AI Data Spending Lifts Software, But Budgets Are Tight

AI spending is shifting from pilots to production, and the fight for data gravity now stretches from databases to observability to graphics-processor access. That puts database providers, observability platforms, and developer clouds directly in the slipstream of generative AI workloads. The challenge is that most of these companies sell into the same enterprise budget lines, and that budget is finite.

Recent figures show premium valuations across the group, which raises the bar for execution. For example, price-to-sales near 14 at Datadog and Snowflake implies investors expect durable growth and margin expansion. In this note, we break down how MongoDB ( $MDB MongoDB, Inc. $294.10 ), DigitalOcean ( $DOCN DigitalOcean Holdings, Inc. $145.34 ), Datadog ( $DDOG Datadog, Inc. $220.94 ), and Snowflake ( $SNOW Snowflake Inc. $227.06 ) say they are capturing AI demand, what the numbers suggest, and the risks the market is weighing.

MongoDB: Turning the database into an AI launchpad

MongoDB, Inc. ( $MDB MongoDB, Inc. $294.10 ) is leaning into AI with its Atlas cloud database and integrated vector capabilities. Company disclosures highlight expanded vector search, access to embedding models, and a broader AI partner ecosystem designed to simplify building retrieval-augmented generation and agentic applications on one operational data platform. Management’s message is that consolidating data and AI tooling in Atlas shortens time to production and helps control costs.

The growth baseline remains solid, with revenue up 22.8% year over year on the latest available figures, and shares still trade at a price-to-sales ratio around 12.2. The market appears to be weighing two questions. Will AI-driven workloads offset macro softness in classic application demand, and can Atlas adoption accelerate enough to defend margins? If the AI features drive more workload consolidation onto Atlas, that would support mix and consumption. The flip side is execution risk if vector and model integrations are not differentiated enough versus hyperscale databases.

DigitalOcean: GPUs for startups and lean teams

DigitalOcean Holdings, Inc. ( $DOCN DigitalOcean Holdings, Inc. $145.34 ) is targeting the AI build-out from the infrastructure side for developers and small to mid-sized businesses. Following the Paperspace acquisition, company materials emphasize on-demand access to NVIDIA H100 graphics processors and a lineup of graphics-accelerated Droplets for training and inference, including configurations marketed around NVIDIA L40S performance, a data center graphics processor geared for AI inference and professional graphics workloads. The pitch is straightforward. Make accelerated compute and machine learning tooling accessible without enterprise complexity or long-term commitments.

On the numbers, DigitalOcean’s revenue grew 15.5% year over year and the shares trade at a price-to-sales ratio near 4.9, a discount to higher-multiple software peers. Profitability also stands out with an operating margin around 17%, which could give the company flexibility to invest behind graphics-processor supply and platform features. The main watch item is capital intensity and pricing. Graphics capacity is costly, and if usage skews to short bursts or promotional credits, unit economics can tighten. Competition from larger clouds offering reserved H100 instances is the other pressure point.

Datadog: Assistants and model monitoring

Datadog, Inc. ( $DDOG Datadog, Inc. $220.94 ) is pushing AI in two directions. First, it embedded generative capabilities into operations with Bits AI, a conversational assistant meant to accelerate incident triage by learning from a customer’s own observability data. Second, its AI product pages describe tooling to monitor model performance and pinpoint anomalies, speaking directly to the observability demands of large language model applications. The strategy ties AI to the existing Datadog data plane, which can deepen seat expansion and reduce churn if it shortens resolution times.

Financially, Datadog’s revenue grew 27.7% year over year and the shares trade around a 13.8 price-to-sales ratio. That premium suggests investors expect AI features to increase stickiness and wallet share across infrastructure, logs, and application monitoring. Risks include competition from point tools in model monitoring and the broader consumption sensitivity of Datadog’s usage-based model if customers reduce telemetry volumes. Execution also matters. AI assistants must reduce toil without hallucinations to win sustained adoption.

Snowflake: From AI Data Cloud to first-party models

Snowflake Inc. ( $SNOW Snowflake Inc. $227.06 ) is using AI both as a product layer and a marketing frame for its platform. Company releases showcase Snowflake Arctic, an open enterprise-focused large language model (LLM), and an expanding Cortex AI suite that brings large language models (LLMs), search, and unstructured analytics into SQL workflows. Snowflake also closed its acquisition of Observe to add AI-powered observability, extending the platform beyond analytics into operations. The unifying idea is to keep data, governance, and AI services in one control plane.

The growth base is still significant, with revenue up 29.2% year over year and a price-to-sales ratio near 13.9. The bear case points to profitability and competition. Operating margin is negative, and hyperscalers are pushing their own integrated AI data stacks. If Cortex and Arctic accelerate developer traction while observability adds a new budget pool, that could support expansion. If customers favor simpler point solutions or push workloads to native cloud services, Snowflake’s narrative would face a tougher test.

Signals that would validate AI-driven adoption

For MongoDB, adoption of Atlas vector search and embedded model workflows in larger enterprise deployments would be a clear signal that AI is moving beyond proofs of concept on the platform. For DigitalOcean, sustained demand for H100 and graphics-accelerated Droplets with healthy utilization is the marker that AI workloads are sticky rather than promotional.

At Datadog, investors may focus on measurable improvements from Bits AI in resolution times and user attach rates to validate AI-driven upsell. For Snowflake, the breadth of Cortex AI adoption and early customer stories that tie Arctic to concrete productivity wins will be important, along with the integration of Observe into cross-sell motions.

Across all four, the same two constraints apply. First, enterprise budgets are finite, which can favor platforms that consolidate tools and data. Second, premium multiples in parts of this group leave little room for missteps. Investors may want to monitor evidence that AI features are translating into higher consumption, new workloads, and improving margins over the next few quarters.