationally validated. Dress To Impress and Grow a Garden are both proof points that breakthrough experiences can resonate across age cohorts, with the over-13/over-17 cohorts driving the highest absolute monetization contribution. The narrative shift from "kids platform" to "all-ages platform" is now data-supported, not aspirational.

7. Creator Economy at Scale: $316M DevEx, 18 Creators Earning $10M+

DevEx in Q2 was $316.4M, up +52% YoY, a new quarterly record and nearly double the level of two years ago. The letter notes developers have earned more than $1B over the trailing twelve months. The top 1,000 creators averaged roughly $1M in trailing 12-month earnings; the top 10,000 averaged $110K+. Most notably: 18 creators earned over $10M in the trailing 12 months.

"And finally, our economy, which we're really building to help creators thrive and earn, for example, in the last 12 months, 18 creators earned over $10 million on the platform. There's a lot of momentum in new content in our ecosystem for really up and down the size of creators."
— David Baszucki, CEO and Co-Founder

The new Creator Rewards Program restructures the incentive system to reward creators for driving organic traffic to their experiences rather than for raw engagement time. This realigns creator incentives toward sustainable, long-term-value-creating behaviors (viral content, social acquisition, off-platform marketing) and away from short-term engagement-gaming tactics (artificial padding, repetitive grind loops).

Assessment. The creator economy is now at a scale where Roblox is a meaningful career path for thousands of independent developers globally, not just a hobbyist platform. This is the supply-side network effect: more economic opportunity attracts more creator talent, which produces more high-quality content, which attracts more users, which generates more bookings, which funds more DevEx. The flywheel is operating at scale; the question is acceleration trajectory rather than directional viability.

8. CFO Transition: Mike Guthrie → Naveen Chopra at Peak Operating Moment

This was Naveen Chopra's first earnings call as CFO: appointed June 9, 2025 and effective June 30, he succeeds Mike Guthrie, whom Baszucki opened the call thanking for "such an incredible tenure at Roblox." Chopra arrives from Paramount Global, where he was EVP & CFO through the company's legacy-media-to-streaming transition, with earlier finance leadership roles at Amazon, Pandora, and TiVo. It is a credentialed succession with a strong capital-discipline reputation. But the transition timing is conspicuous: it lands at the strongest operating moment in RBLX's public history, with no overlap on a full earnings cycle, and Guthrie's relationship with the buy-side was deeply established.

"Look, I came to Roblox not just because of the amazing business that exists today, but really because of where I think it can go in the future. Today, I think, as most of you know, Roblox is a highly scaled consumer platform with massive engagement and what I think of as an enviable financial model… And I particularly love the fact that it's a team that embraces a long-term orientation and intentionally chooses to tackle really hard problems."
— Naveen Chopra, CFO

Chopra's introduction was tonally well-paced: balanced enthusiasm for the operating model with explicit framing of his thesis (UGC tailwinds, AI-driven creator productivity, market headroom, financial scale optionality). He flagged Q4 conservatism multiple times across his prepared remarks and Q&A responses, which is appropriate for a first-quarter CFO but is also a meaningful tonal departure from Guthrie's more confident guidance-setting posture.

Assessment. Chopra's profile fits RBLX's next chapter: Paramount's CFO seat through a structural business-model transition, plus platform-scale finance experience at Amazon, maps well onto a UGC platform mid-inflection. But the transition risk is real: no overlap on a full earnings cycle, no track record yet with the RBLX investor base, and the timing at peak operating moment means his first 2-3 quarters will be heavily scrutinized for any signs of guidance miscalibration. This is the binding execution risk vector for the next 6 months and a primary reason we are initiating at Hold rather than Outperform.

9. CPO Manuel Bronstein's Departure With No Announced Successor: Governance Overhang

Buried in the prepared remarks: Manuel Bronstein, RBLX's Chief Product Officer since 2021, plans to depart effective September 30, 2025. No successor has been announced. Bronstein has been a key product architect across the last 4+ years, including the discovery algorithm redesign, the studio tooling improvements that enabled the Grow a Garden launch trajectory, and the safety platform innovations.

"In our press release, we shared that Manuel Bronstein has decided to move on from Roblox. I just want to share that Manuel has been at Roblox for more than 4 years, and he's been a key part of our growth and maturation. He's helped us build incredible teams across product and partnerships. He plans to take a break before pursuing personal and entrepreneurial activities, but he'll stay with us through the end of September."
— David Baszucki, CEO and Co-Founder

The lack of an announced successor (with Bronstein's last day 60 days away) is a governance flag. Either (a) a successor has been identified but is not yet ready to be announced (most likely an internal promotion still in negotiations), or (b) a search is underway and the company is comfortable with an interim period. Either way, the CPO seat is critical during a moment when the product roadmap (discovery, safety, studio tooling, AI integration) is the binding driver of the multi-year thesis.

