The public software basket is back green for 2026. After the roughest start to the year for software stocks… ever.
At the index level, the aggregate losses from the sell-off have been pared, and a chunk of the group is now putting up a monster year.
- One thing to be precise about up front: this is a huge bounce off the March 2026 bottom, not a return to old highs
- The infra names carry the index, so the cap-weighted basket reads green
- The median software company is still climbing out of a deep hole
- Both things are true at once, and that’s the whole story
Salesforce is up 9% today on a clean earnings beat. HubSpot is up 10%. Snowflake ripped 49% (!) this week on AI-driven results and dragged the whole tape up with it. The names everyone left for dead two months ago, Twilio, Okta, Datadog, ServiceNow, have all strung together big runs off the lows.
And the broader market? Nasdaq is up 20% on the year. Software didn’t just rejoin the rally. It’s now leading parts of it.
Here’s where things actually stand in 2026.
The YTD winners are eye-popping
- DigitalOcean (DOCN): +227.22%
- Datadog (DDOG): +76.08%
- CrowdStrike (CRWD): +54.67%
- Okta (OKTA): +40.73%
- Twilio (TWLO): +32.74%
A 227% year from DigitalOcean. A cloud infra name the market had written off as a commodity is now the single best performer in the group. Datadog up 76%. CrowdStrike up 55%.
Notice the pattern. The companies that own observability, security, and the raw compute layer are getting paid. These are the picks-and-shovels of the AI buildout. When every enterprise is racing to deploy agents, somebody has to monitor them, secure them, and run them somewhere. That somebody is winning.
This is not a uniform recovery. It’s a bifurcation.
Now flip to the other side of the ledger, also year to date:
- Klaviyo (KVYO): -51.83%
- HubSpot (HUBS): -45.83%
- Monday.com (MNDY): -44.66%
- Zscaler (ZS): -38.88%
- Atlassian (TEAM): -35.64%
The index is green. The dispersion is enormous. The spread between the best and worst name in this set is nearly 280 points. That is not a tide lifting all boats. That is the market making a very specific bet about which software survives the AI transition and which gets compressed.
The bet is simple. Infrastructure and consumption-based businesses get a premium. Seat-based application software that an AI agent could plausibly replace gets discounted, even when the underlying numbers are fine. HubSpot is down 46% on the year and it just grew revenue 23% with 83%+ gross margins. The business is excellent. The market is pricing the model risk, not the quarter.
The worst-hit names are bouncing the most
Here’s the tell. Look at the single-day gainers:
- Okta: +28.47%
- ServiceNow: +13.42%
- Atlassian: +11.86%
- Workday: +10.96%
- Rubrik: +9.70%
Atlassian is down 35% on the year and up almost 12% in a single session. HubSpot is down 46% YTD and up 10% today. The deepest cuts are seeing the sharpest snap-backs. Okta up 28% in a day is not a normal move for a company that size. That is short covering, dip buying, and a re-rating all firing at once.
When roughly 75% of software stocks screen as oversold (which they did at the bottom), the bounce is violent and fast. That’s what you’re watching right now. Keep it in perspective: these are explosive recoveries off March lows, not victory laps at fresh highs. HubSpot up 10% in a day is still 46% underwater on the year. The hole was deep enough that even a great bounce leaves most of these names well short of where they started 2026, let alone their 2021 peaks.
March Was The Low Point
The narrative that gutted software was seat compression: the idea that AI agents would replace human seats, and seat-based pricing would collapse with them. For a while the market priced the worst-case version of that story across the entire sector.
What’s reversing it is evidence. ServiceNow got rewarded for monetizing AI productivity instead of getting eaten by it. Salesforce beat and led with Agentforce traction. The durable players are showing they can sell AI as a new line item rather than watch it cannibalize the old one.
And there was a floor. At the March bottom, forward software multiples hit 22.7x, below the S&P 500 for the first time in the cloud era. EV/revenue bottomed near 3.1x to 3.4x after sitting around 7x a year earlier. At those levels the private equity firms started circling with take-private bids, and cash-flow-positive companies trading at distressed multiples found a bid. That’s how bottoms form: not on good news, but on prices that got too cheap to ignore.
The Last 30 Days: All Green
What this means if you’re building
1. The market now reprices software on AI-defensibility, not growth rate alone. Two companies growing 23% can trade 200 points apart. The question every buyer is asking is whether your product gets more valuable as agents proliferate, or less. Build for the first answer.
2. Consumption beats seats in this regime. The winners are usage-based or infra-adjacent. If your pricing is tied to work getting done rather than humans logging in, you’re aligned with where agents take the world, not against it.
3. Great numbers do not float a tough story. HubSpot proved you can beat and still get cut if the market questions the model. Founders raising private rounds should expect the same scrutiny on AI exposure that public companies are getting now.
What this means for private companies and the funding market
The public tape sets the price for everything downstream. When public multiples re-rate up off a bottom, three things follow that matter more to most founders and GPs than any single stock move.
- The exit window cracks back open. With multiples crushed in Q1, the IPO pipeline was frozen and strategic acquirers sat on their hands. A green index changes the math. The same private equity firms that started circling cash-flow-positive names with take-private bids near the bottom now have public comps trending the right way, which makes deals easier to price and finance. M&A and IPO activity tend to lag a public bottom by two to four quarters. The March low was the signal that clock started.
- Late-stage marks get a comp to defend. When the public infra and consumption names re-rate, the best private companies in those categories get a defensible benchmark for holding or raising their valuations. The flip side is real too. If your private company looks like the seat-based application names trading at distressed multiples, expect that discount to show up in your next round. The public market is now the reference price your next lead will anchor to.
- The “is software dead” overhang on Series B and C lifts … somewhat. For two quarters, the AI-disruption narrative made growth-stage software a hard sell at any price. That fog is clearing. Capital is rotating back toward companies that can show AI as expansion rather than cannibalization. The bar is not lower. It’s different. Investors are funding AI-defensible, consumption-aligned businesses and staying skeptical of anything that looks replaceable, which is exactly what the public market is rewarding and punishing in real time.
The public bottom in March was also the private financing bottom. The companies that raised through the ugly part, at sane prices, on real metrics, are about to look very smart.
The SaaSpocalypse is over. The issues though aren’t. The era it ended, where software got a premium just for being software, is not coming back.
What’s replacing it is a market that pays for durability, consumption, and AI leverage, and punishes everything that looks replaceable. Build for the world that rewards you.
