When can I eat meat again?

By Claire Yip, co-founder of Cel­lu­lar Agri­cul­ture UK. Th­ese views are my own.

Summary

Timeline: When we can ex­pect highly similar cost-com­pet­i­tive al­ter­na­tives to an­i­mal products

Timeline

  • There is a lot of un­cer­tainty around when we will be able to eat meat grown from cells, and how we should di­vide our efforts be­tween that, plant-based al­ter­na­tives, and other forms of an­i­mal ad­vo­cacy. This post seeks to give sen­si­ble, un­bi­ased views on the fu­ture of al­ter­na­tive pro­teins.

  • How­ever, these views are un­cer­tain too: I have c.40-60% con­fi­dence. Th­ese es­ti­mates are not set in stone. Fac­tors like in­vest­ment and ac­tivity would quicken progress, but things will also prob­a­bly take longer than we ex­pect, for un­ex­pected rea­sons, and we don’t know ev­ery­thing that we don’t know.

  • Th­ese are es­ti­mates for cost-com­pet­i­tive al­ter­na­tives. We will be able to buy these prod­ucts ear­lier than this.

  • In the next 5-10 years, ex­pect to see plant-based ver­sions of pro­cessed meats be­come more wide­spread and deli­cious. Yay, chicken nuggets! You’re also in luck if you want plant-based or an­i­mal-free scram­bled eggs, omelettes, milk, cream, yo­ghurt, or whey pro­tein pow­der.

  • Th­ese plant-based prod­ucts will get even bet­ter in the next 10-20 years, es­pe­cially as they’re blended e.g. with col­la­gen (pro­duced with­out an­i­mals), or real meat cul­ti­vated from cells. De­cent an­i­mal-free but­ter, cheese, and whole egg prod­ucts might be­come a re­al­ity! Pet food pro­duced from an­i­mal and microor­ganism cells will also be more eas­ily available.

  • If you want to eat un­pro­cessed whole meat like ba­con or sashimi with­out hurt­ing too many an­i­mals or blow­ing your gro­cery bud­get, it looks like you’ll have to wait a few decades (30-50 years).

  • My time es­ti­mates are mostly based on pri­vate con­ver­sa­tions with plant-based and cel­lu­lar agri­cul­ture com­pa­nies, at con­fer­ences and through writ­ing a re­port on low-cost cell cul­ture me­dia for the Good Food In­sti­tute, as well as my un­der­stand­ing of the tech­ni­cal progress needed for each tech­nol­ogy. How­ever, my views do not rep­re­sent those of GFI.

Ac­tions you can take

  • Donors: If you want to donate to this space, promis­ing re­cip­i­ents are the Good Food In­sti­tute and New Har­vest (for open ac­cess re­search into cel­lu­lar and acel­lu­lar agri­cul­ture speci­fi­cally).

    • Farmed An­i­mal Fun­ders has highly ten­ta­tive sug­ges­tions on how philan­thropists might al­lo­cate dona­tions/​fund­ing to plant-based al­ter­na­tives.

  • If you want to work in this space:

    • Most com­pa­nies are hun­gry for sci­en­tific and en­g­ineer­ing tal­ent and will con­tinue to hire.

      • Rele­vant dis­ci­plines/​skills for plant-based meat in­clude: bio­chem­istry, food sci­ence, plant biol­ogy, chem­i­cal en­g­ineer­ing.

      • Rele­vant dis­ci­plines/​skills for a/​cel­lu­lar agri­cul­ture in­clude: bio­chem­istry, food sci­ence, plant biol­ogy, chem­i­cal en­g­ineer­ing, tis­sue en­g­ineer­ing, syn­thetic biol­ogy, biore­ac­tor en­g­ineer­ing, cell cul­ture.

      • Soft­ware en­g­ineer­ing will prob­a­bly be­come more use­ful in au­toma­tion and com­pu­ta­tional mod­el­ling.

      • Ex­pe­rience re­quired varies: some ask for a few years of ex­pe­rience in a lab while oth­ers may only hire PhDs.

    • There are more job open­ings at plant-based com­pa­nies, al­though pub­lic in­ter­est is peak­ing, so com­pe­ti­tion is also high (so re­place­abil­ity might be a con­cern).

      • Th­ese are hugely varied, from work­ing on pro­duc­tion lines to op­er­a­tions, HR, data sci­ence, etc.

    • Non-sci­en­tific roles at acel­lu­lar and cel­lu­lar agri­cul­ture com­pa­nies will be more widely available as they ap­proach com­mer­cial­i­sa­tion, in 3-5 years.

    • Cel­lA­gri, GFI, and 80,000 Hours main­tain lists of job op­por­tu­ni­ties.

  • Stu­dents: GFI has a guide for stu­dents which con­tains ca­reer pro­files.

  • Re­searchers/​sci­en­tists: Aca­demic re­search seems valuable, partly be­cause it is of­ten open ac­cess while start-ups in the space tend to keep their re­search highly con­fi­den­tial[1]. GFI and New Har­vest are fund­ing most pro­jects in the space[2] (these links are a good place to look for in­sti­tu­tions and ad­vi­sors, alongside GFI’s aca­demic labs database).

  • In­vestors/​fun­ders: see ‘Notes for in­vestors/​fun­ders’.