Assessment. The CPO departure layered on top of the CFO transition means two C-suite seats are in motion at the platform's peak operating moment. If a credible internal successor is announced within 30-60 days, this overhang dissipates. If a search drags into Q4 with no announcement, the governance discount expands. We model this as a contained, time-bound risk: meaningful enough to support the Hold initiation, not severe enough to support Underperform.

10. Chopra's Q3-Q4 Conservatism Flag: The Most Informative Tonal Signal of the Call

Despite Q2's extraordinary print and the ~$600M FY bookings guide raise, Naveen Chopra was explicit, repeated, and unmistakable in his framing of forward conservatism. Three distinct conservatism flags were embedded in his prepared remarks and Q&A: (a) Q3 guidance conservatively assumes "normalization" of recent viral-hit engagement toward underlying trends, (b) tougher YoY comps in August/September as prior viral hits peaked in 2024, (c) Q4 guidance carries "heightened uncertainty" given concentration of Q4 bookings in the final days/weeks.

"First, although momentum has continued in July, for guidance purposes, we conservatively assume that engagement and spending in recent viral hits will move toward underlying growth trends during Q3. Second, we note the fact that year-over-year growth rates will face tougher comps in August and September as prior viral hits were at their peak during those months in 2024. And then if you look at Q4, we believe that some additional conservatism is warranted. To be clear, everything we're seeing is positive and healthy, but it's just too early to extrapolate Q2's extraordinary trends over a prolonged period of time."
— Naveen Chopra, CFO

For a CFO in his first earnings call, this level of explicit forward hedging is unusual. It can be read two ways: (1) Chopra is establishing a beat-and-raise pattern by sandbagging his first quarter's guidance, or (2) Chopra is genuinely concerned that the Q2 viral-hit dynamics will not sustain and wants to give himself room to deliver in Q3-Q4 without further raises.

Assessment. We read this as ~60% sandbagging / ~40% genuine caution. The conservatism is well-calibrated for a new CFO: even if Q3 prints within the guidance corridor without further raise, the stock can hold; if Q3 modestly beats and the FY remains in the corridor, the stock has room to add another 10-15%. But the bear case (that Grow a Garden engagement tapers materially in Q3 and the FY guide proves out at the low end) is also entirely consistent with Chopra's language. The conservatism flag is the strongest signal that even management does not fully believe Q2's pace sustains. This is a material factor in the Hold (CB) initiation.

11. Capital Position: $3.7B Net Liquidity, Bare-Metal+Cloud Burst Architecture, $176.7M Q2 FCF

The balance sheet carries $4.7B of cash and investments and $3.7B of net liquidity (net of debt). Q2 generated $176.7M in free cash flow (+58% YoY) on $199.3M of operating cash flow, and both figures absorb a $30M developer payout delayed from Q1; ex-delay, FCF grew +85%. Trailing 12-month FCF is approaching $1.0B for the first time. This capital position is at a scale where strategic optionality is meaningful: M&A, buyback capacity against SBC dilution, or product/talent acquisitions are all feasible without straining the balance sheet. Baszucki specifically called out the infrastructure architecture: bare-metal owned servers as the baseline with cloud burst for 4-8 hour peak windows.

"Behind the scenes, the optimal mix for spending is both our bare metal infrastructure with burst mode into cloud providers. And what you're seeing now behind the scenes is somewhat of a really great mix where for 4 to 8 hours, we spin up a bunch of cloud servers and mix it with the constant bare metal of our own infrastructure. So that's worked really well and allowed us to absorb this without paying the CapEx 7 days a week, 24/7 to handle that type of a load."
— David Baszucki, CEO and Co-Founder

The architectural sophistication matters: at 30M+ concurrent peak load, RBLX is operating closer to a hyperscaler than to a traditional game publisher. The bare-metal/cloud-burst mix optimizes the cost-of-compute trade-off: cloud peak pricing only during peak windows, owned-infrastructure economics during steady-state. This is what keeps cash generation compounding alongside the bookings growth even while the recognized-revenue P&L absorbs the deferral mismatch.

Assessment. The capital position de-risks the multi-year investment thesis on several dimensions: (a) ability to fund continued platform investment without near-term cash constraints, (b) ability to absorb a 1-2 quarter macro-driven bookings softness without distress, (c) optionality on inorganic growth, and (d) buyback support for share-count dilution from SBC. This is structurally supportive of the long-term Outperform case even if the near-term entry is suboptimal.

