Amazon.com’s steady expansion of its Amazon Fresh grocery delivery service in Seattle is raising the question of whether the e-commerce giant will ever go national with the business. By doing so, Amazon would position itself to grab a piece of the multibillion-dollar U.S. grocery industry. But there are dangers. The landscape is littered with dot-com era delivery companies (some of which Amazon invested in) that crashed and burned. And the grocery business is known for its razor-thin profit margins.

Amazon Fresh’s bright green delivery trucks have become a common sight on Seattle area streets since the service launched in 2007. The service has steadily expanded its geographic reach and now offers grocery delivery in 49 local ZIP codes, from Bothell in the north to Renton in the south. Amazon Fresh recently indicated it plans to blanket the entire Seattle metro area by the end of the year, though it did not specify the exact borders of the service area.

Doug Herrington, vice president of consumables at Amazon, declined to comment on future plans for Amazon Fresh but continued to classify it as an experiment.

“This is a pilot and we’re working to see if we can figure out how to make the economics work,” Herrington said via email. “It’s a financially difficult business.”

Some observers see a lot of potential in a broader rollout of Amazon Fresh.

Barnaby Dorfman, an Amazon veteran who is now chief executive of Foodista, an online recipe website, said dot-com era delivery firms such as Kozmo.com met with a premature demise because they expanded too rapidly. He said the underlying principles of the grocery delivery model are still strong.

“For a lot of people, shopping is something they don’t like doing,” said Dorfman, whose startup has an investment from Amazon. “The potential to have it show up at your doorstep makes a lot of sense.”

Dorfman said he doesn't think Amazon will go national all at once, but instead establish a presence in another high-density urban area like Manhattan or San Francisco, “places where the truck can go out and over a short distance cover a lot of people.”

It’s not hard to see why Amazon is looking hard at the grocery business.

U.S. online sales of food and beverage products are expected to grow from $8.6 billion this year to $15.4 billion in 2013, according to Forrester Research of Cambridge, Mass.

But Amazon also is well-versed in the pitfalls of grocery delivery, having watched two of its dot-com era investments, Kozmo.com and HomeGrocer.com (later acquired by Webvan) flame out in the dot-com bust. (Herrington, the Amazon executive overseeing Amazon Fresh, worked for Webvan.)

For that reason, Amazon has taken an incremental approach to building Amazon Fresh, and has kept the unit small. The service has a single distribution center, in Bellevue, and a modest fleet of 20 trucks, according to Herrington.

Sucharita Mulpuru, a retail analyst for Forrester, said she doubts Amazon will take grocery delivery national, because of the high costs of setting up food distribution centers in urban centers. Amazon has been selling dry, packaged goods, such as coffee and cereal, for years, but distributes those items through its existing warehouse network. Perishable groceries such as meat and milk require refrigeration and other specialized conditions.

“I don’t see them expanding beyond Seattle,” Mulpuru said.

Amazon Fresh is not alone in the grocery delivery game. Large supermarket chains such as Safeway and Albertsons offer home delivery in select cities. FreshDirect does delivery in New York City, Pink Dot does the same in Los Angeles and Peapod, a unit of Dutch supermarket group Royal Ahold, delivers in various East Coast and Midwest markets.

Over the past two years, Amazon Fresh has tested a variety of features that could inform a future national expansion. The service, for example, tried kiosk-like pickup locations for groceries, but later dumped that idea.

Today, Amazon Fresh offers a “Big Radish” membership program and case sale discounts. It lets customers set up automatic delivery of certain items, and order some non-food items like books with their food deliveries.

Amazon Fresh also added beer and wine to its inventory, an interesting development given Amazon’s much-anticipated plans to sell wine nationally online.

Joseph Park, the founder and CEO of failed dot-com delivery company Kozmo.com who later joined Amazon, said he doesn’t have special insight into Amazon Fresh’s plans but said, “I wouldn’t be surprised if they expand nationally.” He added: “I think (Amazon CEO) Jeff (Bezos) is running that business very well.”

Park, who left Amazon earlier this year to join Bible Gateway, a Bible search engine, conceded that Kozmo expanded too rapidly but noted that it was profitable in three cities -- New York, Boston and San Francisco -- before going under.

“Consumers love convenience and they love instant gratification,” Park said. “That is not going to change regardless of whether you’re buying shoes, buying books, or buying food.”

Follow my updates on Twitter.