“It’s not about how much you earn, but what you’re worth,” said swagger-over-substance billionaire Russ Hanneman in HBO’s Silicon Valley. “And who’s worth the most? Companies that lose money.”

This was satire, I think. The episode aired almost exactly 11 years ago, and if Hanneman’s logic inspired you to immediately plunk cash into iShares Expanded Tech-Software Sector, an exchange-traded fund, you made 278% over the following six years, more than double the S&P 500’s return. There was a lot of talk then about how software companies can scale and stay asset-light, which is true, and why fast-growing ones deserve to trade at high multiples of revenue, never mind profits, which isn’t always true.

For the past five years, returns have stunk, especially following a 28% tumble for the ETF since the end of October. One problem is that after the Covid-19 pandemic, as interest rates climbed from near zero to more ordinary levels, investors decided that they prefer what you might call a reverse-Hanneman approach: valuing companies according to profits or cash flows, not just revenue. (Technically, this mischaracterizes Hanneman’s philosophy. As a venture capitalist, he disliked revenue, too. “If you show revenue, people will ask how much, and it will never be enough,” he warned his software start-up. But you get the idea.)

The other, more immediate problem is that investors have decided that software companies will be victimized by artificial intelligence. And that is partly right, says longtime industry analyst Gil Luria at D.A. Davidson. We spoke this past week about what these companies must do differently to win back stock buyers, and which stocks look likeliest to bounce back.

There’s good news here. Valuations might be low enough to spur takeovers, especially among cash buyers. “Salesforce used to buy companies at 20 times revenue,” says Luria. “It can now go out and buy companies at 20 times cash flow.” Also, value investors who previously stayed away from the sector are sniffing around. But they want more than just rising revenue or even decent cash flows.

First, software companies must stop simply telling investors that they’re going to create winning AI products. Investors can see the flood of money hitting OpenAI and Anthropic, pioneers in AI agents whose recent private funding rounds put their combined valuations at well over $1 trillion. Instead, software companies must help customers incorporate agents from OpenAI, Anthropic, and others into their products.

Second, software makers must show investors that they can use AI themselves to deliver growth with steady or falling head count. That means delivering big upside earnings surprises, especially because upside revenue surprises could be harder to come by. “Software companies are actually in the best position to benefit from the AI revolution because their costs are software development, which can now be done 10 times more efficiently,” says Luria.

Third, cater to value investors. One way to do that is to stop giving away gobs of stock as compensation, especially now that hiring is becoming less difficult.

These things won’t turn around all software stocks. Luria cites companies that make call-center software as an example. “The change in how customer service is going to happen is going to be so radical that it’s going to be very hard for them to make it,” he says. But there are pockets of stocks worth a look.

Start with companies that can continue to produce excellent growth this year, demonstrating that they’ve traded down unfairly with the group. Here, Luria includes Microsoft, down 20% in the past six months and trading at 22 times projected earnings for the next four quarters, and Oracle, down 44% and trading at 21 times. Datadog, down 25%, and Snowflake, down 41%, are in this group, too, although they’re younger companies trading at high multiplies of earnings.

Then there are companies that can use AI to cut costs, including Zeta Global Holdings, down 6% over six months and trading at 17 times forward earnings; Braze, down 14%, trading at 32 times; and Amplitude, down 36% and only thinly profitable. Braze and Zeta also screen as potential takeover targets, along with Klaviyo, down 24% in six months and trading at 21 times earnings.