Datadog Holds BofA's Backing as OpenAI Fear Lingers
BofA is staying constructive on Datadog after a pullback, keeping the OpenAI concentration debate alive a year after Guggenheim's sell call. DDOG last closed at $233.93.
Bank of America has reiterated a bullish stance on Datadog Inc. (NASDAQ: DDOG) after recent share weakness, more than a year after Guggenheim downgraded the stock to sell on concerns that OpenAI, Datadog's largest customer, was building in-house monitoring tools that could remove roughly $150 million of revenue.
Bank of America has reaffirmed a bullish view on Datadog Inc. (NASDAQ: DDOG), using recent share weakness as the argument rather than the deterrent. It is a stance that puts the bank squarely on the other side of the single loudest bear case in observability software: that Datadog's largest customer, OpenAI, is quietly building the tools it currently rents.
Datadog shares last closed at $233.93, up 2.02% on the session, according to market data as of 20:00 GMT on Friday, 7 August 2026. That close sits near the bottom of a wide intraday band of $233.20 to $248.35 — a spread that tells you as much about the stock's current temperament as any research note. The previous close was $229.29.
The accusation that has followed Datadog for a year
In July 2025, Guggenheim downgraded Datadog to sell. The reasoning was concentration risk with a number attached: OpenAI, described as Datadog's biggest customer, was developing internal monitoring tooling that could remove roughly $150 million of revenue from Datadog's base. For a business built on usage-based pricing, where the customer's own growth is the vendor's growth engine, the reverse also holds — a large customer that insources becomes a subtraction that compounds.
That is the specific fear. The broader one it feeds is more uncomfortable for the entire infrastructure-software complex: the biggest AI companies are also the most capable engineering organisations on earth. Anything they buy, they can plausibly build. Observability — the business of collecting logs, metrics and traces so engineers can see why software broke — is not obviously immune.
Why BofA is treating the pullback as an entry point
Bank of America's position, per TheStreet, is that the weakness is the opportunity, not the warning. The implicit argument in any "buy the dip" reiteration on a concentration story is that the market has priced a worst case that is unlikely to arrive in full, and has discounted the rest of the business while doing it.
Three strands usually sit underneath that view:
- Insourcing is rarely total. A hyperscale customer building its own tooling for one workload does not necessarily unwind every contract, and migrations of telemetry pipelines are slow and risky.
- The rest of the book keeps compounding. If growth outside the single named account is intact, the arithmetic of the bear case gets diluted with every quarter.
- Guidance is the referee. Concentration fears are settled by what management is prepared to commit to publicly, not by speculation about a customer's engineering roadmap.
None of that makes the Guggenheim call wrong. It makes it a call about magnitude and timing — how much revenue, how quickly — rather than about whether the risk exists at all. It plainly does.
What the market data does and does not tell you
The tape on 7 August 2026 was constructive for growth software generally. The Nasdaq 100 tracker (QQQ) closed at $723.03, up 1.17%, its strongest showing among the major benchmarks. The S&P 500 proxy (SPY) closed at $773.26, up 0.61%, and the Dow 30 tracker (DIA) at $539.62, up 0.27%. Datadog's 2.02% gain outpaced all three.
The more revealing detail is the intraday range. A stock that traded as high as $248.35 and closed at $233.93 gave back a meaningful part of its day, and the close landed within a dollar of the session low of $233.20. That is the signature of a name where buyers are willing to step in but sellers still control the upper end of the range — consistent with a stock caught between two credible research narratives rather than one.
What the data does not tell you is Datadog's actual dollar exposure to OpenAI, its current growth rate, or what guidance it has issued. Those figures are not in the public record supplied here, and the honest version of this analysis says so rather than filling the gap with an estimate. The $150 million figure is Guggenheim's framing of a downside scenario from July 2025, not a disclosed revenue line.
The concentration problem is now a sector problem
93 gave back a meaningful part of its day, and the close landed within a dollar of the session low of $233.
Datadog is the visible test case, but the structure of the issue extends across software sold to AI labs. Usage-based pricing produced spectacular revenue acceleration when model training and inference volumes exploded. The same mechanism transmits any pullback — insourcing, efficiency gains, a shift in workload mix — directly into the vendor's top line without the cushion of multi-year committed seats.
That is why the analyst split matters beyond one ticker. A sell rating premised on a single customer's build-versus-buy decision is really a rating on how durable AI-era software revenue is when the customer is a peer-level engineering shop. If Guggenheim's version proves right, the read-across hits every infrastructure vendor with a marquee AI logo. If Bank of America's version holds, the market spent a year applying a discount to businesses whose customers turned out to be too busy shipping products to rebuild their monitoring stack.
What to watch from here
For anyone tracking the name, the checkpoints are narrow and specific:
- Any disclosure of customer concentration. Percentage-of-revenue detail on the largest account is the figure that settles the debate; commentary alone will not.
- Guidance versus consensus. Conservative guidance will be read as confirmation of insourcing pressure whether or not that is the cause.
- Whether the $233 area holds. The stock closed at the low end of a wide range. Reclaiming the upper part of that band would suggest the bull case is being funded; losing the base would suggest the opposite.
- Ratings dispersion. When a stock carries an active sell rating and an active bullish reiteration at the same time, volatility around events tends to be higher than the multiple alone implies.
Datadog's problem, in the end, is not that its biggest customer is unhappy. It is that its biggest customer is capable. Investors will be paying for the difference between those two things for some time.
Frequently asked questions
Why did Guggenheim downgrade Datadog to sell?
In July 2025 Guggenheim cut Datadog to sell on customer concentration risk. The firm warned that OpenAI, described as Datadog's largest customer, was developing in-house monitoring tools that could remove roughly $150 million of revenue. Because Datadog's pricing is usage-based, a large customer moving workloads internally reduces revenue directly rather than at contract renewal.
What is Bank of America's current stance on DDOG?
Bank of America has reiterated a bullish view on Datadog, framing recent share weakness as an entry point rather than a warning signal. That puts the bank on the opposite side of Guggenheim's sell rating, with the disagreement centred on how much revenue OpenAI insourcing would actually remove and how quickly it would happen.
Where did Datadog shares last close?
Datadog closed at $233.93, up 2.02% from a previous close of $229.29, as of 20:00 GMT on Friday, 7 August 2026. The session range was wide, from $233.20 to $248.35, meaning the stock finished near its intraday low after trading substantially higher earlier in the day.
How much of Datadog's revenue comes from OpenAI?
Datadog's precise dollar exposure to OpenAI is not established in the material available. The $150 million figure widely cited is Guggenheim's estimate of potential downside revenue at risk from July 2025, not a disclosed company revenue line. Any future disclosure of largest-customer concentration as a percentage of revenue would be the decisive data point.
What is observability software and why does concentration matter?
Observability software collects logs, metrics and traces so engineers can diagnose why applications fail. Datadog is a leading vendor. Concentration matters because usage-based billing ties the vendor's revenue directly to a customer's consumption, so a single very large account insourcing its monitoring can subtract revenue immediately rather than at renewal.
How did Datadog perform against the market on its last trading day?
Datadog's 2.02% gain outpaced all three major benchmarks on 7 August 2026. The Nasdaq 100 tracker QQQ rose 1.17% to $723.03, the S&P 500 proxy SPY gained 0.61% to $773.26, and the Dow 30 tracker DIA added 0.27% to $539.62. Growth software broadly had a constructive session.
Sources
Photo: cottonbro studio · Pexels Licence — source
