Eating at the Bay Area’s First “Home” Restaurant

9 min readJul 26, 2021


I recently visited the first legal home restaurant in Northern California. A home restaurant is exactly what it sounds like — it’s a restaurant run out of a residential home. It can offer dine-in, takeout, or both. Home restaurants have different and lighter regulations than commercial kitchens. For example, cooking areas in commercial kitchens cannot have wooden counters or floors; they cannot have exposed brick. Commercial kitchens also have regulations mandating a minimum size for their sinks, a regulation almost all residential homes would fail¹.

Home restaurants were legalized in a 2019 California bill named “AB 626,” but their origins go back to 2012 when California legalized “Cottage Food Operations” (CFOs). This CFO bill allowed residential kitchens to sell low-risk foods like granolas and baked goods — but not full meals. In 2015 a startup named Josephine built an online marketplace for home restaurants, but home restaurants were still against the law at the time. Josephine (think Uber for home-cooked food) got some traction but ended up being shut down by health authorities. The founders of Josephine were frustrated and partnered with an assemblyman from Southern California to pass AB 626. This bill explicitly allowed the creation of home restaurants, also called “Microenterprise Home Kitchen Operations” (MEHKOs). Counties in Southern California have been most welcoming to this law so far, while Alameda County is the first Northern Californian county to allow MEHKOs to operate. Certification of the home restaurant law is done on an opt-in basis by each county, with Southern California moving the fastest so far.

Before AB 626, regulations forbade you from serving meals out of your kitchen. You could not have sold meals because you would have failed the health inspection by county health inspectors. For example, counties and states do not allow you to use “residential grade” refrigerators, freezers, or dishwashers². You must use “commercial grade” appliances that are much more expensive.

AB 626 relaxes many of these constraints. Home restaurants can use residential appliances and do not require expensive grease traps and hoods over their griddles³. The bill has some bizarre caveats, though. AB 626 places a cap of $50,000 in annual revenue on home restaurants, in addition to a cap of 60 meals sold per week. $50,000 may sound like a lot, but that’s revenue — not profit. $50,000 in revenue may only yield $5,000 in actual profit for the owner. I talked to a few people in the industry, and they expect these limits to be revised upwards as the home restaurant industry gains momentum.

I heard about Bao House (the Bay Area’s first home restaurant) as I was scrolling through Google News one day. I saw a headline about it and was fascinated. I knew I needed to go. Bao House is owned and run by Akshay Prabhu, who is also the founder of an online food marketplace called Foodnome. Foodnome aims to be the platform for home restaurants to sell on — similar to other marketplaces like Amazon or eBay. He created Bao House in order to gain publicity and “prime the pump” for his own marketplace. It would be like the founders of Uber driving cars for customers to help test the product and spread awareness. So, Akshay is the founder of an online marketplace called Foodnome — and he is the founder of Bao House, which is one of the early restaurants on the platform. He became involved in this movement when he launched a food cart at UC Davis to make extra money when he was in college. There were no health complaints from customers, but local health authorities shut his food cart down anyway. He felt the journey to launching a food business was bureaucratic and filled with red tape. Foodnome (and Bao House) are the result of those early frustrations.

After hearing about Bao House, I went to, secured a reservation, and invited a friend to come with me. We showed up for dinner on Friday, July 16. From the front it was an unassuming house in North Berkeley. The front door had a sign instructing us to walk around back for Bao House. The backyard was not massive, but it fit about 20 guests. It had an eclectic Berkeley hippie vibe: The table we sat at had a book about meditation for customers to read. It was a charming, avant-garde ambiance. Akshay lived in the house we were eating at.

My friend and I sat down and chatted for 15 minutes before our food arrived. Cambodian food was the cuisine; they served us steamed chicken buns with apple and cucumber salad as the side. For dessert they served sweet buns with berries and powdered sugar inside. If nothing else, they were consistent. I hadn’t tried Cambodian food before, and I enjoyed it. The food was decently priced — about $10–12 each for the meal. While it’s too early to tell, I bet home restaurants will be cheaper than commercial restaurants long-term because they are not paying a restaurant’s #1 cost: rent. The restaurant owner is not paying for a second lease, and they can use appliances which they already use in their home. Leasing space is the biggest cost for most restaurants, so eliminating that will (theoretically) lower the overall cost structure. We didn’t have to pay a tab, because we had already paid through the platform (Foodnome) beforehand. Once we were done we thanked them and went for a walk.

Before AB 626, a fledgling restaurant entrepreneur would have had the following three broad options to start a business: opening a physical restaurant, buying a food truck, or leasing a “ghost kitchen.” A ghost kitchen is a commercial kitchen used for food delivery only. Ghost kitchens do not have sit-down service and only exist on delivery apps like UberEats and DoorDash. A physical restaurant will cost upwards of $250k and 6 months of “lead time” before reaching opening day. A food truck may cost $60–80k and take 4 months⁴. Renting a ghost kitchen will cost $5–7k per month in rent,⁵ and it could take a couple months to find the right location and sign a lease. A home restaurant requires less than $1k, and you can start in a couple weeks.⁶ An entrepreneur still needs to secure health permits to operate a home restaurant, but the permitting process is much less onerous than opening a physical restaurant.

