Databricks' $190 Billion Vote of Confidence

Databricks closed a $5 billion round at a $190 billion valuation on August 13, 2026 — a 42% jump in six months — betting that the next fight for enterprise AI will be won underneath the chatbot, in the plumbing that lets agents remember, reason, and stay on budget.

Six months earlier, in February 2026, Databricks raised $5 billion at a $134 billion valuation and quietly took on $2 billion in new debt capacity. By mid-July, the number had already climbed to $188 billion. On August 13, the company confirmed it had crossed the next threshold: $190 billion, a private-market valuation now within shouting distance of Salesforce and roughly double Snowflake's public market cap. The company also disclosed it had crossed a $7 billion revenue run-rate, growing more than 80% year over year in its second quarter — a growth rate that, if it holds, would put Databricks among the fastest-scaling software companies at that size in history.

What happened

The round was led by Coatue, joined by Blackstone, MGX, funds advised by T. Rowe Price, and new investor Sixth Street Growth, according to Databricks' own announcement. Other new backers included BOND, Clearlake Capital, Point72, Premji Invest, and TPG, alongside a long list of existing investors — Andreessen Horowitz, Goldman Sachs Alternatives, Thrive Capital, Fidelity, Franklin Templeton, GIC, Morgan Stanley Investment Management, NEA, Ontario Teachers' Pension Plan, and Temasek among them. Bloomberg and CNBC both confirmed the terms independently, and the UAE's MGX sovereign fund co-led the round alongside Coatue and Blackstone, per The National.

The timeline matters as much as the number. Databricks raised $5 billion at $134 billion in February. It disclosed a target of $188 billion in a raise announced in July. And it closed at $190 billion in August — three valuation markers in six months, each one larger than the last, each one drawing a fresh syndicate of growth investors and sovereign wealth funds willing to write bigger checks at a higher price. That is not the cadence of a company topping up a war chest between IPO windows. It's the cadence of a company that keeps discovering more demand than it planned for, the same compression Fireworks AI's own valuation jump showed at a smaller scale earlier this year.

Databricks says the money will go toward three products it is treating as a single bet: Lakebase, a serverless Postgres database built specifically for AI agents that need fast, transactional reads and writes rather than the batch-oriented queries a data warehouse is built for; Genie, an AI assistant that turns a company's internal data into answers and, increasingly, actions; and Unity AI Gateway, a control layer that routes AI traffic across models and enforces spending limits so a runaway agent doesn't quietly burn through a department's token budget. In the same announcement, Databricks said Lakebase has already crossed $100 million in revenue run-rate, that its Lakehouse data-warehousing product has passed $1.5 billion in run-rate while growing over 100% year over year, and that it now has more than 1,000 customers spending above $1 million a year and over 100 spending above $10 million.

Co-founder and CEO Ali Ghodsi framed the pitch around what enterprise agents actually need to function: "That requires real-time operational data with Lakebase, context from across the business with Genie, and multi-AI cost controls with Unity AI Gateway," he said. Coatue co-founder Thomas Laffont, whose firm has backed Databricks since 2019, put the pace of execution at the center of his rationale for leading the round: "They've compressed R&D timelines that used to take years into months, more like a research lab than a typical software company."

The competitive backdrop

None of this is happening in a vacuum. Databricks' longtime rival, Snowflake, has spent 2026 making its own case for owning the "system of intelligence" layer that sits between enterprise data and AI agents, building out products like Cortex Sense and Horizon Context to compete for the same workloads. The two companies increasingly resemble mirror images: Snowflake pushing deeper into governance and orchestration from its warehouse roots, Databricks pushing an entire agent stack — Genie, Lakebase, Agent Bricks, Unity AI Gateway — out from its lakehouse core. Both are converging on the same bet: that agents, not dashboards, will be the primary unit of enterprise software going forward, and that whoever controls the data layer underneath the agent controls the economics of the agent itself.

The valuation gap between the two companies is now the story in itself. Snowflake trades publicly at a market capitalization in the neighborhood of $100 billion on roughly $5 billion in trailing revenue — a public multiple in the low-to-mid 20s. Databricks, still private, is being priced at roughly 27 times its $7 billion run-rate. Investors are not just paying for revenue; they are paying a premium for growth rate, for staying private long enough to compound before public-market scrutiny arrives, and for a specific thesis about where the AI infrastructure buildout goes next. That thesis — that the money moves from model training toward the "boring" layer of context, memory, and cost control that makes agents usable in a regulated enterprise — is precisely what Lakebase, Genie, and Unity AI Gateway are built to serve.

It also explains why sovereign wealth funds and late-stage growth firms that don't normally chase software valuations at this multiple showed up for this round. MGX, backed by Abu Dhabi, has been an aggressive allocator across the AI infrastructure stack this cycle, from chips to compute to now data platforms. Blackstone and T. Rowe Price bring a different kind of validation — the sort of institutional capital that treats a private round as a pre-IPO position rather than a venture bet. Their presence signals that Databricks' path to a public listing, long rumored and repeatedly delayed, is being priced by people who expect to still be holding the stock when that listing happens.

