In short: Terranova's final trade (43:18). He is "rebuilding positions in data software" and cloud security — "Datadog is the name that you want to own in that space," as semis and software rise together on the day.
Datadog sells software that watches companies' cloud systems and apps and alerts them when something breaks or slows down, with security monitoring on top. Terranova is buying software again now that software and chip stocks are rising together, and names Datadog as the one to own there.
In short: Named by Terranova in the same breath as CrowdStrike, Okta and Fortinet as down on the day "because positioning is full and the expectations could not be exceeded no matter how well they actually performed." Datadog is an observability/monitoring platform rather than a pure security vendor, which is itself the point: the selling is factor- and crowding-driven and does not respect business-model boundaries.
Datadog is an observability platform — it monitors whether software and infrastructure are running properly, which is adjacent to security but not the same business. It gets named in the same breath as the three security vendors and falls with them.
That is itself the useful detail. Factor-driven selling follows how stocks are grouped in investors' baskets, not how the companies actually make money. If your holding sits in a popular basket, it will trade with that basket on the days the basket is being sold, whatever the business does.
In short: The flagship of the contrarian software basket bought while hedge funds were shorting software as a hedge against long-AI books: "we bought a lot of the software names like Datadog, which has done very well." Cybersecurity was the biggest overweight inside that basket.
Datadog sells software that monitors whether companies' systems are running properly. Nothing changed about the business — what changed was who was selling it. Hedge funds piled into AI chip stocks and, to protect themselves, sold software stocks short as the offset. Because AI names are far bigger than any one software company, that hedging crushed software prices mechanically.
Singh bought the wreckage. Datadog "has done very well," and cybersecurity — which he considers the least AI-replaceable corner of software — was his largest position within the basket.
7:50So, what we had was because AI stocks are so much bigger than the individual software stocks, we had very high short interest and systematic selling with software names and we bought a lot of the software names like Datadog, which has done very well. And even cybersecurity was our biggest overweight.
In short: Repurchased the position he sold in June: "it doubled in the spring, we sold it, and we bought it back after earnings because it sold off 17% on positive earnings. Cybersecurity is, I think, less affected by AI, which was the thesis that worked." Morgan Stanley's AI-winners framework independently puts Microsoft and digital-security names like Datadog on the winning side in all three model scenarios. Deck page 58.
Datadog sells software that monitors whether a company's own software and cloud systems are working, and flags security problems. Singh owned it, watched it double in the spring, sold it — and has now bought it back because it fell 17% despite reporting good results.
His reason for owning this corner of software rather than the rest: security and monitoring tools are hard for AI to replace, because the value is in the plumbing and the trust, not in generating text. Morgan Stanley reaches the same conclusion independently, putting Microsoft and security-software names on the winning side of AI in every scenario it models.
Full passage: premium transcript (PDF).
In short: AI concentration bites. Q2 revenue +36% Y/Y to $1.12B ($40M beat), accelerating again from 32% in Q1, with adjusted EPS $0.65 ($0.07 beat), non-GAAP operating margin 23% and free cash flow $279M at a 25% margin; $100K+ ARR customers accelerated 23% to 4,720. Underneath, the business broadened: non-AI customers accelerated into the high-20s from roughly 18% a year ago, 58% of customers now use four or more products, and calls to Datadog through MCP — the standard letting AI agents interact with software — are up 22x since Q4. But Q2 exposed the downside of AI concentration: Datadog's largest customer has begun reducing usage — "a leading AI company that uses 17 products and recently signed a nine-figure renewal" — with the decline built into the second-half outlook. Datadog still raised FY26 revenue guidance to $4.45–$4.47B (from $4.30–$4.34B) and adjusted EPS to $2.50–$2.54, with Q3 above consensus — and shares still fell about 17%. "A single AI customer can now meaningfully move the growth curve… The question for the rest of FY26 is whether that broad-based acceleration can absorb the largest customer's pullback." A disclosed author holding.
Datadog monitors software systems — it tells companies which part of their application is slow or broken. Revenue grew 36%, accelerating again, profits beat, and guidance went up. The shares fell about 17%.
The reason is concentration. Datadog's single biggest customer — described only as "a leading AI company," using 17 of its products and having just signed a contract worth more than $100 million — has started reducing how much it uses. Management has built that decline into its second-half forecast. When AI labs grow explosively, their suppliers grow with them; when they optimise their costs or build tooling in-house, the supplier feels it immediately.
The encouraging half is that everything else is getting better, not worse. Customers outside AI accelerated from around 18% growth a year ago to nearly 30%, and 58% of customers now use four or more Datadog products — a stickier relationship. Usage by AI agents is exploding too: calls into Datadog through MCP, the standard that lets agents operate software, are up 22-fold since Q4.
So the quarter poses a clean question rather than answering one: can broad-based growth from thousands of ordinary customers absorb the pullback of one enormous AI customer? That is why the read is neutral. The author owns it; analysis, not a recommendation.
In short: "One of our best cybersecurity and software name longs," reporting Thursday pre-market.
Full passage: premium transcript (PDF).
In short: His preferred software exposure: "DataDog has crushed it" and trades near all-time highs — vindicating the house view that cybersecurity/observability is the best place in software (moat > AI-disruption risk).
Full passage: premium transcript (PDF).
In short: Endorses Leon's call on the one bright spot in software — infrastructure/data-driven software can win: "a lot of the big model companies like Anthropic are using DataDog's tools. So that's a pretty good tell."
Datadog sells "observability" software — tools that monitor whether a company's apps and servers are running properly. Sacerdote is broadly bearish on ordinary business software, but he carves out an exception for data-driven and infrastructure software, and points to Datadog as a "unique asset" at the heart of the AI buildout.
His tell: the big AI labs themselves, including Anthropic, use Datadog's tools to keep their systems healthy. When the companies driving the boom are your customers, that's a good sign — even if you can debate whether the stock deserves 30× or 40× earnings.
20:04Maybe it's just a matter of time, but maybe it's a culture thing. They don't have the right people. It's a different sales process, because you're selling a service, not software. And it's a different business model. So I don't think software's going to be bouncing anytime soon, but we're watching it really carefully, because we might see a few software companies actually develop and benefit from AI.
In short: Software-ROS example: "Datadog flushed its 2021 all-time high… and promptly died" — a failed breakout he flags in the IGV basket.
Nothing matches this filter.
Verbatim excerpts from the public transcripts (auto-pulled at each mention's timestamp, lightly cleaned). Timestamps link into the video; "source page" opens that commentator's full analysis at this row. Click a mention's header line to expand it (one open at a time). For personal study — not investment advice.