Mark Murphy

Olivier confirmed heterogeneity plays in Datadog's favor, then highlighted the new development of landing hyperscaler AI labs for training workloads, validating a market that was not real a year ago.

Olivier confirmed more applications are being created and reaching production, with signs visible across every layer of the platform. Directly addressed the code-to-production funnel question.

Pomel noted it's too reductive to peg hyperscaler CapEx directly to LLM observability. The conversion rate is unclear but the CapEx points to massive increases in system complexity, which benefits Datadog broadly.

Pomel emphasized Datadog is not built for customer concentration. More large AI companies are inbound, including hyperscaler-level AI labs, driving both cohort size and diversification.

Pomel explained the standard motion: higher commitment yields better unit economics (lower per-unit price), with short-term revenue dip followed by growth from increased usage.

Pomel cited multiple drivers: channel investment, product maturity especially with Flex Logs underpinning Cloud SIEM, and learning how to properly go-to-market for security, all clicking together.

Pomel highlighted three AI agent categories (SRE, coding/dev, security) and noted Toto is a state-of-the-art time series model that demonstrates research capability to be incorporated into products.

Obstler cited successful early-year recruiting execution plus FX impact from Paris R&D center. Q3 OI benefits from DASH timing ($13M in Q2) and cloud efficiency gains. Pomel added AI training/inference spending is growing within R&D.

Pomel argued AI shifts value from writing code to validating and observing it in production, which is core to Datadog's business.

Obstler attributed strength to strong pipeline, expanded go-to-market investment, and product strength enabling consolidation.

Pomel explained AI native renewals involved higher commits, better terms, and usage optimization simultaneously, typical of larger customers at renewal time.

Pomel noted Datadog is meaningfully outgrowing the non-GPU part of hyperscaler business and is driven by digital transformation.

Aidan said headcount has been roughly flat for 2 or 3 years and should stay there, AI inference costs are manageable, and SaaS tooling costs should not grow meaningfully.

Khozema said RCS is growing off a small base and sees marketing and small-business use cases ahead, but did not quantify the economics to Twilio.

Khozema declined to assess rivals and credited developer experience, innovation and data context; Thomas cited $500,000-plus customers up 36%.

CRO said ISV customers are shifting some traffic to RCS with strong results, called it early innings, and expects the holiday season to be a proof point.

CEO said it ranges from bare voice infrastructure to the full ConversationRelay stack; AI-native firms lean on infrastructure, and voice is gross margin accretive.

CFO said daily data through April show no slowdown, travel is a small share of revenue, and conservatism was built into the second half.

CEO said AI companies first attach to the voice stack, then add voice intelligence and Segment profiles for customer care, citing Cedar as an example.

Management said AI use cases span 2FA to vertical AI startups and voice, but AI is not yet a huge revenue tailwind.

CFO said prepayments are episodic, used opportunistically with strategic vendors to secure better pricing and terms, not an every-Q4 pattern.

Aneel emphasized the core is rock solid and acquired agents are generating rapid growth. Organic agents in early access are even more deeply embedded and harder to replicate. Gerrit described 'lights out finance' vision with continuous AI compliance autopilot and document-driven accounting automation.

Carl said CXOs cite data quality, data integrity, and security as top barriers to enterprise AI adoption, none of which are issues for Workday. Customers are returning to trusted platform vendors rather than evaluating point AI solutions.

Carl explained the subsidiary structure demonstrates focus to government customers and has been very well received in D.C. Building a specific cloud environment with higher security levels. Engagement spans DoD, intelligence agencies and civilian agencies.

Carl confirmed Extend Pro more than doubled YoY, with growing customer and partner innovation on the platform. Also highlighted Workday Wellness, Global Payroll Connect, and marketplace expansion.

Carl explained restructuring savings are being redirected specifically into Agent System of Record and AI. David confirmed strong interest from both customers and partners in managing third-party agents.

Carl argued that DOGE push for efficiency will drive migration from on-premises legacy systems to cloud, positioning it as a tailwind rather than headwind.

Zane confirmed employment trends have stabilized but not improved, consistent with their planning assumptions. Still seeing increased deal scrutiny globally.

Shantanu cited product innovation driving new user adoption, enterprise pipeline with GenStudio and AEP growing 30%, and improved ARR-to-revenue translation. Stock business is approximately $450 million; excluding it, ARR growth would have been 11.2% vs 10.9%.

David gave a detailed media company example: customer spending $10M ARR on creative products added $7M in Firefly services and Foundry. Models trained in 2-3 months as managed services, driving content production efficiency and new revenue-bearing opportunities like social shorts personalization.

David described expanding Stock with creative missions for video content. Emphasized Firefly's advantage in enterprise selection due to quality, controllability, and IP safety.

Shantanu confirmed video model is part of Firefly app and will contribute to the AI book of business, with more color to come at FA meeting.

Shantanu emphasized digital transformation remains fundamental regardless of macro turbulence, with technology helping companies on both top-line growth and efficiency.

Tom said this was always the strategy, deploying distributed inference and compute platform desired by enterprises across the spectrum. Akamai's customer base features world's largest enterprises that spend 10x or more on compute than traditional delivery and security services. Platform has reached the point where it can deliver these results, with more expected going forward.

Tom said no fundamental change in competitive dynamics, Akamai has competed successfully against hyperscalers for over a decade in delivery and security, with 2 of 3 being large Akamai customers. Now adding compute where all 3 use Akamai for distributed performance. Ed added it's not zero-sum, new inference workloads are expanding the market. Customers need ultra-low latency within a couple hundred miles, which differs from even traditional CDN requirements. Akamai is becoming a real player in this growing market.

Ed listed life sciences, manufacturing, health care, and industrials as newer verticals, companies that typically don't have large websites but spend heavily on compute and are good security customers. Mix of channel (including NVIDIA referrals) and direct motion. Strong customer interest and curiosity with many proofs of concept underway. Customers believe Akamai has a right to win in this space.

Ed confirmed mid-to-high single digits organic with about 1% from inorganic M&A. Not anticipating large acquisitions, more tuck-ins like Noname and Guardicore.

Tom said DeepSeek validates their edge compute strategy, with entities already running it on Akamai. Lighter-weight models enable more edge use cases.

Management said deals are priced with purchase orders and power costs locked in, protecting margins; answered directly.

Management expects step-function model performance gains as Grace Blackwell clusters scale; answered directly.

Management said the constraint is powered shells, not power itself, and expects ample power for years; answered directly.

Management clarified one expansion is in Q2 backlog and the other will land in Q3; answered directly.

Management declined to name specific clients but pointed to a broad cross-section of enterprise adoption; partial detail.