Twilio shares soared more than 20% on Friday and headed for their biggest gain since the early days of the Covid pandemic after the cloud communications software vendor issued an uplifting profit forecast for the coming years.

The stock jumped to $140.12 as of midday trading, which would be its highest close since 2022.

Twilio revealed its new guidance at an investor event Thursday, a little over a year after the company named Khozema Shipchandler as CEO. Shipchandler, who had been Twilio’s president and before that spent 22 years at GE, replaced co-founder Jeff Lawson after a battle with activist investors.

Twilio now sees its adjusted operating margin widening to between 21% and 22% in 2027 as part of a three-year framework for guidance. That’s higher than Visible Alpha’s 19.68% consensus. Twilio’s adjusted operating margin in the most recent quarter was 16.1%.

At Thursday’s event, company executives committed to generating $3 billion in free cash flow over the next three years, compared with approximately $692 million in free cash flow for 2022, 2023 and 2024. The Visible Alpha consensus for Twilio’s 2025 through 2027 was $2.76 billion.

“If we execute well in 2025, I think we write our own story from 2026 on,” Shipchandler told CNBC ahead of the investor gathering.

Twilio, which sends text messages and emails for customers, did not issue a revenue growth target for 2027 at its Thursday event.

But Shipchandler did tell analysts at the investor event that “we’re orienting the company to deliver against double-digit growth over time.”

For 2025, the company said it expects $825 million to $850 million in free cash flow and the same amount in adjusted operating income, with 7% to 8% revenue growth year over year. The Visible Alpha consensus was $814 million in adjusted operating income and about $808 million in free cash flow. The 2025 revenue forecast was in line with LSEG consensus.

Twilio went public in 2016 as a high-growth software company taking advantage of the transition to the cloud. It was one of the big early beneficiaries of the Covid remote work boom as more companies relied on mobile communications to keep in touch with employees and clients. The stock surged more than 240% in 2020.

But in 2022, the stock lost more than 80% of its value as investor focus shifted to profit over growth to reckon with rising interest rates and soaring inflation. Twilio cut 17% of its workforce in early 2023, and activist investors Anson Funds and Legion Partners Asset Management agitated for a sale of Twilio or one of its business units, CNBC reported.

Since activist firm Sachem Head Capital Management won a Twilio board seat in April, the company’s stock has jumped about 81%, as revenue growth has accelerated and losses have narrowed.

By expanding into new areas, such as conversational artificial intelligence, Twilio says it can sell into a $158 billion total addressable market by 2028, compared with $119 billion when only focusing on the communications and customer data platform categories.

Twilio’s preliminary results for the fourth quarter show 11% revenue growth, with adjusted operating income that exceeds the top end of the $185 million to $195 million range that the company issued in October. Analysts surveyed by LSEG had expected 7.9% revenue growth and, according to Visible Alpha, the adjusted operating income consensus was about $190 million.

Baird analysts William Power and Yanni Samoilis upgraded their stock to the equivalent of buy from the equivalent of hold in a Friday note to clients, raising their price target to $160 from $115. The analysts said they “expect a potential beat-and-raise cadence to continue to push shares higher, particularly with the strengthening profitability, cash flow, and capital returns.”

WATCH: Twilio CEO says its time to supercharge innovation cycle

Meta CEO Mark Zuckerberg on Friday announced the company plans to invest around $60 billion to $65 billion in capital expenditures in 2025 as it continues to build out its artificial intelligence infrastructure.

Zuckerberg said 2025 will be “a defining year for AI” and that Meta is building a large datacenter that “would cover a significant part of Manhattan” to power its AI offerings. Additionally, Meta will bring on around 1 gigawatt in compute and end the year with more than 1.3 million graphics processing units, he said.

“This is a massive effort, and over the coming years it will drive our core products and business, unlock historic innovation, and extend American technology leadership,” Zuckerberg wrote in a post on Facebook.

Meta shares hit a new all-time high on Friday during intraday trading after the announcement.

The company has been pouring billions of dollars into AI and ramping up related research and development in recent years, but it’s a fiercely competitive market and will take time before investors begin to reap those benefits. In an April call with investors, Zuckerberg said he expects to see a “multiyear investment cycle” before Meta’s AI products will scale into profitable services, but he also noted that the company has a “strong track record” in that department.

Shares of Meta plunged 16% at the time. The company still generates the vast majority of its revenue from digital advertising.

Zuckerberg said Friday that he expects the company’s Meta AI digital assistant to become the “leading assistant serving more than 1 billion people.” Meta is also building an AI engineer that will contribute “increasing amounts of code to our R&D efforts,” Zuckerberg added.

“We have the capital to continue investing in the years ahead,” he wrote in his Facebook post.

A little-known AI lab out of China has ignited panic throughout Silicon Valley after releasing AI models that can outperform America’s best despite being built more cheaply and with less-powerful chips. 

DeepSeek, as the lab is called, unveiled a free, open-source large-language model in late December that it says took only two months and less than $6 million to build, using reduced-capability chips from Nvidia called H800s. 

The new developments have raised alarms on whether America’s global lead in artificial intelligence is shrinking and called into question big tech’s massive spend on building AI models and data centers. 

In a set of third-party benchmark tests, DeepSeek’s model outperformed Meta‘s Llama 3.1, OpenAI’s GPT-4o and Anthropic’s Claude Sonnet 3.5 in accuracy ranging from complex problem-solving to math and coding. 

DeepSeek on Monday released r1, a reasoning model that also outperformed OpenAI’s latest o1 in many of those third-party tests.

“To see the DeepSeek new model, it’s super impressive in terms of both how they have really effectively done an open-source model that does this inference-time compute, and is super-compute efficient,” Microsoft CEO Satya Nadella said at the World Economic Forum in Davos, Switzerland, on Wednesday. “We should take the developments out of China very, very seriously.” 

DeepSeek also had to navigate the strict semiconductor restrictions that the U.S. government has imposed on China, cutting the country off from access to the most powerful chips, like Nvidia’s H100s. The latest advancements suggest DeepSeek either found a way to work around the rules, or that the export controls were not the chokehold Washington intended.

“They can take a really good, big model and use a process called distillation,” said Benchmark General Partner Chetan Puttagunta. “Basically you use a very large model to help your small model get smart at the thing you want it to get smart at. That’s actually very cost-efficient.”

Little is known about the lab and its founder, Liang WenFeng. DeepSeek was was born of a Chinese hedge fund called High-Flyer Quant that manages about $8 billion in assets, according to media reports.

But DeepSeek isn’t the only Chinese company making inroads. 

Leading AI researcher Kai-Fu Lee has said his startup 01.ai was trained using only $3 million. TikTok parent company ByteDance on Wednesday released an update to its model that claims to outperform OpenAI’s o1 in a key benchmark test. 

“Necessity is the mother of invention,” said Perplexity CEO Aravind Srinivas. “Because they had to figure out work-arounds, they actually ended up building something a lot more efficient.”

Watch this video to learn more.