When Netflix founder Reed Hastings spun off the streaming video box he was developing to a little-known start-up called Roku in 2008, he thought it would probably fail.

“There was Xbox and PlayStation and Samsung and Apple TV,” Hastings said in an interview. “Frankly, we didn’t think Roku had much of a chance.”

After first meeting at a conference, Roku CEO and founder Anthony Wood pestered Hastings for months to let his company make a streaming video box for Netflix. Hastings at the time wanted to build the box in-house at Netflix. So the two struck a deal — Wood took a part-time job at Netflix to make the device while remaining CEO of Roku, which had about 15 employees.

That experiment lasted nine months. Hastings wanted Netflix to be available on all sorts of streaming devices, such as Microsoft’s Xbox, Sony‘s PlayStation, and Apple TV. Those companies felt Netflix’s hardware posed a threat to their own businesses. Moreover, people surveyed in focus groups said they wanted a box that could stream more than just Netflix.

So Hastings decided to spin out the division to Roku. Wood received an unfinished device, patents, 20 to 30 Netflix employees (more than doubling the size of Roku) and some cash. In return, Netflix received about 15% of Roku’s equity.

Netflix would later sell its Roku shares to venture capital firm Menlo Ventures to avoid the perception of being conflicted by favoring one streaming distribution manufacturer over another. When Netflix sold its stock in 2009, it claimed a $1.7 million gain on a $6 million investment.

If Netflix had held, its stake would be worth nearly $7 billion today. Roku has been one of the pandemic’s big winners. Shares have have gained more than 480% from March 17, 2020, as the media world shifted to focus on streaming video. Today, Roku’s market capitalization is more than $45 billion.

Wood, who owned more than 28% of Roku at its initial public offering but now owns less than 15% of shares outstanding after various sales through the years, has an estimated net worth of about $7 billion.

“Obviously in hindsight, we missed a fortune,” said Hastings.

To call Roku the offspring of Netflix is literally and figuratively true. While it’s not a carbon copy of its parent, Roku took more than just hardware from Netflix — it took a strand of its corporate DNA.

Wood downplays the comparison. “My relationship to Netflix was obviously very important to Roku,” he said in an interview. “But I only worked there nine months.”

But Roku and Netflix have become market-leading companies worth tens of billions of dollars by out-competing media and technology giants. Both companies could have been acquired in their early days for a fraction of what they’re worth today. Both pivoted their businesses to adapt for streaming video. And both have unusual corporate cultures that can alienate employees who say they live in fear of being fired.

In fact, until recently, Roku’s headquarters were literally next door to Netflix in Los Gatos, California.

Just as Netflix defied the odds to dominate entertainment, Roku overcame widespread industry confusion and doubt to become the U.S. market leader in streaming video distribution. As the media industry has reorganized en masse for a direct-to-consumer world, Roku has become an indispensable intermediary that can guarantee distribution to more than 50 million households.

For its next act, Roku could misdirect the media and technology world again to build its content business — the same kind of move that propelled Netflix to world-beating success.

Just as Netflix began as a DVD rental company, Roku’s first attempts at business bear little relationship to how it makes money today.

Wood, who graduated from Texas A&M with a degree in electrical engineering, founded Roku in 2002 as a maker of high definition video players. Wood initially funded Roku himself with money he had earned from selling other businesses, including DVR maker ReplayTV, which digital audio device maker SonicBlue bought for $120 million. (SonicBlue has since gone out of business.)

Wood then added streaming audio devices to compete against Apple iPods. Unfortunately, Spotify didn’t exist yet.

“I was a little early on that one,” Wood acknowledged.

Next, he added digital signs — common in sporting event concession areas and even used by CNBC for background monitors. Wood eventually spun that unit out to a separate company called BrightSign.

Then came the Netflix deal.

Wood saw a future where Roku would be a centralized distribution platform for digital television. Although Roku seemed like a hardware company, Wood actually envisioned Roku as a services company, making its revenue from channel store fees and a share of advertising from every TV app carried by the platform.