Guidance & Outlook

Read the two halves of this table together. Bookings, operating cash flow, and free cash flow were all raised substantially; the net-loss guide widened and the Adjusted EBITDA guide was cut from $205-265M to $(5)-55M in the same release. Management's footnote is explicit that both moves are a consequence of the bookings raise itself: the operating expenses supporting the incremental bookings are recognized immediately while the vast majority of the associated revenue is deferred over the 27-month paying-user life. This is a timing mismatch, not a demand or cost problem, and the raised cash-flow guides are the tell.

Q3 shape. The explicit Q3 bookings guide of $1,590-1,640M implies +41-45% YoY, a deceleration from Q2's +51%. This is consistent with Chopra's "normalization" language: the guide conservatively assumes the recent viral-hit engagement moves toward underlying growth trends during Q3, with tougher August-September comps against 2024's viral-hit peaks.

Guidance style: Chopra explicitly framed the guide as conservative, and flagged Q4 with additional caution given lower visibility. With the ~$600M raise being the largest in company history, calling the implied Q3-Q4 trajectory "conservative" is meaningful: it suggests management's internal base case sits at or above the high end, with the low end serving as a hard floor. The beat-and-raise pattern likely continues if Q2's underlying dynamics sustain through August-September.

Analyst Q&A Highlights

Infrastructure Capacity for 30M+ Concurrent Users: Bare-Metal vs. Cloud Burst

The opening question pressed management on how the infrastructure has absorbed the surge from Grow a Garden's 20M+ peak concurrent load, a scale that exceeds the prior Guinness World Record for any single game. The question framed prior CapEx commentary about platform resiliency and asked whether further capacity investments would be required to sustain this scale. Baszucki's answer leveraged the bare-metal/cloud-burst architecture as the structural cost-control mechanism: owned infrastructure for steady-state, cloud burst for 4-8 hour peak windows, no 24/7 over-provisioning.

Q: "Maybe if you could just give us a little color about how much capacity the system has to sort of absorb the surge in demand that we've seen? Because I know you guys are very focused on the long term. But a couple of years back, I think you sort of increased your CapEx to increase the capacity and resiliency of the platform. Can you just give a little bit of color on how you're thinking about sort of the CapEx and the platform itself?"
— Jason Bazinet, Citi

A: "Great question. I want to highlight, we are pushing numbers above 30 million concurrent players at the same time and Grow a Garden in the past few weeks has gone over 20 million concurrent. Prior to that, I believe, in the Guinness Book of World Records, the peak concurrent of any game was in the 14 million to 15 million. So we're pushing some really big numbers. Behind the scenes, the optimal mix for spending is both our bare metal infrastructure with burst mode into cloud providers. And what you're seeing now behind the scenes is somewhat of a really great mix where for 4 to 8 hours, we spin up a bunch of cloud servers and mix it with the constant bare metal of our own infrastructure. So that's worked really well and allowed us to absorb this without paying the CapEx 7 days a week, 24/7 to handle that type of a load."
— David Baszucki, CEO and Co-Founder

Assessment: The bare-metal/cloud-burst architecture is the operational reason RBLX can grow bookings +51% YoY while holding the cash cost curve: the platform is not over-provisioning to handle peak load; it is dynamically scaling. This is structurally analogous to how mature hyperscalers (AWS, GCP, Azure) operate at the infrastructure layer — RBLX has built a similar elastic-compute capability for its specific platform use case, which is the cost-curve advantage that enables the operating leverage.

Platform Vision: Implications for the 2026+ Growth Trajectory

A direct question on whether the current platform improvements translate to a structurally faster run-rate of growth in 2026 and beyond. The framing was sophisticated: acknowledging that visibility is limited to the next 2 quarters but pressing for the directional logic on whether developer investment + discovery improvements compound. Baszucki's response framed the technology stack as still mid-build: the vision of a single-build experience that runs across all device tiers and all languages is not yet fully delivered, meaning the platform compounding has further room to run.

Q: "It's hard for me to imagine that, I guess, professional developers aren't taking note of the changes to the platform, the earnings opportunity and with more of that being spread to non-top 10 developers… how does this maybe not translate a faster run rate of growth for Roblox in sort of 2026 and beyond? I think we've seen the proof points for how developer investment and sort of the discovery engine translates to bookings growth. So I'm wondering, I guess, where there could be sort of incremental friction here?"
— William Lampen, BTIG

A: "We're on the track that we started really annotating 6 quarters ago, which is raw perf, scale, tech, discovery, economy, live ops. We still believe we have a lot of technology in the pipeline to fully realize our vision of what it's going to take… The vision on Roblox is ultimately to support a single build from a creator that runs in any language on any device and scales all the way from a low-end 2-gigabyte Android phone maybe in a difficult networking environment, all the way to an experience that looks great on a gaming PC and starts to go higher in res, maybe with avatars that -- our avatars that don't look like Roblox avatars and look like a wide range of that. So we're not quite there yet. So we do want to focus on that technology. We think over time that's going to accelerate the genre expansion that we're talking about."
— David Baszucki, CEO and Co-Founder

Assessment: Baszucki's "we're not quite there yet" is the most important strategic signal. The implication is that the current Q2 print — the strongest in company history — was achieved with the technology stack only partially delivered. The "single build across all device tiers" vision (Android low-end through gaming PC high-end with avatar fidelity scaling) implies further engagement-deepening across the over-13/over-17 cohort that monetizes at the highest tiers. This supports a 2026 bookings growth trajectory that does not need to decelerate sharply from the 2025 +34-37% pace.