  • En­trepreneurs: see ‘Notes for en­trepreneurs/​start-ups’.

Introduction

The in­creas­ing availa­bil­ity of tasty plant-based al­ter­na­tives in de­vel­oped coun­tries sug­gests fewer an­i­mals will have to suffer for food[3] and peo­ple are more able to ex­er­cise con­sump­tion choices.

But de­spite the ap­par­ent ex­plo­sion of Beyond Meat and Im­pos­si­ble Foods, the pro­por­tion of plant-based meat be­ing sold out of all meat[4] is only 1% in Europe and 2% in North Amer­ica.[5] The pro­por­tion of cul­ti­vated meat is 0%. And meat aside, have you ever seen a plant-based soft-boiled egg?

Alter­na­tives to tra­di­tional an­i­mal prod­ucts have a long way to go, es­pe­cially if the goal is 100% mar­ket share. How­ever, there is a lot of un­cer­tainty around when we will be able to eat meat grown from cells, and how we should di­vide our efforts be­tween that, plant-based al­ter­na­tives, and other forms of an­i­mal ad­vo­cacy. This post seeks to give sen­si­ble, un­bi­ased views on the fu­ture of al­ter­na­tive pro­teins by mak­ing semi-con­fi­dent pre­dic­tions about when we can see highly similar (to an­i­mal prod­ucts), cost-com­pet­i­tive prod­ucts from al­ter­na­tive pro­tein tech­nolo­gies, by both type of tech­nol­ogy and by type of product. It is in­tended for both tech­ni­cal and non-tech­ni­cal au­di­ences; if you want to un­der­stand how these tech­nolo­gies work, see ‘Fur­ther read­ing’.

I recom­mend read­ing the sum­mary and the rele­vant part of ‘Ac­tions you can take’ if you are in­di­rectly in­ter­ested, e.g. as an an­i­mal-welfare-fo­cused donor look­ing into al­ter­na­tive pro­tein de­vel­op­ment. If you are an in­sti­tu­tional in­vestor or sci­en­tist/​pro­fes­sional/​stu­dent look­ing to re­search or work in the space, I sug­gest read­ing the whole ar­ti­cle, or if you are already com­mit­ted to a product type/​tech­nol­ogy, just the parts rele­vant to that.

Definitions

Plant-based prod­ucts are made mainly from plants, though the term also in­cludes fungi (e.g. mush­rooms, Quorn), al­gae, and bac­te­ria. In this post, I gen­er­ally talk about plants. ‘Plant-based’ may mean veg­e­tar­ian, ve­gan, or nei­ther, but in gen­eral (in 2020) and in this post it tends to re­fer to ve­gan. There may be some an­i­mal use in­volved in these tech­nolo­gies (e.g. con­duct­ing an­i­mal tests for in­gre­di­ents), so I won’t call them ve­gan to avoid de­bates over ter­minol­ogy.

Cul­ti­vated meat is meat made from ac­tual an­i­mal cells, grown in vitro (out­side the or­ganism). This pro­cess is called cel­lu­lar agri­cul­ture.

In acel­lu­lar agri­cul­ture, the prod­ucts are not cells but pro­teins and fats. Cells or microbes (like yeast or bac­te­ria) are used as ‘fac­to­ries’ to pro­duce these. This gives us ac­tual dairy and eggs; just as the in­sulin we use to­day is pro­duced in bac­te­ria rather than live pigs.

Pro­cessed meat here refers to red meat, poul­try, or seafood prod­ucts that are rel­a­tively ho­moge­nous e.g. burg­ers, sausages, nuggets, some fish fingers. This ex­cludes meat prod­ucts where the tis­sue struc­tures have been main­tained, like ba­con.[6]

Whole mus­cle meat here refers to meat prod­ucts where the tis­sue struc­ture of meat is pre­served, e.g. chicken ten­ders, beef steaks.

I in­tend the cat­e­gories of ‘pro­cessed’ and ‘whole tis­sue meat’ to be use­ful for dis­t­in­guish­ing com­plex­ity and struc­ture, so they differ from con­ven­tional defi­ni­tions for pro­cessed meat.

The timelines given are for highly similar, cost-com­pet­i­tive prod­ucts. How­ever, ac­cessibil­ity to these prod­ucts will also de­pend on fac­tors like reg­u­la­tion, mar­ket­ing, and dis­tri­bu­tion.

Pre­dic­tions by tech­nol­ogy type

Timeline: When we can ex­pect highly similar cost-com­pet­i­tive al­ter­na­tives to an­i­mal products

Timeline

Plant-based

Since ~2018, a surge of in­no­va­tive uses of plant pro­teins (e.g. pea pro­tein burg­ers and mung bean scram­bled eggs) am­plified by savvy mar­ket­ing and dis­tri­bu­tion have led to rel­a­tively con­vinc­ing plant-based ver­sions of pro­cessed/​more ho­moge­nous an­i­mal prod­ucts. This strat­egy will con­tinue to be highly effec­tive for the next 5-10 years as there is much un­tra­versed ter­rain[7].

