SAN FRANCISCO -- For years now, Silicon Valley has been searching in frustration for the next big breakthrough.

Innovative businesses that launched a decade ago, such as Uber and Airbnb, are finally making money or heading toward public offerings - but only after years of bruising fights with regulators and costly operational challenges.

At the same time, the dominance of Facebook, Google, Apple, and Amazon, has made it harder for new startups to burst on the scene. And the impact of powerful and much touted emerging tech, such as artificial intelligence and crypto-currencies, has been more incremental than explosive.

Analyst Benedict Evans, of the venture capital firm Andreessen Horowitz, says that despite these challenges, society is in the midst of a major wave of innovation, and more is coming next.

In his annual slide deck -- a document that many use to read the tea leaves in Silicon Valley -- Evans walks the line between a tech cheerleader and a realist.

He points out that while tech companies have been successful at putting smartphones in everyone’s pocket and connecting people via social media, they have not been as good at convincing people to shop or conduct transactions on these platforms. Just under five percent of consumer spending is conducted online -- let alone through smart refrigerators, microwaves, and other appliances.

“It means we’re only at the beginning of that penetration,” he said. “We’re at the end of the beginning, not at the end.”

But other experts have said that these persistently low numbers may be due to limitations in the way physical experiences that can be digitized. These e-commerce analysts doubt all commerce will move online or onto smartphones because many consumers simply prefer to shop in person.

In another part of his presentation, Evans noted that well-known technology companies that launched in the late nineties and early 2000s, such as Google, Facebook, Yelp or Zillow, have focused on indexing massive amounts of information and presenting it to people in more efficient ways.

Now, he said, a new crop of startups has expanded to delivering services on top of that information-indexing. If Zillow provided real estate listings and comparison shopping, five-year-old startup OpenDoor will buy and sell your house for you, Evans noted in an interview. If Yelp provided a directory of local businesses, today any number of food marketplace apps will give you a similar restaurant search experience - and then deliver the food right to your door. Airbnb isn’t just indexing homes for travelers and Uber isn’t just indexing a directory of cars; they are using the Internet-enabled, location-aware smartphone to create new types of transactions and value in the physical world. Even more recently, startups are now becoming insurers, lenders, and credit providers. Not long ago they would have merely organized financial information.

These digital-to-physical businesses have faced immense challenges in their quest to go beyond software bits. They’ve disrupted entire industries, but their growth has been slower than their predecessors, because their services come with large logistics and regulatory costs.

“We’re moving from business models that didn’t need much capital and were about information arbitrage to providing the entire meal,” he said. “It costs a lot more money to make cars than to tell you what car to buy, but that is just a consequence of the opportunity. You’re also completely changing what the experience of buying a car might look like.”

Evans still believes that the next phase of innovation will be in blockchain, cryptocurrencies, and AI. But like the Internet in 2003, he said, it’s hard to see all the possible applications and ways that these tools will be used.

Blockchain-based apps will be able to overcome the dominance of technology giants by enabling businesses to grow without relying on the major platforms for distribution, he said, echoing a wish among many venture capitalists that something will come along to break the current cycle. Others have called that wishful thinking.

Artificial intelligence, he said, will create more sophisticated ways to understand the intent of customers, which will in turn engender new models for commerce.

For example, one of his slides showed a photo of a chair. No company has the tools to recognize the chair, he said. Amazon’s technology, he argued, effectively sees the chair as a skew number or a bar code, and it can recommend similar products based on what other customers bought. Facebook can show a person which of their friends liked the chair, and sort consumers into profiles based on their tastes. Google can find the chair across many different websites by looking at links and what you type into the search bar.

Once companies can teach artificial intelligence to recognize objects visually, software will be able to understand the realm of intentions and desires that go along with it. More sophisticated nudges will change commerce.

“Once computers can see, it can really tell what you like, and what you don’t like, and it lets you answer questions that were not answerable before,” he said.