Netflix was Roku’s first customer, followed by Amazon Video on Demand and MLB TV. More recently, Roku added HBO Max, NBCUniversal’s Peacock, Disney+ and many other subscription streaming services — including Roku’s own The Roku Channel. Roku has become the operating system for more than 15 brands of smart TVs, baking its software directly in consumer’s TV sets — just as Wood predicted more than a decade ago.

The pandemic has accelerated Roku’s foothold in American households. With more than 53 million active accounts, Roku has consistently been the leader among all streaming platforms in the U.S., although Amazon is catching up, based on data from Parks Associates. Roku has taken a 33% to 39% market share every year since 2015. In the first quarter of 2021, Amazon Fire TV tied Roku for No. 1 at 36%. Apple TV was third with 12%, followed by Google Chromecast at 8%.

Wood credits some of Roku’s success to Clayton Christensen’s famous business concept of “The Innovator’s Dilemma” — where incumbent companies couldn’t focus on streaming video because they were too busy protecting their older, linear cable TV models. Christensen’s book just happens to be one of Hastings’ favorites, too.

Wood also noted that Roku’s relatively unchanging user interface and simple remote control have appealed to customers because users want simplicity.

“Many companies just don’t really understand the attitude people have when they’re watching TV,” said Wood. “People want to sit there, drink their beer, and watch TV.”

As Wood envisioned, Roku now makes the majority of its money from services — much of which comes from taking a share of every media company’s total streaming advertising time and selling it. When Roku agreed to distribute Peacock, NBCUniversal‘s streaming service, it took about 10% of what would have been Peacock’s ad inventory to sell for itself, according to people familiar with the matter who spoke on condition of anonymity because details of the deal are private.

Using its viewership data, Roku is developing its own advertising technology to better target commercials than what’s possible on linear television. In March, Roku acquired Nielsen’s advanced video advertising business to begin dynamically inserting linear TV advertising, which increases the number of ads that can be showed on a given show or movie and can be used to better target ads to users.

More recently, Roku has invented two content arms of its own. The Roku Channel licenses content from other media companies and has acquired some original programming, including the content that used to be Quibi, the short-lived streaming service founded by Jeffrey Katzenberg and Meg Whitman. Roku sells advertisements against the programming. Roku is also launching an advertising brand studio to help companies make their own original content.

Last year, Roku made about $510 million from its hardware and branded smart TVs. It made $1.3 billion from platform services.

“We focused on the idea that all TV was going to be streaming,” Wood said. “It was obvious. I’m not sure why there were skeptics.”

For years, Wood struggled to find outside financing. Venture capitalists consistently told Roku it was a hardware maker, and hardware wasn’t a good business. Some potential early investors were taken aback by Roku’s modest headquarters in Saratoga Office Center, in Saratoga, California — an uncommon starting spot for Silicon Valley darlings.

The only person who seemed to believe was Menlo Ventures partner Shawn Carolan.

“Silicon Valley does not like to invest in hardware companies,” Carolan told CNBC. That’s because hardware can often be easily replicated and frequently costs nearly as much to manufacture and market as it does to sell. Roku’s hardware, even today, is a zero-profit margin business, according to a person familiar with the matter.

But Carolan saw a clear go-forward strategy based around services.

“I remember this PowerPoint deck I presented around 2009, 2010 where I kind of laid it all out,” Carolan said in an interview. “We called it our popcorn strategy, because movie theaters don’t make money off movies, they make money off the popcorn. How are we going to continue to incrementally add services revenue?”

Wood financed Roku’s Series A round himself. Netflix pitched in $6 million for the Series B as part of the 2008 box transaction. Roku’s Series C, split in two parts in 2008 and 2009, featured one venture capital firm — Menlo Ventures. Carolan and his partners would reinvest again in 2011’s Series D, 2012’s Series E and finally 2015’s Series H — the last round needed before Roku’s IPO.

