Oracle is placing a huge bet on OpenAI. Will it succeed, or is it part of a tech bubble set to burst?
Oracle has taken a significant turn by aligning itself with AI, specifically partnering with OpenAI. Unlike other AI giants like OpenAI itself or Anthropic, Oracle isn’t directly building foundational models. Instead, it’s shifting to a model where it offers software-as-a-service, taking on a bold bet on AI’s future. This move comes as their traditional business is on a gentle decline.
Oracle’s older than most of its AI competitors, except for Microsoft, and is leaning into a vast computing agreement with OpenAI, which bizarrely isn’t generating profits yet. OpenAI’s ability to fulfill its promises largely hinges on securing enough funding and achieving profitability.
Oracle runs the risk of investing heavily in building data centers for OpenAI, only for OpenAI to default on the $300 billion agreement. Neither company gave comments on these developments.
The strategy with OpenAI, and broader AI shift, suggests that training AI models isn’t where the money is—rather, it lies in inference, applying models to new data. Oracle views many startup ventures as features it could embed within its existing infrastructure.
Oracle’s well-established enterprise operations, experienced sales teams, and business relationships position it to push its vision forward, which may lead to a consolidation of the AI stack under major players.
Wall Street sees potential in AI, but OpenAI’s lack of public shares means investing in Oracle or Microsoft is the way. Microsoft offers a complex, broader business; Oracle is more focused on AI, offering a purer investment in this technology’s growth. Oracle’s shares and credit default swaps offer insights into the market’s sentiment about AI’s trajectory.
However, Oracle’s history underscores the gap between vision and implementation.
Larry Ellison, Oracle’s founder, known for his grand ambitions, drives the AI focus. Oracle has pivoted away from its comfortable, profitable database business towards riskier, high-capex ventures into AI and cloud computing.
Paul Kedrosky notes that Ellison was likely motivated by wanting excitement. In the 1990s, Ellison was a visionary, predicting cloud-based services long before the technology ripened. Despite this foresight, Oracle didn’t lead the cloud revolution, falling behind companies like Salesforce and AWS.
Oracle’s cloud business is seeing growth with fast expansion but low margins. As ChatGPT ignited AI enthusiasm in Silicon Valley, Oracle was ready to jump into the action. This led to a huge $300 billion deal with OpenAI for data centers, marking one of the largest cloud deals.
Oracle isn’t attempting to compete with Nvidia in chip-making but has good ties with AMD. Ellison’s competitive nature and past disdain for Microsoft have likely influenced this move, positioning Oracle to swoop in as OpenAI distanced from Microsoft.
But aligning closely with OpenAI, notorious for its turbulence, makes Oracle’s fate tied to another’s strategy—risky business. Oracle’s share prices dropped following several large-scale agreements by OpenAI. The deal tethered Oracle’s fortunes to OpenAI’s trajectory, creating a proxy to bet on OpenAI’s outlook.
Nick Patience from Futurum argues that this AI venture is more solid than Oracle’s past missteps like the network computer. However, relying on OpenAI under Sam Altman might be precarious.
Oracle’s OpenAI connection fortuitously arose when a LinkedIn message opened up discussions right as Elon Musk’s planned collaboration faltered. The subsequent deal exceeded Oracle’s prior expectations by a great margin.
To meet the demands of the deal, Oracle is set to construct immense data centers, demanding substantial resources and time. There’s already slippage in deadlines, and Oracle’s aggressive gamble is unmatched in the industry with other firms like Microsoft sitting this out.
Oracle’s skepticism towards the cloud’s profitability, highlighted by former CEO Safra Catz, contrasts with Oracle’s current AI thrust under new leadership. OpenAI needs Oracle for creditworthiness in its ambitious expansion, while Oracle operates under a looming debt burden post-infrastructure plan.
Oracle doesn’t build data centers itself but leases them out, passing on development risks to companies like Crusoe. However, OpenAI tops the list of Oracle’s performance obligations, tying its significant fortunes to OpenAI.
Facing internal upheavals, OpenAI must also contend with challenges in funding and partnerships as it navigates its way to going public. The tumult raises doubts about OpenAI’s stability, impacting Oracle’s outlook.
Enterprises like Anthropic maintain focus, contrasting with OpenAI’s scattered pursuits from consumer applications to enterprise AI. OpenAI’s lawsuits and ventures could strain it further even as AI infrastructure drives up costs.
Oracle’s alignment with OpenAI could prove challenging as they venture collectively into uncertain waters, facing regulatory, environmental, and geopolitical issues like the Iran conflict.
Communities are increasingly resisting data center expansions over concerns of environmental impacts and municipal accountability. In New Mexico and Wisconsin, opposition to Oracle’s projects has led