Why this VC thinks 2026 will be ‘the year of the consumer’

Funding in shopper tech startups has been in a downturn since 2022, as a turbulent macroeconomic local weather and rising inflation have made VCs skittish about shopper spending energy. For the previous couple of years, most AI funding has targeted on successful over enterprise clients, who present fats checks, multi-year contracts, and fast paths to scale.  

However one VC sees the buyer sector gearing up for a comeback in 2026. 

“That is gonna be the 12 months of the buyer,” mentioned Vanessa Larco, accomplice on the enterprise agency Premise and a former accomplice at NEA, on this week’s episode of the Fairness podcast.

Larco says that despite the fact that enterprises have large budgets and a frantic want to implement AI options, adoption typically stalls as a result of “they don’t know the place to start out,” Larco says. 

“The enjoyable factor about shopper and prosumer…is that individuals already take into consideration what they wish to use it for,” Larco continued. “And they also buy it, and if it meets the necessity, they only preserve utilizing it.”

In different phrases, adoption is faster, and startups constructing AI merchandise don’t must guess whether or not they’ve truly achieved product-market match or have simply received a contract. 

“For those who’re promoting to customers, you’ll know in a short time if it’s becoming a necessity or not, and also you’ll know rapidly whether or not it is advisable to pivot or make some adjustments to your product or completely scrap it and begin one thing completely totally different,” Larco mentioned. 

And in immediately’s anxiety-inducing financial system, shopper tech merchandise that handle to scale show an particularly robust product-market match. 

There are early indications that shopper tech is having a second. Late final 12 months, OpenAI launched apps in ChatGPT, permitting customers to buy with the Goal app, scour the housing market with Zillow, ebook journeys with Expedia, or make a Spotify playlist, all via the ChatGPT chatbot expertise. 

“AI is gonna really feel like concierge-like providers, which is able to do every part for you that you’ve in thoughts,” Larco mentioned. “The query is, which of it ought to be specialised, and which ought to be normal function?”

Or put in another way, as OpenAI works to make ChatGPT the brand new working system of the buyer web, which legacy firms — like Tripadvisor or WebMD — will live on in their very own proper, and which is able to get eaten by OpenAI?

Whereas Larco does suppose 2026 goes to be a “gangbuster” 12 months for M&A, she’s concerned with investing in startups that “OpenAI isn’t going to wish to kill.”

“OpenAI doesn’t handle real-world property,” she mentioned. “I don’t suppose they’ll construct an Airbnb competitor as a result of I don’t suppose they’re gonna wish to handle houses…I don’t suppose they’re going to construct any of those marketplaces that require actual people as a result of they don’t wish to handle the people.”

Except for which startups can fill the gaps, Larco is watching out for what occurs if OpenAI “decides to drag an Apple or Android the place they take a 30% minimize of all of the site visitors they ship you.”

“Is Airbnb gonna wish to play ball with that?” she requested. 

Total, Larco predicts new monetization methods and recent enterprise fashions will emerge from the advanced shopper expertise on-line. 

‘Social has to alter’

Whereas doomscrolling on Instagram about Trump’s seize of Venezuelan chief Nicolás Maduro, Larco observed one thing. She had come to the platform to get information on the escalating disaster, however as an alternative she was overwhelmingly flooded with AI-generated Maduro slop. 

Whereas deepfakes have been steadily turning into mainstream on social media, this was one of many first main information occasions the place AI-generated slop muddied the waters of the reality. 

“At that time, I used to be like, if I’m simply gonna be watching AI-generated movies and images, I need it to be humorous,” she mentioned.

Larco says she has been inundated with sufficient realistic-looking AI movies on social media that she simply assumes it’s all AI at this level, and he or she’s not alone. If all of us begin to assume that nothing we see on Meta’s platforms or TikTok is actual anymore, the query will likely be, the place do you get the actual stuff?

Larco says others may fill within the gaps of the place to seek out truthful, non-AI content material as platforms like Reddit and Digg make strikes to confirm humanity. However for Meta? Perhaps it simply turns into an leisure firm, a platform for user-generated quick movies.

“I feel we should always transfer on from getting your information from [Meta],” Larco mentioned. “You might be simply getting humorous movies from there. It’s not social media. It’s simply gaming and leisure media.”

‘Some issues are higher with voice than a display screen’

When Meta acquired AI agent startup Manus final week, many noticed it as an enterprise play. Larco thinks it could possibly be a transfer geared at enhancing Meta’s Ray-Ban sensible glasses, a product the VC is a large fan of as a result of they permit her to reply cellphone calls, reply to messages, take images and movies, and ask Meta AI questions, all with out having to drag out her cellphone and navigate a display screen. 

Larco says she thinks really helpful voice AI assistants are lastly “on the cusp of occurring,” fueled by extra superior tech and extra strong compute. 

“Some issues are higher with voice than a display screen,” she mentioned. “And since voice sucked, we wanted the display screen as a crutch. However I might love to start out separating out what issues are actually higher on a display screen and what issues are simply higher with audio.”

Getting solutions to questions her children ask about what the tallest constructing is? Positively voice. Taking out her cellphone to sort within the query now feels “archaic,” Larco mentioned. 

“I feel it’ll be actually enjoyable for designers as a result of they lastly get to select and select what type issue is best for what use instances,” she mentioned.