They will get even bet­ter in the next 10-20 years, es­pe­cially as they’re blended e.g. with col­la­gen (pro­duced with­out an­i­mals), or real meat cul­ti­vated from cells, alongside pro­cessed prod­ucts made wholly from cel­lu­lar agri­cul­ture (see be­low). In fact, Im­pos­si­ble Foods’ burger is tech­ni­cally a blended plant-based and acel­lu­lar agri­cul­ture product, as they add heme pro­duced with yeast, which is one of the com­po­nents that makes meat taste iron-rich and look red.

How­ever, it’s difficult to use plants to repli­cate ‘whole tis­sue’ an­i­mal flesh and more com­plex pre­sen­ta­tions of eggs and dairy. Plant fat, pro­tein, and cells tend to be sim­pler than the an­i­mal ver­sions, and meat de­rives its flavour from thou­sands of unique flavour molecules.

If we do man­age to make con­vinc­ing cost-com­pet­i­tive al­ter­na­tives for these, they will prob­a­bly emerge from acel­lu­lar and cel­lu­lar agri­cul­ture. Th­ese ap­pear to be ~10-20 and ~30-50 years away re­spec­tively (with good mar­gin for er­ror).[8]

Acel­lu­lar agriculture

Acel­lu­lar agri­cul­ture is already a known com­mer­cial pro­cess used for in­sulin, food en­zymes, and in­dus­trial en­zymes. The clos­est com­par­i­son for acel­lu­lar prod­ucts are food en­zymes, which lie in be­tween phar­ma­ceu­ti­cals and in­dus­trial en­zymes in terms of pu­rity, cost, and scale. Food en­zymes such as amy­lase are cur­rently sold for as lit­tle as $3/​kg, which is already com­pa­rable to dairy and eggs (note: I’m un­cer­tain about the re­li­a­bil­ity of this cost es­ti­mate). How­ever, to build a pro­cess for dairy and egg pro­teins, sig­nifi­cant work is still needed to bring costs down while scal­ing pro­duc­tion up, as well as prop­erly fold­ing, as­sem­bling, and har­vest­ing the var­i­ous pro­teins. This is due to dairy and egg pro­teins be­ing con­sumed in much larger quan­tities[9] and the ac­tual pro­teins be­ing larger and more bulky than with food en­zymes. [10]

Acel­lu­lar agri­cul­ture prod­ucts have already passed reg­u­la­tion in the US to be com­mer­cially sold:[11] Perfect Day has se­cured ap­proval from the Food and Drug Ad­minis­tra­tion for its whey pro­tein to be ‘Gen­er­ally Rec­og­nized As Safe’, and has sold small batches of ice cream.[12] This might not be as easy in other coun­tries, but the US is a ma­jor tar­get mar­ket for many of these com­pa­nies.

Cel­lu­lar agriculture

While the ba­sic tech­nique of cel­lu­lar agri­cul­ture is an es­tab­lished field in sci­ence, the lab pro­cess of mul­ti­ply­ing hu­man cells for or­gan trans­plant[13] is very differ­ent from com­mer­cial­is­ing the large-scale fer­men­ta­tion of an­i­mal cells for food, in terms of costs and pro­cess[14]. Areas to be ad­dressed in­clude au­toma­tion, in­vent­ing new pro­duc­tion meth­ods, and trans­form­ing parts of the life sci­ence sup­ply chain.

One of the key challenges in cel­lu­lar agri­cul­ture is cell cul­ture me­dia. At cur­rent costs, GFI es­ti­mates[15] that cell cul­ture me­dia would ac­count for c.55-95% of the cost of cul­tured meat pro­duced at scale. Sig­nifi­cant time and in­vest­ment are needed to bring cell cul­ture me­dia from a few hun­dred dol­lars per litre down to <$1/​litre, which is ap­prox­i­mately the price needed for cost-com­pet­i­tive cul­ti­vated meat. Ad­di­tional hur­dles that need to be over­come in­clude cells used, scaf­fold­ing, biore­ac­tors, and reg­u­la­tion. Reg­u­la­tion might be harder for cel­lu­lar agri­cul­ture prod­ucts than acel­lu­lar agri­cul­ture prod­ucts, es­pe­cially as the USDA will also be in­volved in reg­u­lat­ing cul­ti­vated meat in the US.

How­ever, de­vel­op­ing pro­cessed meat prod­ucts is much eas­ier than build­ing whole tis­sue prod­ucts. In sum­mary: the com­plex­ity is com­pounded for whole tis­sue prod­ucts.

  • A key differ­ence is that in pro­cessed prod­ucts, small tis­sue pieces are grown on scaf­folds, through which nu­tri­ents, oxy­gen, and waste can diffuse. In com­par­i­son, whole tis­sue prod­ucts re­quire the de­vel­op­ment of vas­cu­la­ture, i.e. chan­nels that carry nu­tri­ents, oxy­gen, and waste to and from cells, like blood ves­sels.

  • Pro­cessed prod­ucts may re­quire just a few types of mus­cle and fat cells, while whole tis­sue prod­ucts could re­quire many more cell types, which each re­quire op­ti­mi­sa­tion and ex­per­i­men­ta­tion.