By 2017, including the Netflix shares it bought, Menlo owned about 35% of all Roku shares. Carolan stayed on Roku’s board from 2008 to 2018.

As the company gained scale, it proved it could make money from its channel store, through revenue shares with media companies, and advertising. Wood expected to hear from other companies interested in acquiring Roku, but few came calling.

Roku held talks with Intel when it toyed with developing OnCue, an Internet-based TV platform, in 2012, according to people familiar with the matter. Intel was eventually willing to pay about $450 million for Roku, but Wood asked for $1.5 billion, according to one of the people. Wood, who several co-workers acknowledged had a quirky personality, told an Intel executive he asked for $1.5 billion because he wanted to open a university in Texas, and that price would cover the expense, according to a person familiar with the talks. The large gap in value doomed the transaction.

About a year later, Amazon approached with an initial offer of about $300 million for the company. Those talks progressed in seriousness, leading Roku to drop its ask all the way to about $690 million, one of the people said. Still, the gap proved too large to cement a transaction.

After that, the offers basically stopped.

“We’ve had less acquisition offers than is normal for a company as successful as Roku,” said Wood, who said he didn’t remember details about the Amazon and Intel offers. “I think it’s because people don’t understand the company. For a long time, they didn’t.”

Waverley Capital managing partner Daniel Leff, who sat on Roku’s board from 2011 to 2018, said the lack of takeover interest from big technology and media companies was stunning.

“Lots of CEOs of big media companies came to spend time with Roku to figure out what it is, what’s streaming, how is it going to disrupt my business?” Leff said. “And I will say, unequivocally, there wasn’t one media executive — and they’re all very smart in their own right — there wasn’t one who believed Roku would be successful, even when it was generating hundreds of millions of dollars in revenue. Even when it went public.”

Roku first attempted to go public in 2014, but bankers told Wood there wouldn’t be appetite for investment until services revenue was 50% of total sales.

“They told us we couldn’t get out, or not at a good price, until we could prove that platform revenue was real,” Carolan said.

So Roku got serious about its platform business. When Roku released its S-1 filing — the document all companies must publish before going public — player revenue in the first half of 2017 represented 59% of total revenue and declined 2% year over year, while platform revenue represented 41% of total revenue and grew 91% from a year earlier.

When Roku went public on Sept. 28, 2017, Carolan broke down in tears.

“I thought, wow, the world finally sees what my partners and I have seen for the last ten years,” Carolan said. “It was just super emotional. And for the past few years, obviously more and more people are finally getting it.”

Wood said he’s spending much of his time now on charting out a strategy for The Roku Channel.

Most of the content on Roku’s channel is licensed from other media companies and studios — and it’s not necessarily their best stuff. The 40,000 free movies and TV shows are largely back-end library content that media companies have deemed unimportant for own streaming endeavors. When Roku can get its hands on more popular content, it tends to be limited — for instance, it only has one season of “The Bachelorette” (Season 13, starring Rachel Lindsay).

In addition to licensed content, Roku has begun dabbling in original programming. Earlier this year, Roku bought more than 75 shows that Quibi created for its short-lived service. It also acquired “This Old House,” which is still making new episodes in its 42nd season. Roku has programming for both kids and adults, building offerings for anyone in the family.

There’s some evidence the original programming is finding an audience. The top ten most-watched programs on The Roku Channel from May 20 to June 3 were all Roku originals. Since adding the Quibi library last month, according to Roku’s own data, more Roku users have seen that programming in two weeks than Quibi users in its six-month lifetime.

The strategy at this point looks a lot look like — surprise — Netflix. In Netflix’s early days, it was happy to license whatever content media companies would give it. Former Time Warner Chief Executive Officer Jeff Bewkes famously called it “The Albanian Army,” emphasizing its small stature at the time.

Now, Netflix spends $17 billion on content a year.

Roku plans to spend more than $1 billion on content next year, according to a person familiar with the matter. Wood declined to comment on the exact total, but did admit the budget will grow next year and in years to come.