Grow a Garden as a Developer Launch Pad & Top-10 Rotation Drivers

A two-part question: (a) Is Grow a Garden's success serving as a launch pad for newer prospective developers? and (b) What is driving the unusually high rotation at the top of the platform, where 4 of the 5 experiences above 10M DAUs launched within the past 12 months? Baszucki's response confirmed both dynamics: developer awareness of the platform's viral potential has materially increased ("the word is out"), and the algorithmic discovery system is the structural driver of the rotation rather than promotional campaigns.

Q: "Are experiences like Grow a Garden serving as a sort of a promotional launch pad for newer prospective developers? So maybe could you just give us an update on developer growth and what you're hearing from the community? And then I just want to look past Grow a Garden a little bit on hone in on the other experiences that are gaining real traction. Could you maybe talk about why you're seeing such a high rotation in the top 10 experiences? Is it search and discovery? Is it sort of the halo effect from Grow a Garden? Is there something else? Maybe talk about the durability of some of those dynamics."
— Benjamin Black, Deutsche Bank

A: "I do feel the word is out, in that most creators in the gaming space when one experience goes over 20 million concurrent, they take note of that. So I do think the word is out. I want to highlight Grow a Garden shows that when the conditions are right for growth, metaphorically Grow a Garden is that, we see interesting things… On the other experiences on the platform, we're seeing continued really interesting stuff. Once again, 5 experiences with more than 10 million DAU. But what Naveen shared that I think is really interesting is the growth all the way down to 1,000 and the breadth of what we have coming in the pipeline. We think there's a lot of headroom as we look at genres that we think should someday be really bigger on the platform, sports, racing, battle, RPG… The other thing I'll note is, without giving any forward predictions, we want to move as much money as we can from the top line to the creator community."
— David Baszucki, CEO and Co-Founder

Assessment: The "word is out" framing is operationally meaningful — it signals that the supply-side network effect (creator influx) is now self-reinforcing. The genre expansion roadmap (sports, racing, battle, RPG) addresses the historical gap that prevented older-cohort engagement parity with traditional gaming platforms. If these genre expansions deliver in 2026, the over-17 cohort growth could compound further from the current trajectory.

Long-Term Monetization Mix: Commerce vs. Ads vs. Subscription

A strategic question on how the monetization mix evolves as the platform matures: how should investors think about the balance between commerce (physical and virtual goods), advertising (Rewarded Video, sponsored tiles), and subscription as growth vectors? Baszucki's response framed the answer as a platform-provided menu: not RBLX picking winners among monetization models, but RBLX providing the full toolkit and letting individual creators optimize for their experiences. Chopra's add-on emphasized the flywheel dynamic: engagement growth amplifies all monetization vectors simultaneously.

Q: "When you think about longer term monetization of the platform and maybe for a mixture of Dave and Naveen, how do you think about the building blocks when you move from sort of product by product to more of this platform approach and you're building this developer ecosystem of how you're thinking about the building blocks for commerce as opposed to advertising and maybe the longer term for subscription aspects to the business model? And how should we think about mix of monetization evolving as the platform does?"
— Eric Sheridan, Goldman Sachs

A: "We really are building an ecosystem and a platform here, and we see a market where individual creators have a range of monetization vehicles to choose from and optimize for their own experiences. As we roll out Rewarded Video ads, we believe some developers will focus very heavily on those, whereas some creators will stay with a virtual economy type model. So we really want to have the best range of tools, both freemium, virtual currency, paid access gaming, ad-supported, and have an ecosystem where creators can do a pickup on that. And that ultimately includes physical shopping as well where I believe some of our creators are making most of their money now by selling physical items… I will highlight there are certain genres out there amongst older users that naturally monetize much high -- more highly than Roblox does. And so the expectation could be that as we start to see content in those genres amongst older users, with the market supporting that, we may and should see higher monetization within some of those."
— David Baszucki, CEO and Co-Founder

Assessment: The "platform-provided menu" framing is strategically correct — RBLX should not pick winners among monetization models for its creators. The implicit ARPU expansion thesis is genuine: certain mature gaming genres (RPGs, battle shooters) monetize at $50-100 per active user annually on traditional platforms versus RBLX's current ~$50 blended. As over-13/over-17 cohort growth in those genres compounds, blended ARPU has structural room to expand — potentially toward $65-80 over 3-5 years if the genre expansion roadmap delivers.

