If you have noticed ChatGPT rolling out new features faster than usual, the secret might just be a software tool from Bellevue, Washington. OpenAI confirmed on Tuesday that it has acquired product experimentation startup Statsig to accelerate its consumer software division. The deal highlights a major shift in the artificial intelligence industry, where building the smartest language model is no longer enough. To win the market, companies now have to build applications that feel polished, update seamlessly, and rarely break.
The Bellevue Startup That Caught OpenAI’s Attention
Statsig is not a household name for the average internet user, but software developers know exactly who they are. Founded in 2021 by former Facebook engineering leader Vijaye Raji, the company built its reputation on making software experimentation fast and reliable. They achieved rapid growth by helping companies test new features on small segments of users before launching them globally. By May 2025, they had secured a Series C funding round led by ICONIQ Growth, pushing their private valuation past the billion-dollar mark.
OpenAI decided to bypass a vendor relationship and simply buy the entire operation. The artificial intelligence firm agreed to an all-stock transaction worth $1.1 billion to bring the technology in-house. Before the ink dried on the acquisition, Statsig was already generating a reported $75 million in annual recurring revenue. This revenue stream came from a loyal customer base that relied heavily on their specific toolset.
The platform provides three core functions that software engineering teams require to move quickly:
- Feature flags that let teams turn specific app functions on or off without writing new code.
- A/B testing environments to run two versions of a software update simultaneously.
- Real-time analytics dashboards that immediately highlight if a new feature is crashing the application.
- Gradual rollout controls to release updates to five percent of users before a full global launch.
Those tools are critical for any company trying to scale digital products. S. “Soma” Somasegar, Managing Director at Madrona Ventures, noted the regional impact of the deal, stating that it serves as clear validation that Seattle remains a primary hotbed for engineering talent and software innovation.
A Rare Acqui-Hire With Real Engineering Muscle
The transaction brings approximately 155 Statsig employees into the OpenAI organization, marking one of the largest talent grabs the company has executed to date. The move points to a strategy focused heavily on acquiring proven leadership. Statsig’s founder, Vijaye Raji, is not just handing over the keys and exiting the industry. He is stepping into OpenAI as the newly created Chief Technology Officer for Applications.
In this new executive position, Raji will report directly to Fidji Simo, the current CEO of Applications at OpenAI. His mandate is to bridge the gap between complex machine learning research and everyday consumer tools. OpenAI leaders recognize that raw computing power requires an elegant user interface to actually secure paying subscribers.
“Vijaye has a remarkable record of building new consumer and B2B products and systems at scale. His leadership will help turn that progress into safe applications.”
OpenAI was already an internal user of Statsig’s feature flagging tools long before the purchase was finalized. They knew the technology worked because their own engineering teams relied on it to deploy ChatGPT updates safely. That inside knowledge likely made the billion-dollar price tag much easier to justify to their board of directors.
Despite the change in ownership, the Bellevue office will remain fully operational. OpenAI, which previously opened its own office in the Seattle area in October 2024, plans to let Statsig continue serving its independent clients. Major tech firms like Notion and Atlassian will not face sudden service interruptions.
The Expanding Blueprint for a Trillion-Dollar Valuation
This buyout did not happen in a financial vacuum. In March 2025, OpenAI raised $40 billion at a $300 billion valuation, giving them immense leverage to buy smaller companies using their own stock rather than cash. By October 2025, secondary share sales had pushed that internal valuation even higher, reaching half a trillion dollars. With numbers that large, an all-stock deal allows OpenAI to acquire premium startups without draining their cash reserves.
The Statsig acquisition is just one piece of an aggressive shopping spree aimed at diversifying their business model. They are looking far beyond web browser interfaces and smartphone apps. Over the course of 2025, they have targeted companies in entirely different sectors of the technology industry to build a comprehensive ecosystem.
| Acquired Company | Deal Value | Strategic Purpose |
|---|---|---|
| Statsig | $1.1 Billion | Software testing and application infrastructure |
| io (Jony Ive’s Startup) | $6.5 Billion | Physical consumer hardware devices |
| Roi | Undisclosed | Personal finance application integration |
| Rockset (2024) | Undisclosed | Analytics database and real-time search |
Buying former Apple designer Jony Ive’s hardware startup proved they intend to build physical devices that compete with modern smartphones. Picking up the personal finance app Roi in late 2025 demonstrated an appetite for highly specific, utility-driven consumer products. By adding Statsig to this mix, they now own the exact infrastructure needed to test and deploy all of these new ventures safely.
Faster Updates and Fewer Broken Prompts
For the average person typing questions into a browser, corporate acquisitions usually mean very little. But this purchase directly dictates how reliable your daily digital assistant will be. The market for AI assistants is expected to balloon from $3.35 billion in 2025 to over $21 billion by 2030, according to recent forecasts. Capturing that market requires software that never feels like a beta test.
A 2025 report from Stanford University highlighted that implementing better experimentation frameworks reduces error rates by 25 percent in AI systems. When engineers can test two different language models side-by-side with real users, they can instantly identify which version is providing inaccurate information. They can catch biases and broken responses before a feature goes public.
The speed of product launches is also a major factor. A McKinsey study found that organizations utilizing real-time experimentation data can launch new products up to 20 percent faster. For OpenAI, staying ahead of well-funded competitors like Anthropic and Google means pushing updates constantly without breaking the core experience.
The entire technology sector is shifting its focus from raw research to consumer polish. People no longer care if a language model has more parameters; they care if the mobile app crashes when they try to upload a photo. As the battle for dominance in #ArtificialIntelligence heats up, bringing a dedicated #SoftwareTesting team in-house might be the smartest infrastructure investment Sam Altman has made all year.
