Will AI First-Mover Disadvantage Dethrone ChatGPT?

Most entrepreneurs in new computing markets including generative artificial intelligence prefer to be “first movers,” on the theory that this helps ensure longer-term leadership by creating scale and enhancing network effects. 

Of course, much hinges on the metrics used to estimate market share. Recurring users, habitual and regular users and samplers are all different ways of measuring share. Estimates of routine use might suggest a market with ChatGPT models having 40-percent share.

Model	Market Share
OpenAI (DALL-E, ChatGPT)	40%
Google (Imagen, Bard)	25%
Stability AI (Stable Diffusion)	15%
Midjourney	10%
Other (Adobe, Baidu)	10%

If we measure using the "ever used it once, even if you no longer do so" metric, ChatGPT might t end to rank higher. 

So computing giants are spending big to get big. And, right now, most observers would tend to agree that ChatGPT is the market share leader.

But the history of new computing markets actually suggests the opposite: pioneering companies that create new product categories often don't become the eventual market leaders. The "first-mover disadvantage" might help us avoid the mental trap that the early innovators will inevitably lead the market longer term. 

In fact, the pattern is so frequent it might come as a surprise, given the attention paid to early-mover strategy emphasized by venture capitalists, for example. 

Computing Market Pioneers and Ultimate Market Leaders
Product Category	Notable Pioneer(s)	Year	Pioneer's Fate	Ultimate Market Leader	Year	Key Advantage of Later Entrant
Personal Computers	Altair 8800 (MITS)	1975	Company sold in 1977, eventually disappeared	IBM PC/Compatible makers (Dell, HP)	1981+	Open architecture enabling third-party development
PC Operating Systems	CP/M (Digital Research)	1974	Marginalized after failing to secure IBM PC deal	Microsoft Windows	1985+	Secured IBM partnership; better graphical interface
Spreadsheet Software	VisiCalc (Personal Software)	1979	Company sold; product discontinued	Microsoft Excel	1985+	Better features, integration with Office suite
Word Processing	WordStar (MicroPro)	1978	Declined in 1980s, company bankrupted	Microsoft Word	1983+	WYSIWYG interface, better Windows integration
Web Browsers	Mosaic/Netscape Navigator	1993	Lost browser wars, sold to AOL	Google Chrome	2008+	Faster performance, better security features
Search Engines	AltaVista, Yahoo	1995	AltaVista absorbed by Yahoo; Yahoo search declined	Google	1998+	Superior algorithm and minimalist interface
MP3 Players	Diamond Rio PMP300	1998	Limited storage and features; company exited market	Apple iPod	2001+	Larger storage, better design, iTunes integration
Smartphones	IBM Simon, Palm, BlackBerry	1992-2002	Market share collapsed after iPhone introduction	Apple iPhone/Android devices	2007+	Full touchscreen UI, app ecosystem
Social Networks	Friendster, MySpace	2002-2003	User exodus, both eventually failed	Facebook (Meta)	2004+	Better reliability, features, and network effects
E-commerce	CompuServe Mall, Internet Shopping Network	1984/1994	Early initiatives failed to gain traction	Amazon	1995+	Customer-centric approach, broader selection
Tablet Computers	Apple Newton, Microsoft Tablet PC	1993/2001	Newton discontinued; Windows tablets had limited success	Apple iPad	2010+	Mature touchscreen technology, app ecosystem
Streaming Video	RealPlayer (RealNetworks)	1995	Overtaken by competitors, lost market relevance	YouTube, Netflix	2005, 2007+	Better user experience, content libraries
Voice Assistants	IBM Simon, Microsoft SPOT watches	1992/2004	Limited capabilities, poor market reception	Amazon Alexa, Apple Siri	2011, 2014+	Cloud computing advances, better natural language processing
Virtual Reality	Sega VR, Nintendo Virtual Boy	1991/1995	Technical limitations led to commercial failures	Meta Quest, Valve Index	2016+	Superior technology, computing power, content ecosystem
Cloud Storage	Xdrive, MediaMax	2000-2003	Early services closed due to business model issues	Dropbox, Google Drive	2008, 2012+	Better synchronization, freemium business model

That might be the biggest cautionary tale for today’s early generative AI market share story. It is too early to know which firms will eventually emerge as the market leaders.