Each challenge of a/​cel­lu­lar agri­cul­ture can take months or years to ex­plore and perfect,[16] and this may have to be re­peated to find the op­ti­mal com­bi­na­tion of vari­ables. Cost-com­pet­i­tive dairy and eggs from acel­lu­lar agri­cul­ture are at least a few years away and pro­cessed meats from cel­lu­lar agri­cul­ture will prob­a­bly take at least a decade. With the com­pounded com­plex­ity of whole tis­sue cel­lu­lar agri­cul­ture, this will likely take at least twice as long – while whole tis­sue is many times more com­plex than pro­cessed meat prod­ucts, this can be miti­gated with in­creas­ing ex­per­tise brought about by ex­pe­rience and greater sci­en­tific in­ter­est, a more ma­ture sup­ply chain, fund­ing in­creases fol­low­ing proof of con­cept, and ex­ist­ing reg­u­la­tory ap­provals. For more de­tail on costs, see the ap­pen­dices.

Pre­dic­tions by product type

Table: Distinc­tions be­tween an­i­mal product types

Table

Caveats to the table

I am un­cer­tain about the dis­tinc­tions be­tween ‘Lower’ and ‘Medium’ difficulty. For ex­am­ple, it might have taken a long time to de­velop good scram­bled egg al­ter­na­tives sim­ply be­cause there isn’t much de­mand for them (since chicken eggs tend to be seen as healthy and not too bad for the en­vi­ron­ment).

‘Ap­prox­i­mate share of to­tal vol­ume’[17] refers to the cur­rent split of tra­di­tional an­i­mal prod­ucts as of 2018. How­ever, I am very un­cer­tain about their ac­cu­racy and I was un­able to tri­an­gu­late them. If any­one can cor­rob­o­rate the ac­cu­racy/​in­ac­cu­racy of any of these figures, please let me know. Fur­ther­more:

  • In this data, the defi­ni­tions of ‘pro­cessed’ and ‘whole tis­sue’ meat are from the meat pro­cess­ing in­dus­try. The key differ­ence is that ‘pro­cessed’ is a larger cat­e­gory here and in­cludes any meat prod­ucts with added in­gre­di­ents, so for ex­am­ple chicken ten­ders and meat in sauce fall un­der ‘pro­cessed’. This is some­what bal­anced out by the lower pro­por­tion of meat in pro­cessed prod­ucts (vol­umes are for the to­tal foods).

  • I used North Amer­ica and Western Europe to rep­re­sent ‘de­vel­oped coun­tries’. Devel­oped coun­tries in Asia may have differ­ent pro­por­tions, but it’s harder to gen­er­al­ise across them.

#1, #2, #3 – Beef, pork, chicken

Very good ver­sions of plant-based burger pat­ties, sausages, and meat­balls ex­ist. How­ever, in or­der to repli­cate the full sen­sory and visceral ex­pe­rience of eat­ing an­i­mal flesh, it’s un­likely that we can get the same tex­ture, taste, aroma, ap­pear­ance, and mouth­feel wholly from plants. If we do, it’s prob­a­bly as difficult as cul­ti­vat­ing ac­tual meat cells, so we might as well do that. (I have lower con­fi­dence in this of c.30%. It’s quite pos­si­ble that we’ll dis­cover novel plant in­gre­di­ents or pro­cess­ing tech­niques that hugely in­crease similar­ity.)

While there is less re­search on avian cell cul­ture than mam­malian, chick­ens are rea­son­ably close to mam­mals so some of the re­search is trans­fer­able. Fur­ther­more, chicken cells may be even eas­ier to cul­ture than mam­malian cells, so chicken prod­ucts from cel­lu­lar agri­cul­ture may be closer to mar­ket than beef and pork. This is be­cause a lot of time can be saved in ‘dis­cov­ery’ work, de­vel­op­ing new scaf­fold­ing ma­te­ri­als, new meth­ods for iso­la­tion and cul­ture, etc.[18]

#4 – Seafood (fish and aquatic in­ver­te­brates)

Seafood is similarly difficult to repli­cate with plants, but cul­ti­vated seafood also comes with a set of challenges. The set-up costs are higher in terms of time and money: there’s much less sci­en­tific liter­a­ture on non-mam­malian cell cul­ture and cul­tur­ing cells of fish and in­ver­te­brates is very differ­ent from grow­ing pig mus­cle. So re­searchers and com­pa­nies have less of a knowl­edge base about the cell cul­ture me­dia, growth fac­tors, and scaf­folds to use. How­ever, fish and in­ver­te­brate cells may ac­tu­ally be eas­ier to cul­ture than mam­malian and avian cells, be­cause of higher tol­er­ance to low lev­els of oxy­gen, low tem­per­a­tures, and changes in pH.

#5 – Dairy

Milk is the star of the al­ter­na­tive pro­teins move­ment. Sim­pler prod­ucts like cream and yo­ghurt are fol­low­ing in its foot­steps, but the plant-based ver­sions of more com­plex prod­ucts like cheese and but­ter are not grow­ing as much. Ve­gan cheese is no­to­ri­ously bad. This is be­cause cheese is com­plex, much more so than milk: cheese in­volves a great va­ri­ety of bac­te­rial cul­tures and con­sumers have nu­mer­ous taste and tex­ture re­quire­ments (e.g. melt, stretch). It’s difficult to catch up to cheese’s head start of mil­len­nia of ex­per­i­men­ta­tion on bac­te­rial cul­tures and dairy.[19]

How­ever, this might ac­tu­ally make acel­lu­lar cheeses and yo­ghurts eas­ier to pro­duce than milk, be­cause the microbes do more work in pro­duc­ing flavour than the ac­tual an­i­mal com­po­nents them­selves.