Wood also said The Roku Channel creates a virtuous cycle. Roku sells advertising against every ad-supported application on its platform. With its own channel, Roku can offer advertisers another way to market brands. That’s more money, which can be used for more content, making the channel a bigger draw for consumers — and more appealing to advertisers.

There’s real money to be made in free ad-supported video. ViacomCBS’s Pluto TV will top $1 billion in ad revenue next year, CEO Bob Bakish said at a recent investor conference.

Roku announced in March it was raising $1 billion — money that ex-board member Leff expects will go largely toward content. With a market capitalization above that of media companies like Discovery, which is merging with WarnerMedia, and ViacomCBS, Roku is a theoretical buyer for Lionsgate and AMC Networks, said MoffettNathanson media analyst Michael Nathanson.

For the time being, Wood is talking like a CEO who wants to stay under the radar. Wood emphasized Roku was a distribution platform first and a content company second. But if content producers don’t watch out, Roku may “eat their lunch” — just like Netflix did, predicted Nathanson.

“This reminds me so much of Netflix in its early days,” Nathanson said. “I used to interview [Netflix Co-CEO] Ted Sarandos at conferences ten years ago, and he’d say, ‘oh, we’re happy with just one or two original shows.’ Meanwhile, they’d be laddering up into better content. I’d argue companies giving Roku content are digging their own grave.”

Hastings told CNBC he isn’t worried about Roku as a competitor because its goals as an advertising-supported service will be different than Netflix, which is subscription based and has no commercials.

“They’re not a big threat for us,” Hastings said.

Wood agreed with Hastings that The Roku Channel isn’t in competition with Netflix. Roku is looking to capture a person’s attention so it can sell advertising — but it doesn’t need to spend so much on content to keep a person paying $5, $10 or $15 each month. The Roku Channel is available on Amazon Fire TV, Apple iOS and Google’s Android, though the company prefers users watch on Roku’s platform, where it can better monetize viewership data.

“We have less expensive content than a subscription service because it’s not required for us to be successful,” Wood said. “For us, it’s about helping users discover content that appeals to them.”

Still, Roku may be able to increase the quality of licensed content over time. Direct-to-consumer streaming apps need global distribution, and Roku has a roadmap to enter countries around the world. So far, Roku is also in about one-third of all smart TVs in Canada and is the second-largest operating system for smart TVs in Mexico. Europe is its next likely expansion opportunity, said Nathanson, where Google’s Android TV is the dominant incumbent.

As Roku signs new carriage agreements, it could start demanding that each company give it better content for the Roku Channel. Roku asked for quality titles in its negotiations with WarnerMedia and NBCUniversal, according to people familiar with the matter, but it was rebuffed. It settled on paying for a few older, relatively unpopular series, such as NBCUniversal’s “Coach” — for now.

In recent years, Roku has become more aggressive with its carriage agreement demands, including asking for more advertising inventory, higher app store fees, and better content for The Roku Channel. That’s led to delays in reaching agreements with both HBO Max and Peacock. In April, Roku dropped the YouTube TV app from its platform for new customers in a dispute over manipulating search results and hardware requirements. The main YouTube app remains for everyone, but that deal is up later this year — and could test Roku’s leverage.

“They have to be careful,” said Leff. “Netflix is still one of their biggest partners. They don’t want to compete too hard against all of their content partners.”

Then again, if media companies don’t work with Roku, who can they turn to for distribution? Apple, Google and Amazon are still bigger long-term threats, rich with both data and cash, with the power to outspend legacy media for content if they desire. Roku has used its “we’re just the little guy” approach to its benefit throughout its existence.

For now, Roku’s media partners aren’t worried.

“I don’t think they’re challenging to do business with given their market scale,” said Steve MacDonald, president of global content licensing for A+E Networks. “They’re very collaborative and open about information about how we can better monetize our relationship together. They promote our content. They’re good partners.”

That’s what the media industry used to say about Netflix.

Disclosure: Comcast-owned NBCUniversal is the parent company of CNBC.