Google Ad Partnership: Rewarded Video Rollout & Third-Party Demand

A specific question on the state of the Google ad partnership: where is Rewarded Video in its rollout, what's the prioritization within the broader product roadmap, and is RBLX planning to bring additional third-party demand sources beyond Google? Baszucki confirmed the eligibility window has expanded with strong creator pickup and performance metrics on both iOS and Android. Chopra added that additional ad partners are in the medium-term plan, with the Google partnership demonstrating the integration template.

Q: "Maybe we could just follow up on Eric's question here and maybe a little finer tuned. Could you talk about where you are with the Google partnership on the advertising side, the implementation time line and how we should think about the opportunities for kind of ad monetization and maybe where that is on the priorities… And then within that, are you -- how much are you thinking bringing other sources of third-party demand on the ad side?"
— Ken Gawrelski, Wells Fargo

A: "We did open the window of which creators can now offer Rewarded Video. We have been very fastidious in watching the performance of that product on both iOS and Android, making sure even on low memory devices that we offer the performance and scale that we're used to, and we've been doing that relentlessly and we're now live. We've -- I've seen the internal results of the growth rate of the pickup of Rewarded Videos and there is strong and continued pickup of it. We're not splitting out the size of that, but there is, once again, a wide range of creators who this tends to line up for. And we have a lot of creators now starting to experiment with this. I'm really excited about the opportunity to grow that. So stay tuned. We hope, as we start to imagine being 10% of the global gaming content market running through our platform, this will naturally form a monetization component for a chunk of developers that have experiences that make sense for this."
— David Baszucki, CEO and Co-Founder

Assessment: Rewarded Video is not yet a material bookings contributor in Q2 but the structural setup is in place. With "almost 100 publishers" onboarded and the eligibility window now expanded, the H2 2025 Rewarded Video pickup curve will determine whether ad-supported monetization becomes a meaningful incremental contributor in 2026. Modeling Rewarded Video at 3-5% of incremental 2026 bookings is the appropriate base case; upside if multi-DSP integration accelerates beyond Google-only.

Creator Rewards Program Incentive Shift & Sponsored Tiles

A two-part question: (a) How does the new Creator Rewards Program change incentives for creators versus the old engagement-based payouts? (b) When does RBLX make a more aggressive push into sponsored tiles within the discovery surface? Both responses were strategically clarifying. Baszucki framed the Creator Rewards shift as an alignment of incentives toward long-term platform health (rewarding traffic generation rather than raw time-on-platform), and characterized sponsored tiles as a "big future" with discovery-optimization work continuing.

Q: "Maybe just on the Creator Rewards Program that you talked about in the shareholder letter. How does that change incentives for creators versus the old engagement-based payouts? And should that drive any near-term meaningful uptick in the amount of DevEx that you're paying out? That's the first question. The second one is just on discovery. Obviously, that's been such a huge part of driving deeper engagement and spending across the content portfolio. At what point do you make a more aggressive push into sponsored tiles? Because that really seems like an obvious commercial opportunity within that discovery engine. So I guess what inning are you in with that? And when do you push more on that front?"
— Matthew Cost, Morgan Stanley

A: "Let's start with Creator Rewards. We're a systems company. We support a complex set of incentives for creators. And as you correctly mentioned, with Creator Rewards, we want to fully optimize the alignment of those incentives with behavior that we think makes sense for the long-term health of the platform. This is a bit of a move from engagement-based payouts that pay for raw time. And in certain cases, paying for raw time can create incentives that maybe aren't optimal, whereas with our new Creator Rewards system, we really want to make it good for creators to drive organic traffic to the platform… On the discovery side, I do want to highlight we have adopted the notion that with discovery, we want to drive long-term health of the platform, both for users, for creators and the overall ecosystem. And that means being very thoughtful about what we reward for discovery. We're getting more and more transparent with what the things are that we reward for creators… And you are spot on that over time as we get better targeting with sponsored tiles, it's a very economical way for creators to bring more traffic to their game… So we do think, long term, there's a big future for sponsored tiles."
— David Baszucki, CEO and Co-Founder

Assessment: The Creator Rewards incentive realignment is a strategically important move — it explicitly pivots the creator economy away from short-term engagement-padding tactics (which generated quality concerns) toward sustainable user-acquisition behavior. Sponsored tiles as a future commercial vector is a meaningful incremental monetization line; we model 2-3% of bookings contribution potential from sponsored tiles by 2027 if execution lands.