#6 – Eggs

There is a strong eth­i­cal case for non-an­i­mal al­ter­na­tives to eggs. There may be some com­pa­nies that can suc­cess­fully repli­cate cer­tain in­stances, like JUST’s scram­bled eggs, but eggs in gen­eral will prob­a­bly be more suc­cess­fully pro­duced through acel­lu­lar agri­cul­ture. While eggs are ex­tremely ver­sa­tile in­gre­di­ents, the ac­tual pro­teins and fats are ac­tu­ally fairly sim­ple, as an­i­mal pro­teins go, and are thus rel­a­tively easy to pro­duce via acel­lu­lar agri­cul­ture. And if you repli­cate the ex­act pro­teins of an egg, you can sub­sti­tute it for all ap­pli­ca­tions. In con­trast, you need heaps of plant-based sub­sti­tutes: one to fry, one for mer­ingues, an­other for emul­sifi­ca­tion, and an­other for bind­ing.

Devel­op­ing biolog­i­cally iden­ti­cal egg pro­teins also has an ad­van­tage in a key area of re­duc­ing an­i­mal suffer­ing, in func­tional in­gre­di­ent re­place­ment. Ad­di­tion­ally, eggs are cur­rently per­ceived as benefi­cial for hu­man health, so con­sumers are less ea­ger for plant-based sub­sti­tutes.

#7 – Pet food

Plant-based pet food has grown more slowly than plant-based hu­man food. Con­sumers need to put more thought and effort into switch­ing cats and dogs into veg­e­tar­ian/​ve­gan food, be­cause of their biolog­i­cal needs, less di­verse diet, and the difficulty in notic­ing how it af­fects their health. As a re­sult, only the sec­ond most hard­core veg­e­tar­i­ans/​ve­g­ans have put their pets onto plant-based diets. (The most hard­core don’t have pets.)

So, while pet food is more pro­cessed and it’s eas­ier to de­velop plant-based sub­sti­tutes that match the tex­ture, taste, aroma, and ap­pear­ance, and in fact have low pro­por­tions of an­i­mal meat any­way, it will be difficult to get con­sumers to switch to 100% plant-based prod­ucts. This is par­tic­u­larly true for pre­mium wet food which has ‘whole tis­sue’ an­i­mal flesh, which is a low-vol­ume, high-value product that can only be sub­sti­tuted with cel­lu­lar agri­cul­ture.

Pet food in gen­eral is a large op­por­tu­nity, as dogs and cats con­sume a sig­nifi­cant pro­por­tion of an­i­mal meat. In the US, they ac­count for about a third of all calories from an­i­mals. Three caveats: first, a lot of pet food is made of by-prod­ucts that hu­mans don’t eat, so switch­ing away from these will have lit­tle effect on de­mand for an­i­mal meat. Se­cond, pets out­side the US eat less meat be­cause the US pet food mar­ket is much more pre­mium-fo­cused. Third, I have medium to low con­fi­dence in the method­ol­ogy of this par­tic­u­lar study.

#8 – Insects

The mar­ket for in­sects as a food is small[20]. It is grad­u­ally grow­ing in the de­vel­oped world amidst con­sumer reser­va­tions. How­ever, there are po­ten­tially large eth­i­cal con­cerns around rais­ing live in­sects for food and we have no plant-based in­sects be­cause in­sects are not pop­u­lar enough.

A pos­si­ble al­ter­na­tive is cel­lu­lar agri­cul­ture: a team of sci­en­tists at Tufts is work­ing on in­sect cel­lu­lar agri­cul­ture (Kyle Fish is on the team and spoke about their work at EA Global in 2018). Like with fish, in­sect cells are much eas­ier to cul­ture than mam­mal and bird cells. One of the main ad­van­tages seems to be in ap­ply­ing learn­ings from land in­ver­te­brate to aquatic in­ver­te­brate cell cul­ture, as the former is more es­tab­lished than the lat­ter, e.g. there are hun­dreds of im­mor­tal­ised in­sect cell lines, but no such crus­tacean cell lines.

There are no com­pa­nies work­ing on in­sect cel­lu­lar agri­cul­ture com­mer­cially, prob­a­bly be­cause of the difficulty of as­cer­tain­ing con­sumer de­mand and rais­ing fund­ing for it. I am similarly un­cer­tain and don’t have a view on the timeline for com­mer­cially available in­sect prod­ucts from cel­lu­lar agri­cul­ture.

#9 – Bonus: ex­pen­sive, low-vol­ume products

We’ll see cost-com­pet­i­tive ver­sions of high-value, low-vol­ume cel­lu­lar agri­cul­ture prod­ucts sooner, like foie gras and fish maw, be­cause it’s eas­ier to achieve their high prices. The an­i­mal welfare benefit from these is lower, but they could be very valuable for proof of con­cept and rev­enue and fund­ing gen­er­a­tion.

It may also be good for cel­lu­lar agri­cul­ture prod­ucts to start out as ex­pen­sive and as­pira­tional, then be­come cheaper, in terms of build­ing con­sumer per­cep­tions.

