Databricks Hits $188B Valuation, Proving AI’s Future Is Infrastructure, Not Apps

Databricks, the enterprise data and AI platform, achieved a $188 billion valuation, according to an investigation by the newsroom based on the Tech Startups bulletin dated July 20, confirmed with market sources. The figure places the company at the top of the technology ecosystem and reveals a decisive shift in the AI market: while the fight for end users—the retail side of artificial intelligence—generates headlines, the real money is concentrated in the wholesale side of data and infrastructure.

In the same week, TSMC, the world's largest chipmaker, raised its investment projections to meet AI processing demand, according to the same report. The combination of these two events paints a picture in which the physical race for artificial intelligence—chips, data centers, energy—takes center stage, making it clear that the new frontier is not on the smartphone screen, but in the foundations that support the models.

Databricks: From Big Data to the Engine of Enterprise AI

Founded by the creators of Apache Spark, Databricks built its reputation by simplifying the processing of large data volumes. In recent years, the company repositioned itself as a unified platform for data analytics and artificial intelligence, competing directly with giants like Snowflake and the cloud divisions of Amazon, Microsoft, and Google.

The $188 billion valuation—which surpasses the market capitalization of many publicly listed companies—reflects the perception that data infrastructure is the true fuel of generative AI. Models powering chatbots and assistants rely on proprietary, curated, and governed data sets within enterprises. Databricks offers exactly that: an environment where corporations can store, organize, and train models on their own data without exposing it to third parties.

The valuation jump is also tied to an investment round whose size and participants were not detailed in the initial bulletin. The move, however, aligns with the thesis that venture capital has migrated from the application layer to the infrastructure layer—where barriers to entry are higher and margins more predictable.

The Race for Physical Infrastructure: Chips, Data Centers, and Energy

The capex increase announced by TSMC, as reported by Tech Startups, reinforces that AI expansion increasingly depends on physical capacity. The Taiwanese company, already dominant in advanced chip manufacturing for Nvidia, AMD, and others, needs to expand factories and develop even more efficient processes to sustain demand.

Industry analysts point out that AI's bottleneck has shifted from software to silicon. Each new generation of models requires clusters with thousands of GPUs or specialized accelerators, consuming electricity at industrial scale. Major data center operators, such as Blackstone and other infrastructure funds, have also recently increased investments, according to the same roundup, signaling that the entire supply chain behind AI is inflating in unison.

Energy is another critical component. Hyperscale data centers already face supply constraints in regions like Virginia (USA) and Ireland. Pressure for sustainable sources has led companies like Microsoft and Amazon to sign multi-billion-dollar deals with nuclear and renewable power generators. The message is simple: those who control physical infrastructure will control the next decade of AI.

The Contrast with AI Retail: Devices and Users

While Databricks symbolizes the billion-dollar wholesale market, the retail side of artificial intelligence tries to win over users with more tangible features. Samsung, for example, announced in June its new UFS 5.0 storage technology for smartphones. The component promises to accelerate on-device AI tasks—such as real-time translation and image editing—and reduce battery consumption by over 40%, according to the company.

The innovation is legitimate and should appear in premium devices from Q4 2026 onward. But in market logic, this is a battle for user experience differentiation, not a structural transformation. The value generated by hardware improvements for on-device AI is marginal compared to what is at stake in corporate data platform contracts and chip sales for data centers.

Reorganization at Big Tech: Cuts and Resource Reallocation

Another side effect of capital concentration in infrastructure appears in workforce management. Google employees delivered a petition in July signed by more than 4,500 colleagues requesting protection against layoffs driven by AI task automation. At Microsoft, a new round of cuts is expected to affect less than 2.5% of the workforce, impacting areas like sales and consulting, as reported by Olhar Digital.

These cuts do not indicate financial crisis—Alphabet and Microsoft remain highly profitable. They reflect a reallocation of resources: the money that funded support teams and product development now flows into GPU purchases and data center construction. Pressure for efficiency also responds to growing geopolitical complexity. The United States accused Chinese startup Moonshot AI of using Nvidia GB300 chips obtained via Thailand to train models, illustrating how control over hardware has become a strategic weapon.

What emerges is a two-speed ecosystem. On one side, the billion-dollar race for raw infrastructure, where Databricks positions itself as a data platform provider and TSMC and Nvidia as the suppliers of computational muscle. On the other, the retail side of AI, which fills keynotes with demonstrations of real-time translation and image generation but depends entirely on the physical foundation being built beneath the market. For companies and investors, the signal from 2026 is crystal clear: the real pot of gold lies where data meets chips—and that is where Databricks plants its $188 billion flag.