Grow a Garden Stickiness & Whether It's Driving Age-Up

A multi-faceted question on Grow a Garden specifically: How sticky is the title likely to be? Is it driving meaningful new-user acquisition? Does it skew older than the platform average and accelerate the aging-up trajectory? Baszucki and Chopra both engaged with depth. Baszucki framed Grow a Garden as evidence of universal-appeal experiences resonating across age cohorts. Chopra surfaced two specific data points: Grow a Garden skews older than platform average, and early stickiness signals (retention, co-play with friends, live-ops successes) are consistent with sustainable evergreen-hit dynamics.

Q: "Can you talk about the new customer acquisition you saw during the quarter? How impactful was Grow a Garden to that new customer acquisition? Do you believe it's sustainable? And then specifically, double-clicking on Grow a Garden, how sticky can that game be? Can you talk about how these new players you've acquired behave relative to other players you've acquired previously? What does the age demo look like? Are you seeing some of the new content migrating to maybe an older age. I know you guys have been trying to age up for a while. Any of these viral titles reaching a generally older audience?"
— Brian Pitz, BMO Capital Markets

A: "There are some interesting dimensions in which Grow a Garden looks a little different than the platform as a whole. Dave touched on this. From an age and a user tenure perspective, we did see Grow a Garden kind of skew a little older than platform average. So I think that's compelling audience expansion. And then in terms of the durability of the experience, I mean, it's obviously still early days, but we see some early signals that this can be a very sustainable level of engagement for the title… It's a game that is very social in nature, and that's the ingredient that we have seen be an indicator of sustainability. Users who play with friends tend to engage at anywhere from 1.5x to 2x the rate of non co-play users? And the game has very successfully taken advantage of live ops, including the Travis Kelce event that Dave mentioned from last weekend. So there's definitely the potential for evergreen hits."
— Naveen Chopra, CFO

Assessment: Chopra's "skews a little older than platform average" combined with "very social in nature" is the most underappreciated structural data point. Social co-play at 1.5-2x engagement of non-co-play users means Grow a Garden's user base is naturally driving the friend-graph expansion of the platform — bringing in older friends and family members who then engage with the broader platform. This is precisely the demographic-expansion mechanism the aging-up thesis requires.

What They're NOT Saying

1. No FY2026 guide framework: Management did not provide any preliminary 2026 framework or directional commentary on whether the +34-37% FY2025 bookings growth rate is structurally repeatable, decelerates moderately, or normalizes to the +20% historical range. With H2 2025 visibility still limited, this is appropriate, but it leaves a wide range of plausible 2026 modeling assumptions and contributes to the post-print valuation uncertainty.
2. No quantitative breakout of Rewarded Video bookings contribution: Baszucki noted "strong and continued pickup" and explicitly declined to split out the dollar contribution; Chopra added that "almost 100 publishers" are now onboarded. This is appropriate at this stage but means the ad-monetization optionality is not yet visible in the disclosed metrics.
3. No quantitative disclosure of Grow a Garden's specific bookings contribution: Management referenced Grow a Garden extensively but did not disclose what % of Q2 incremental bookings came from the title specifically. This makes the platform-wide ex-Grow a Garden spending figure (+36% YoY) the cleanest read on underlying-platform growth.
4. No specific timeline on CPO Manuel Bronstein successor: The departure is announced with 60-day notice but no named successor and no stated search timeline. The silence on the successor profile is the most consequential omission from the call.
5. No commentary on potential M&A or capital allocation framework: With $3.7B in net liquidity and the position as the largest UGC gaming platform, the company has meaningful inorganic optionality. Management did not address whether M&A is part of the H2 2025 or 2026 framework.
6. No engagement/safety regulatory disclosure: Despite ongoing regulatory scrutiny of online platforms (particularly around child safety) in the US, EU (DSA), UK (Online Safety Act), and Australia, management did not provide explicit framework on regulatory cost or trajectory. The Trusted Connections / age estimation rollout is mentioned as a product update, not as a regulatory response.
7. No 2026 capex or investment framework: Despite the bare-metal/cloud-burst architecture being structurally efficient, the multi-year capex trajectory was not characterized. With APAC infrastructure investments continuing and AI/Cube 3D model investments ramping, capex direction matters for the FCF trajectory.