Notes for in­vestors/​funders

As a re­minder, the timelines given are for highly similar, cost-com­pet­i­tive prod­ucts. Com­pa­nies will see rev­enue much ear­lier and most start-ups will look to tran­si­tion from a high-value low-vol­ume strat­egy to a low-value high-vol­ume one. Ad­di­tion­ally, in­cre­men­tal in­vest­ment and ac­tivity in the space would bring these timelines for­ward.

In­sti­tu­tional in­vestors less fa­mil­iar with the space may find this use­ful as a start­ing point for think­ing about their risk ap­petite, time scale for in­vest­ment, and the kind of com­pany they want to in­vest in (e.g. B2C or B2B).

This can also be valuable for con­sid­er­ing the com­pet­i­tive dy­nam­ics be­tween an­i­mal prod­ucts made from plants, an­i­mal cells, and live an­i­mals, par­tic­u­larly the risk of can­ni­bal­i­sa­tion. Cur­rently, a sig­nifi­cant pro­por­tion of the in­vest­ment in the space is from im­pact in­vestors, meat pro­ces­sors, and gi­ant in­vestors like sovereign wealth funds. As more main­stream in­vestors move into the space, biotech and pharma ven­ture cap­i­tal firms may be the first to find they are a good fit as in­vestors be­cause of their fa­mil­iar­ity with ex­tended pre-rev­enue pe­ri­ods in­volv­ing high R&D costs and risks.

For reg­u­lar in­di­vi­d­ual in­vestors, there is limited op­por­tu­nity to get in­volved be­cause there are few pure play­ers that in­di­vi­d­u­als can buy shares in – the only ones I am aware of are Beyond Meat and Tofutti (trades over the counter). Buy­ing shares in food com­pa­nies that are de­vel­op­ing al­ter­na­tive pro­tein prod­ucts is un­likely to en­courage them in this spe­cific di­rec­tion. I am also scep­ti­cal of im­pact in­vest­ing in this way be­cause it does not seem likely to have sub­stan­tial im­pact (ex­plained in #2 here). While Beyond In­vest­ing has a ve­gan in­dex fund, it is tar­geted more at ve­g­ans want­ing to ex­clude com­pa­nies en­gaged in ‘an­i­mal ex­ploita­tion, defense, hu­man rights abuses, fos­sil fuels ex­trac­tion and en­ergy pro­duc­tion, and other en­vi­ron­men­tally dam­ag­ing ac­tivi­ties’.

Notes for en­trepreneurs/​start-ups

The views pre­sented are on a long-term mar­ket level and have limited use­ful­ness for de­ter­min­ing the next start-up that should be founded. For ex­am­ple, a plant-based pork pro­ducer is prob­a­bly go­ing to see a lot more rev­enue than a cul­ti­vated fish com­pany in the next five years.

Scott Weathers et al posted on the EA fo­rum about their re­search into a prospec­tive al­ter­na­tive meat start-up. They con­sid­ered var­i­ous pos­si­bil­ities in plant-based and cul­ti­vated meat and ul­ti­mately de­cided against all of them, mainly be­cause of per­sonal fit – their team’s lack of tech­ni­cal ex­per­tise and com­par­a­tive ad­van­tage el­se­where. This is also a short-term view look­ing at the marginal im­pact of one more start-up. They had a struc­tured ap­proach and I recom­mend read­ing through it for any­one con­sid­er­ing start­ing a com­pany in this area.

Char­ity En­trepreneur­ship has a re­port on food tech­nol­ogy as an in­ter­ven­tion. This is a more short-term view as­sess­ing the marginal im­pact of food tech­nol­ogy and I largely agree with their con­clu­sion that plant-based seafood is po­ten­tially very high-im­pact, but best left to the pri­vate sec­tor. I think they are overly con­cerned about re­place­abil­ity and com­pe­ti­tion, given ex­tremely low lev­els of pen­e­tra­tion of plant-based prod­ucts at the mo­ment.

One thing to note is that the al­ter­na­tive pro­tein land­scape may ap­pear less promis­ing in a few years, but the level of risk will not be too dis­similar. Most start-ups fail, and with the re­cent surge in new start-ups across these tech­nolo­gies, there will prob­a­bly be a wave of failures in the next few years. This will hap­pen as start-ups run out of fund­ing, which will be more pro­nounced for cul­ti­vated meat com­pa­nies which re­quire greater up-front in­vest­ment be­fore gen­er­at­ing rev­enue.

Limi­ta­tions and ar­eas for fur­ther research

I’m rea­son­ably con­fi­dent about my pre­dic­tions in gen­eral (c.40-60%, but I’m also un­con­fi­dent in my level of con­fi­dence).Th­ese es­ti­mates are not set in stone and fac­tors like in­vest­ment and ac­tivity would quicken progress. How­ever, the di­rec­tion in which I ex­pect to be wrong is in things tak­ing more rather than less time, be­cause things will prob­a­bly take longer than we ex­pect, for un­ex­pected rea­sons, and we don’t know ev­ery­thing that we don’t know.

A key limi­ta­tion is that my method­ol­ogy for pre­dicted timelines is not based on me­thod­i­cally as­sess­ing in­puts, but rather in­tu­itive weight­ing of sources (see ‘Sources’).