Who this changes things for

For enterprise buyers, the message is that "picking a data platform" and "picking an AI agent platform" are collapsing into the same decision. A company that has already centralized its data in Databricks' lakehouse now has a default answer for where its agents live, what they're allowed to query, and how their model spend gets capped — without stitching together a separate vector database, a separate LLM gateway, and a separate governance layer from three different vendors. That convenience is exactly what Databricks is charging a premium for, and exactly what a $100 million run-rate on a two-year-old product like Lakebase demonstrates buyers are willing to pay for.

For Snowflake and the broader field of point-solution vendors — API gateways, vector databases, agent-governance startups — the message is less comfortable. Every dollar an enterprise commits to Unity AI Gateway is a dollar it isn't spending on a standalone LLM router. Every workload that lands in Lakebase because it's already sitting next to the enterprise's Unity Catalog metadata is a workload a specialized Postgres-for-AI startup doesn't get to bid on. Platform incumbency, once again, is proving to be worth more to buyers than best-of-breed assembly — at least for the parts of the stack that touch governance, compliance, and cost.

For the investors who wrote checks at $134 billion in February and are now sitting on a position marked up 42% by August, the round is a reminder of how compressed the AI infrastructure cycle has become. Six months used to be a rounding error between funding events for a company of this size. In 2026, it's long enough for a full valuation cycle. That compression cuts both ways: it rewards investors who move fast and punishes anyone still running a traditional 18-month diligence process on a company whose growth rate resets the comps every quarter.

Why it worked

Three things made this round possible in a market that has otherwise grown skeptical of software multiples. First, Databricks paired the fundraising announcement with hard revenue disclosure — the $7 billion run-rate, the 80%-plus growth, the free cash flow positivity — rather than asking investors to take valuation on faith. Growth-stage AI rounds in 2026 increasingly require this kind of receipts-first pitch, a shift visible even in how Anthropic has pitched its own scale to investors; the era of valuation by narrative alone is over for any company writing checks this large.

Second, Databricks did not wait for Lakebase, Genie, and Unity AI Gateway to become theoretical AI-hype products. It showed them already generating revenue at meaningful scale — Lakebase's $100 million run-rate is a real number attached to a product that only recently shipped, and it is the single most concrete proof point in the entire announcement. Investors are no longer pricing AI infrastructure companies on total addressable market slides; they're pricing them on whether a specific new product line can already stand on its own.

Third, the round assembled a syndicate that spans venture, sovereign wealth, and traditional asset management in a way that signals durability rather than momentum-chasing. When Blackstone, T. Rowe Price, and MGX all show up in the same round as Andreessen Horowitz and Thrive Capital, it tells the market that this isn't just growth-stage FOMO — it's a bet multiple types of capital are willing to hold through an eventual IPO.

The takeaway

The lesson for anyone building at the intersection of data and AI is not that valuations are irrational — it's that the market has decided where the durable value in the agent stack actually sits. It isn't in the chat interface, which is increasingly commoditized and swappable between model providers. It's in the layer underneath: the database that gives an agent memory, the catalog that gives it context, and the gateway that keeps its spending from spiraling. Whoever owns that layer owns the renewal, regardless of which model happens to be answering the prompt this quarter.

Databricks spent a decade building the infrastructure layer for data science before most of the market cared. It is now betting $190 billion of investor conviction that the same playbook — be early, be boring, own the plumbing — works again for agents. If Lakebase's trajectory from zero to $100 million in run-rate is any indication, the market seems to agree, at least for now. The real test will come when the next funding cycle asks whether that growth rate can survive the scrutiny of a public listing rather than the enthusiasm of a private one.

Frequently Asked Questions

How much did Databricks raise and at what valuation? Databricks closed a $5 billion strategic funding round on August 13, 2026, at a $190 billion valuation, up 42% from the $134 billion valuation it held just six months earlier in February 2026, according to the company's announcement.

Who led Databricks' $5 billion funding round? Coatue led the round, joined by Blackstone, MGX, funds advised by T. Rowe Price, and new investor Sixth Street Growth. Additional new investors included BOND, Clearlake Capital, Point72, Premji Invest, and TPG, with existing backers like Andreessen Horowitz and Thrive Capital also participating.

What is Lakebase, and how is it performing? Lakebase is Databricks' serverless Postgres database purpose-built for AI agents that need fast, transactional data access. It has surpassed $100 million in annual revenue run-rate, one of the fastest-scaling new products in the company's history.

How does Databricks' valuation compare to Snowflake's? At $190 billion on a roughly $7 billion revenue run-rate, Databricks trades at about 27 times revenue as a private company. Snowflake, its closest public rival, has a market capitalization near $100 billion on roughly $5 billion in trailing revenue, putting its multiple in the low-to-mid 20s.

Editor's note — sources consulted but not linked inline: Yahoo Finance, Silicon Republic, Investing.com, SiliconANGLE, TechCrunch.