Start-ups rose to prominence disrupting the status quo with groundbreaking, industry-transforming technology innovations. But recent developments in generative artificial intelligence are flipping the script, with Big Tech incumbents like Microsoft, Google, Apple and others neutralizing early-stage start-up threats through rapid capability integration and consolidation. How can small A.I. start-ups survive to innovate and compete in this kind of landscape?

A.I. wrapper start-ups, who build novel features on top of pre-existing A.I. technology, are proving to be particularly vulnerable to Big Tech updates. These fledgling companies rely on APIs to design products that work as add-ons for Large Language Models (LLMs) like ChatGPT. Late last year, dozens of wrapper start-ups were pronounced “victims” after a ChatGPT update provided users with a newly integrated PDF processing feature – previously a capability only offered by A.I. start-ups as an add-on plugin. With ample resources to throw at development, the biggest players in the space (OpenAI, Meta and Anthropic) can continue to rapidly expand their offerings by integrating the core use cases and features of wrapper start-ups, potentially rendering those businesses and their products redundant or obsolete.

A recent opinion piece in the New York Times took a close look at the dominance of incumbent Big Tech specifically in the generative A.I. space. The authors found that partnerships between A.I. start-ups and Big Tech (like the one between Microsoft and OpenAI) are largely killing competition through consolidation and co-option efforts – acquiring or co-opting start-ups and then either granting or limiting their access to bottomless incumbent capital, data, networks and even hard-to-source hardware, like A.I. chips. And while M&A has always been part of the business world, the wealth of today’s tech giants has all but obliterated the idea of a level playing field (bringing up occasional questions of antitrust violations). In addition to legal and regulatory concerns, eliminating healthy competition in gen A.I., the authors stress, has potential for other negative fallout: stalling progress, sapping creativity and ultimately inhibiting innovation.

Apart from future policy changes that limit or challenge mergers between tech giants and start-ups, how can gen A.I. start-ups create unique value in this complex and challenging landscape? Unique may be the key word in finding a way forward. Right now, start-ups that have access to exclusive or untapped data sets may have a real competitive advantage, especially for training purposes. Areas that seem like “quick wins” for gen A.I. already have a ton of competition (like marketing and legal content generation, chat, customer support) and chances are, the best features from those use cases will soon be built-in by the giants and integrated into their platforms, if they haven’t been already. Apple and OpenAI’s recent agreement to integrate ChatGPT A.I. into their iOS ecosystem is likely to be another example of mass start-up casualties in the face of Big Data capabilities integration. For now, the gen A.I. start-ups that want to stick around long enough to compete need to break new ground, looking for novel A.I. applications that go deeper to solve more complex, nuanced problems.