Market Reaction

• Pre-print setup (July 30 close): $124.94, itself the 52-week closing high. The stock had rallied +24% over the trailing 30 days (from $100.65 on July 1). YTD return entering the print was +116% from the December 31, 2024 close of $57.86; trailing 12-month return was +201% from $41.52.
• Options-implied move: Approximately 14%, reflecting elevated positioning into a widely-anticipated Grow a Garden quarter.
• Open (July 31): $149.60, a +19.7% gap on the bookings/DAU beats and the ~$600M FY bookings guide raise. Intraday high $150.59 (+20.5%), a new all-time intraday high clearing the November 2021 peak of $141.60.
• The gap faded through the session: intraday low $135.65 (+8.6%) before a close of $137.79, up +10.3% (+$12.85), still a new all-time closing high (prior: $134.72 on November 19, 2021).
• Volume: 24.7M shares, 3.1x the trailing 30-day average. Heavy institutional participation.
• Tape: the S&P 500 was -0.4% on the day. META closed +11.3% on its own Q2 print from the prior evening; PINS +0.8% and SNAP +1.2% drifted with the tape.

A +10.3% close on the strongest operating quarter in RBLX's public history is structurally justified, and the intraday shape carries the real information: the market opened the print at +20% and then spent the session marking it down as the GAAP-line misses (revenue, EPS, the EBITDA guide re-base) and Chopra's explicit Q3/Q4 conservatism were digested. The Q2 print moved the FY2025 bookings trajectory up ~$600M on the company's own guide; the question is whether multi-quarter execution validates the elevated FY2026 modeling the new price embeds.

The 37%-in-30-Days Compression Problem. The stock moved from $100.65 on July 1 to $137.79 at the July 31 close, a +37% move in one month. For investors who initiated at ~$100 in early July or lower in the spring, the rally has fully captured the operating thesis upgrade. For investors entering today at ~$138, the risk-reward is roughly symmetric: our bull case supports $180-205 over 12 months (+31-49%), while our bear case supports $90-110 (-35% to -20%), with the base case bracketing the current price.

Street Perspective

Debate: Does Q2's Pace Sustain Into Q3-Q4, or Does Chopra's "Normalization" Framing Prove Correct?

Bull view: The Q2 print is structural rather than viral-hit-driven. Non-top-10 hours +47% YoY and platform-wide spending +36% YoY even ex-Grow a Garden prove the underlying platform is growing at +35%+ regardless of any individual hit. APAC accelerated to +75% YoY bookings with multi-year runway; aging-up is delivering +54% YoY DAU growth in the over-13 cohort. The ~$600M FY bookings guide raise is conservative against current trajectory, and Chopra is sandbagging his first quarter to establish a beat-and-raise pattern. H2 2025 will deliver above the guide midpoint, and Q4 specifically benefits from holiday seasonality plus continued viral-hit pipeline.

Bear view: Grow a Garden's contribution to Q2 was material (plausibly $80-120M of the $482M in incremental bookings); the +51% bookings growth without Grow a Garden was likely closer to +38-42%. As Grow a Garden engagement normalizes in Q3-Q4 (which historical viral-hit patterns universally show within 4-8 quarters of peak), the underlying bookings trajectory reverts to +25-30% growth, below the guide range. The tougher Aug/Sept comps (versus 2024 peak viral hits) plus the Q4 booking concentration risk that Chopra flagged are real. The FY guide will prove out at the low end, the stock will compress on Q3 results, and the post-print rally will give back 15-20%.

Our take: Bull view is closer to right on the underlying-platform growth rate (non-top-10 dynamics genuinely confirm structural acceleration); bear view is closer to right on the magnitude of the Grow a Garden contribution. Net: Q3 likely prints within the guidance corridor with bookings growth of +35-40% YoY (above the low end, below the high end). Q4 carries meaningful uncertainty and is the binding swing factor for whether FY proves out at mid versus high end. The probability-weighted FY outcome is at the midpoint to slightly above, validating the guide while not justifying further aggressive upgrades.

Debate: Has the Concentration-Risk Bear Thesis Been Permanently Retired?

Bull view: Five experiences over 10M DAUs simultaneously, four launched in the last 12 months, with non-top-10 +47% engagement growth: the platform's content diversification is now sufficient that no individual hit fade poses a bookings-trajectory risk. The discovery algorithm's long-term-value optimization, the studio tooling improvements, and the creator economy incentive realignment together form a viral-hit production engine. The 1.5-2 multiple turns of historical concentration discount should compress over the next 12-18 months as 2-3 additional quarters of multi-hit data confirm the structural shift.

Bear view: The 5-hits-over-10M-DAUs data point is real but the bookings concentration within that top tier is still meaningful: the top 5 experiences likely account for 35-45% of total platform bookings, which is structurally similar to the top-2 concentration of two years ago, just spread across more titles. A single 5M-DAU hit fade would still be a measurable bookings headwind. The concentration discount compression already happened in the post-print rally; further multiple expansion requires data that has not yet been disclosed (specific bookings concentration percentages, hit-by-hit attribution).