Other rea­sons why my pre­dic­tions will be wrong/​in­com­plete in­clude:

  • There will be ex­cep­tions, es­pe­cially as the al­ter­na­tive pro­tein mar­ket con­verges with phar­ma­ceu­ti­cals, where sin­gle dis­cov­er­ies or in­no­va­tions can have out­size im­pact on a com­pany’s suc­cess.

  • Cel­lu­lar and/​or acel­lu­lar agri­cul­ture could fail or be de­layed by decades, e.g. if a safety is­sue sets back con­sumer con­fi­dence and hence gov­ern­ment reg­u­la­tion and in­vestor ap­petite, or if it just takes ex­tremely long to re­duce costs and scale up

  • Some prod­ucts will be very suc­cess­ful as blended prod­ucts, e.g. col­la­gen (acel­lu­lar agri­cul­ture) with soy pro­tein (plant-based).

    • This may not be wholly good. Big meat pro­ces­sors such as Tyson and Pur­due have re­leased blended prod­ucts. Based on pri­vate con­ver­sa­tions with one of these com­pa­nies, one of their chief ob­jec­tives is ad­di­tional in­cre­men­tal de­mand for an­i­mal pro­tein, while ac­cess­ing the flex­i­tar­ian mar­ket – that is, they are look­ing to sup­plant fully plant-based/​an­i­mal-free prod­ucts like Beyond and Im­pos­si­ble, rather than an­i­mal meat prod­ucts.

  • There are other cat­e­gories I haven’t con­sid­ered, such as fun­gus (Quorn).

  • There will be re­gional differ­ences and these could change a lot in the next 50 years.

    • Cur­rently, these tech­nolo­gies are con­cen­trated in the US, Europe, Is­rael, and cer­tain coun­tries in Asia (Sin­ga­pore, China, Ja­pan). An­i­mal ad­vo­cacy and tech­nolog­i­cal ca­pa­bil­ity may need to grow fur­ther to foster the cli­mate needed.

    • Reg­u­la­tion will be eas­ier and/​or quicker in some coun­tries than oth­ers.

    • Coun­tries have differ­ent at­ti­tudes to ge­netic en­g­ineer­ing, which could greatly im­pact on how quickly these tech­nolo­gies can ad­vance. Food com­pa­nies in Europe and Is­rael are mostly avoid­ing ge­netic en­g­ineer­ing,[21] while those in the US are ex­ploit­ing it fully.

    • ‘Cost-com­pet­i­tive’ is differ­ent for each coun­try.

  • The prices of an­i­mal prod­ucts could change over the next few decades which would af­fect the defi­ni­tion of ‘cost-com­pet­i­tive’. This could go in ei­ther di­rec­tion.

  • Be­cause a/​cel­lu­lar agri­cul­ture is such a nascent field, a lot of things could change, e.g. some of the tech­niques and re­quire­ments dis­cussed here could be ob­so­lete in a few years.

I wrote this mainly be­cause I don’t want to hurt an­i­mals but I want to eat an­i­mal prod­ucts. I have sought to bal­ance out my in­di­vi­d­ual yearn­ing for these tech­nolo­gies.

Fur­ther read­ing/​listen­ing/​watching

Even fur­ther reading

Th­ese sources will be more com­pre­hen­si­ble and use­ful if you have a rele­vant tech­ni­cal back­ground. I also recom­mend them as a start­ing point to un­der­stand­ing the sci­ence be­hind cel­lu­lar agri­cul­ture.

Sources

My time es­ti­mates are mostly based on pri­vate con­ver­sa­tions with se­nior staff at plant-based and cel­lu­lar agri­cul­ture com­pa­nies. Th­ese have come about through at­tend­ing con­fer­ences and the pro­cess of writ­ing a re­port on low-cost cell cul­ture me­dia and growth fac­tors for GFI in 2018, in which I in­ter­viewed and sur­veyed cul­ti­vated meat start-ups about their com­mer­cial­i­sa­tion timelines. I’ve weighted aca­demic and pri­vate set­tings and sources more heav­ily than busi­ness and pub­lic con­texts, as the lat­ter have strong fi­nan­cial and mis­sion in­cen­tives to be overly op­ti­mistic (the more op­ti­mistic they are about their timelines, the more fund­ing they will get, so the more their op­ti­mism will be­come re­al­ity)[22].

The main char­i­ties in­volved in a/​cel­lu­lar agri­cul­ture, GFI and New Har­vest, used to give for­mal es­ti­mates for timelines, but have done this less since c.2018. How­ever, they have pub­lished re­search about the pre­cise challenges in cel­lu­lar and acel­lu­lar agri­cul­ture. GFI’s pa­per on cell cul­ture me­dia is par­tic­u­larly use­ful for this as it pro­vides in­for­ma­tion on the ac­tual costs that need to be re­duced to bring cell cul­ture me­dia down to <$1/​litre, which is ap­prox­i­mately the price needed for cost-com­pet­i­tive cul­ti­vated meat.

I’ve also taken into ac­count re­ports on cul­ti­vated meat timelines from An­i­mal Char­ity Eval­u­a­tors (2017) and Open Philan­thropy (2015). Since the pub­li­ca­tion of these, more start-ups have been launched, more fund­ing has flowed into the space, and more clar­ity on fu­ture timelines has sur­faced. Fur­ther­more, their sources gen­er­ally fall in three cat­e­gories: pub­lic es­ti­mates from cul­tured meat pro­po­nents (Mosa Meat, GFI), out­dated pub­lished aca­demic es­ti­mates, and pri­vate con­ver­sa­tions with sci­en­tists who are not work­ing on cul­tured meat. Th­ese are all likely to be very in­ac­cu­rate in differ­ent di­rec­tions and ways, which is why the es­ti­mates given are so broad and un­cer­tain.

