How Leading AI Startup Investors Approached Artificial Intelligence In 2023

Editor鈥檚 note: In 2023, SA国际传媒 News interviewed active startup investors in artificial intelligence. Below, we publish highlights from those interviews. Read the full interviews with General Catalyst, Bessemer Venture Partners, Accel, Insight Partners, Index Ventures, Sequoia Capital, Section 32, M12 and Sapphire Ventures.

Over the past year, we鈥檝e chatted with more than a half-dozen leading investors about the AI space and where they are focusing their efforts.

For many, the foundation-level models 鈥� unless you are an early investor 鈥� are challenging for venture funding due to the vast sums of capital required. That鈥檚 because it costs billions of dollars to develop foundation models.

But the tools required to manage these models and data are garnering investor interest. And the application-layer software applying these models for consumers and businesses is a use case that will play out over the next decade.

Repeatedly, what we heard from investors is that the immediate and massive adoption by consumers of AI tools like and image-generation models in the past year was a critical off-to-the-races moment for this industry.

鈥楢ttention is all you need鈥�

We鈥檝e now had four or five eras of artificial intelligence, , president at , told us recently. This latest is undoubtedly the generative AI era, he said: 鈥淲e are just in the first innings of what generative AI can do.鈥�

鈥淐hatGPT is a meaningful step forward,鈥� , a managing director at venture capital firm , told us in our far-ranging conversation earlier this year.

More important than the technology, he said, 鈥渋t has fired up the imaginations of nontechnical people. It鈥檚 probably the fastest thing that ever got 100 million users using it all at once.鈥�

For , a partner at , the technology was not a surprise 鈥� it鈥檚 been around for a few years 鈥� but she was surprised at how quickly ChatGPT took off, and then made its way into the boardroom. Index Ventures is an investor in , one of the foundation model companies that found the market flung wide open up after the launch of ChatGPT.

Meanwhile, from the corporate side, had first invested in ChatGPT developer in 2019.

鈥淲e were already looking for uses of generative AI before the whole thing exploded,鈥� , a partner at , the software giant鈥檚 venture fund, told us.

The moment that drove home the value of this technology for Stewart was not the public launch of OpenAI鈥檚 GPT 3.5 just over a year ago. It was earlier in 2022, when connected the power of GPT-3, a prior version, in a highly automated marketing tech platform that got marketers very excited.

The current technology evolution is driven by the continued evolution in AI models, , a partner at聽 , told us in our conversation earlier this year. 鈥淭he storage and processing power has been getting cheaper,鈥� he said. 鈥淭he models get better, the chipsets and the processing power get cheaper, and then the storage gets cheaper. That鈥檚 created a flywheel here that鈥檚 allowed these massively trained LLMs that have caught everyone鈥檚 imagination.鈥�

However, 鈥渋n the end the stuff is just software,鈥� he said.

Adoption cycles

The 鈥渁doption curve on this one will be mind-blowingly fast,鈥� , a partner at , told us 鈥斅�predicting that AI will be easier to adopt than previous platform shifts like mobile and cloud computing. 鈥淭his time it鈥檚 an API call to a large language model.鈥�

, a partner at , spoke about the firm鈥檚 strategy to follow developers and talent.

The firm has seen more AI companies forming, she said. 鈥淲e鈥檝e seen the rise in early-stage AI investing 鈥� most notably, pre-seed and seed-stage AI companies that we鈥檙e actively investing in right now.鈥�

Price-Wright, who previously worked at and partnered with organizations to manage and understand their data, said the opportunity has expanded to incorporating AI.

鈥淭here鈥檚 this similar challenge over the next decade of how existing enterprises with large data sets and large customer bases and complex operations start to incorporate AI to make their products better, to make their operations more efficient, to make them more streamlined, to make them faster and more nimble at decision-making,鈥� she said.

鈥淗ow effectively organizations are able to really leverage data and decision-making; I think we鈥檙e still in that transition,鈥� she added.

from New York-based investment firm told us the firm is seeking out 鈥渃ompanies that are building domain-specific models on top of private data sets with great user experiences and workflows.鈥�

Harrison of Section 32 noted that 鈥渋t鈥檚 expensive software because of the compute and because of the talent required. So it has to be that much better. Or it has to do something we couldn鈥檛 previously do with software in order to drive high margins. Otherwise it鈥檚 low-margin software.鈥�

He predicts the GPU crunch will alleviate itself in the next 12 to 18 months and the costs of the models will come down.

, a partner at , predicted that AI will have a dramatic impact on many existing companies. 鈥淵ou鈥檒l see new companies for which AI makes the difference between the product being so-so and potentially [being] a game changer in its category,鈥� he said.

He predicted: 鈥淵ou鈥檒l see a lot of companies that in hindsight will look like they鈥檝e replaced existing software players.鈥