Our take: The directional thesis disconfirmation is real and structural (5 hits versus 2 is genuinely a different concentration profile), but the magnitude of the multiple expansion has likely already been priced in the post-print rally. The remaining concentration discount (roughly 0.5-1 multiple turn versus the no-concentration scenario) is appropriate compensation for the bookings volatility that even a diversified hit portfolio retains. Net: bull thesis is structurally correct, but the additional upside from further compression is now modest.

Debate: Does the CFO/CPO Transition Pairing Justify the Hold (CB) Initiation, or Is the Governance Concern Overdrawn?

Bull view: Naveen Chopra is a credentialed CFO (Paramount Global's CFO through its streaming transition, with Amazon and TiVo before that) and joined a month before the print, giving him a clean slate and no legacy anchoring. CPO Manuel Bronstein's departure is meaningful but the product roadmap is shaped by Baszucki and the broader product/engineering organization, not by the CPO seat alone. A successor announcement within 60-90 days resolves the overhang. The governance concern is overdrawn relative to the magnitude of the operating execution that occurred under this leadership team.

Bear view: The pairing of CFO + CPO transitions at peak operating moment is conspicuous. CFO transitions during periods of strong performance are particularly difficult because the new CFO inherits an "elevated expectations baseline" without the institutional knowledge of how guidance has historically been set. The CPO departure with no announced successor at a moment of unprecedented platform innovation is genuinely material: product roadmap continuity is the binding driver of multi-year thesis. If the CPO role remains unfilled into Q4 or Chopra's first independent guidance set in Q3 misses calibration, the governance discount expands and the stock compresses 10-15% on multiple contraction alone.

Our take: The bear view captures the asymmetric risk correctly. CFO + CPO simultaneously in transition is structurally different from either in isolation. The asymmetric exposure: ~25% probability of a 10-15% multiple-contraction-driven drawdown if both seats are not filled cleanly within 90 days, against ~10% probability of meaningful upside if both transitions land cleanly within 30 days with a strong internal CPO promotion. This is the binding factor in the Hold (CB) initiation: we are not underweight the long-term thesis, we are reflecting the binding near-term execution risk vector.

Model Update & Valuation Framework

Valuation framework. At the $137.79 post-print close, RBLX carries a fully-diluted market value of roughly $102B (739.1M fully diluted shares per the letter), approximately 17x the $5.92B FY2025 bookings guide midpoint and 14x our $7.4B FY2026 base case. For a consumer internet platform growing bookings +30%+ YoY with accelerating FCF, the multiple is rich but not unanchored. It already reflects the multi-hit ecosystem confirmation and the FY guide raise; further multiple expansion requires either (a) FY2026 bookings tracking above $7.4B (toward the bull case $8.0-8.5B), or (b) FCF conversion continuing to outrun bookings growth as the FY25 guide implies.

Revised risk-reward. At $137.79: base case PT $140-160 implies +2% to +16%; bull case $180-205 implies +31% to +49%; bear case $90-110 implies -35% to -20%. The risk/reward has compressed to roughly symmetric, supportive of a Hold rating, with the constructive bias reflecting the long-term operating-thesis confidence rather than the near-term entry attractiveness.

Thesis Scorecard & Initiating Coverage Framework

As this is our initiating coverage report on RBLX, we establish the foundational thesis pillars and signposts that will be tracked through subsequent quarters.

Bull Case Pillars (4)

Bear Case Pillars (3)

Upgrade Triggers (to Outperform)

• Q3 2025 print delivers within or above the guide corridor ($1,590-1,640M bookings), AND multiple does not expand materially (i.e., the stock is at or below ~$140)
• CPO successor announced within 60-90 days with credible internal/external profile
• Chopra's Q3 guidance set establishes a clean beat-and-raise pattern (Q4 guide implies further upside vs. current FY corridor)
• Stock retraces to $115-125 on macro or rotation pressure with fundamentals undamaged

Downgrade Triggers (to Underperform)

• Q3 2025 bookings below the guide corridor (<$1,590M, i.e. under +41% YoY), indicating Grow a Garden tapering faster than expected with no offsetting hit emerging
• CFO Chopra's Q3 guidance materially below Street consensus, suggesting internal trajectory has weakened
• CPO seat remains unfilled into Q4 2025, signaling search difficulty or strategic disagreement
• Material regulatory action (DSA enforcement, US child-safety legislation) imposing significant cost or restricting under-13 platform access

Overall verdict: Thesis is fundamentally sound and the long-term trajectory is intact. Near-term execution risks (CFO/CPO transitions) and entry asymmetry compression (a 37% rally in 30 days through the reaction close) justify the Hold (CB) initiation. We are constructively biased and would upgrade on the conditions above.

Action: Initiate at Hold (CB). Existing holders: hold. New positions: wait for Q3 print confirmation or a pullback to $115-125. Sized positions: consider trimming top quartile to neutral if portfolio weight has expanded materially with the post-print rally.