The obvious critique of home restaurants is food safety; people assume home restaurants will be more dangerous than commercial kitchens. However, the quest for “more safety” is complicated. We could make driving safer by reducing speed limits to 15 mph, but the incremental benefit (fewer car accidents) would not be worth the incremental cost (4x more time spent in a car). Put differently, we don’t always want more safety. Instead, we should arrive at the correct “tipping point” where the cost-benefit tradeoff reflects what people actually want.

It’s hard to tell whether home restaurants will be more, less, or similarly safe as commercial kitchens in the long-run. In Southern California, Foodnome claims they have served over 30,000 meals without a single complaint of illness⁷. I guess that rates of food illness will be roughly equal to illness rates in commercial kitchens over the long-term. However, home restaurants may end up being safer because they are hyper-sensitive to the issue and go overboard on food safety. Platforms like Foodnome will proactively address this criticism by monitoring the restaurants coming onto their platform. Critics also assumed there’d be high levels of crime with customers riding in strangers’ cars (Uber) or sleeping in strangers’ homes (Airbnb). Judging by customer adoption of those businesses, those predictions didn’t age well.

A single-minded pursuit of safety is inefficient because there is more than one variable involved. To illustrate my point, let’s assume home restaurants do have a slightly higher level of food illnesses than commercial kitchens. (The facts don’t support this, but it will illuminate the argument better.) Pretend commercial kitchens have a .1% chance of food illness, while home restaurants have a .2% chance. Is the extra .1% chance of safety worth foreclosing on all the entrepreneurial opportunities that home restaurants open up? Would we want a restaurant industry that is .1% more safe, but has prohibitively high barriers-to-entry for new entrepreneurs? We’d have slightly safer food, but dramatically fewer economic opportunities for consumers, entrepreneurs, and employees. Worst of all, high barriers-to-entry fall heaviest on low-income groups. COOK Alliance (a home restaurant nonprofit) “found that the microenterprises in its network kitchens are overwhelmingly run by people from marginalized populations. Eighty-five percent of the cooks were women, and about one-third were first-generation immigrants.”⁸

Furthermore, the limitations on home restaurants should be lifted because health inspectors can’t know the subjective preferences of consumers. Fred may only be comfortable with a .1% chance of illness when he dines out, while Amy would accept a .2% chance of illness. Furthermore, Amy is fine trading an incremental .1% risk for extra time if the nearest commercial restaurant is 15 minutes away, while a home restaurant is only 3 minutes away. Those extra 24 minutes are important to her; she leads a busy life and is fine making that trade-off. Meanwhile, Fred would not make that tradeoff because he is retired and does not have any time pressures. Time is expensive to Amy, while it is cheap to Fred. Thus, Fred should go to commercial restaurants while Amy should go to home restaurants. They have different preferences and risk tolerances. Health inspectors could never know where all these different preferences and risk tolerances balance out. Striking the right balance would be like picking a needle out of a haystack…in the dark.

Letting people “vote with their feet” would allow the industry to develop in a way that matches actual consumer preferences. If home restaurants have a significantly higher chance of illness, say 3% or 5% or 10%, then the movement would collapse before it started. 1,000 early adopters (like me) would try the restaurants, get sick, and spread the word. The movement would fizzle out before large numbers of people tried it. However, if there is not a substantive difference in food safety, then it’s a governmental failure. In this case, the government grossly miscalculated the cost-benefit analysis and kept people in poverty. There were willing buyers and willing sellers, yet the government forbade it. An entire industry was destroyed before it was born.

One might be surprised to learn California is the first state to pass this home restaurant bill; after all, California is derided as a business-unfriendly state. However, the proponents of AB 626 realize that current health regulations harm low-income groups the most. In fact, “illegal” home restaurants have historically been most popular with low-income immigrant groups. A high-income earner may have access to $250k while starting a restaurant, but a low-income immigrant might find it difficult to borrow even $5,000. By defining what a “commercial kitchen” is, regulators inadvertently erect high barriers-to-entry. These barriers-to-entry favor incumbents while discriminating against new entrants.

The future of the home restaurant movement will depend on its level of demand from customers. Maybe home restaurants will disrupt the restaurant industry like Uber disrupted the transportation industry. Or maybe home restaurants become a unique, but niche, part of the overall food landscape. It is too early to tell which will happen. Home restaurants may become more popular in certain demographics than others. They might be popular in cities, but rare in the suburbs — or vice versa. If customers do end up loving home restaurants, the cap of 60 meals per day will likely get expanded along with the $50,000 revenue limit. More states will follow California; Utah passed a home restaurant law recently.⁹ Platforms like Foodnome will have to ensure integrity of their brand, similar to how Airbnb regulates their hosts. Developing the home restaurant industry will be a long journey, and nobody knows where it will go. Different interest groups, like restaurant lobbyists, may not react positively to this new movement. Overall, I had an awesome experience and didn’t get food poisoning. Maybe this idea will work after all.


1. 2021 Alameda County Plan Check Guide For Retail Food Facilities, pages 8 and 16.

2. 2021 Alameda County Plan Check Guide For Retail Food Facilities, page 14.



5. — “Rent for a ghost kitchen, Crocenzi added, can range from $5,000 to $7,000 a month.”


7. — “Is the food safe?”