Thanks to com­ments from Frea Me­hta, Sa­muel Hil­ton, and Saulius Šimčikas.

Ap­pendix 1 – In­ter­pret­ing a/​cel­lu­lar agri­cul­ture cost estimates

Fal­ling cost es­ti­mates are com­monly used to as­sess the progress of a/​cel­lu­lar agri­cul­ture re­search. ACE writes:

‘A num­ber of ground meat prod­ucts have re­port­edly been suc­cess­fully cul­tured. As a gen­eral proof of con­cept, in 2013 a ground beef burger was cul­tured as part of a pro­ject re­port­edly cost­ing $330,000. A few days af­ter that proof of con­cept, the pro­ject head (Mark Post) re­ported that with cur­rent tech­nol­ogy, cul­tured meat could be pro­duced at a cost of $70/​kg—which we be­lieve was in­tended as the es­ti­mated cost if us­ing mass pro­duc­tion tech­niques. In June 2016 Mem­phis Meats re­port­edly cul­tured a meat­ball at a cost of $40,000/​kg. In March 2017 Mem­phis Meats then re­port­edly cul­tured a chicken product at a cost of $20,000/​kg.’

How­ever, large jumps in cost could be in­fluenced by labour costs and mov­ing from bench-scale to larger-scale pro­duc­tion (i.e. buy­ing lab sup­plies in bulk), which may not re­flect true pro­cess im­prove­ments. Mark Post’s burger’s $330,000 price tag was so high be­cause it in­cluded the ‘three months’ of work from ‘three lab­o­ra­tory tech­ni­ci­ans’ and was done at bench scale. I haven’t seen cost break­downs from Mem­phis Meats and other com­pa­nies.

Ap­pendix 2 – How to bring the cost of cel­lu­lar agri­cul­ture down, exactly

Ac­cord­ing to GFI’s anal­y­sis, the cost re­duc­tions needed in cell cul­ture me­dia are (in de­creas­ing or­der of mag­ni­tude):

  • Re­duce use of growth fac­tors (a pro­tein or sig­nal­ling hor­mone that tells cells what to do). Growth fac­tors make up c.90% of the cost of an­i­mal-free me­dia cur­rently used.

  • Scale up pro­duc­tion of growth fac­tors to re­duce costs.

  • Pro­duce growth fac­tors through re­com­bi­nant ex­pres­sion (like acel­lu­lar agri­cul­ture).

  • Source food grade rather than phar­ma­ceu­ti­cal grade basal medium components

  • Sub­sti­tute more ex­pen­sive basal medium com­po­nents with cheaper al­ter­na­tives, e.g. re­plac­ing ascor­bic acid-2-phos­phate with food-grade ascor­bic acid.

Other pos­si­ble av­enues are:

  • Find in­no­va­tive al­ter­na­tives to growth fac­tors, e.g. pep­tide sub­sti­tutes such as pep­ti­domimet­ics, or molecules from the fungi or an­i­mal king­dom such as plant hy­drolysates, which should be eas­ier and cheaper to pro­duce at scale.

  • Use an­i­mal in­stead of hu­man growth fac­tors.

  • Col­lab­o­rate with life sci­ence sup­pli­ers.

  • In­crease rate of me­dia re­cy­cling.

Ad­di­tional hur­dles that need to be met in­clude:

  • Cells: de­ter­mine cell lines and/​or types of cells to use.

  • Scaf­fold­ing: test types of scaf­fold­ing and meth­ods of har­vest­ing cells from scaf­fold­ing.

  • Biore­ac­tors: build or buy biore­ac­tors needed, stim­u­late cells to de­velop mus­cle.

  • Labour: au­to­mate pro­cesses.

  • Reg­u­la­tion: main­tain and demon­strate food safety, sub­mit de­tailed pro­cesses to reg­u­la­tors.

  • Op­ti­mise each vari­able in re­la­tion to ev­ery other vari­able.

Each part of this can take months to ex­plore and perfect, es­pe­cially in a cost-effec­tive and large-scale way, and this may have to be re­peated to find the op­ti­mal com­bi­na­tion of vari­ables. Cost-com­pet­i­tive dairy and eggs from acel­lu­lar agri­cul­ture and pro­cessed meats from cel­lu­lar agri­cul­ture are prob­a­bly at least a decade away.

As a sense check, I took each of the above as a 6-month pro­cess and added them all up, re­sult­ing in 6 x (5 + 4 + 10) = 114 months = 9.5 years. This is a very rough es­ti­mate and of course progress will be made con­cur­rently, but my list is also non-ex­haus­tive.

With the com­pounded com­plex­ity of whole tis­sue cel­lu­lar agri­cul­ture, this will likely take at least twice as long – while whole tis­sue is many times more com­plex than pro­cessed meat prod­ucts, this can be miti­gated with in­creas­ing ex­per­tise brought about by ex­pe­rience and greater sci­en­tific in­ter­est, a more ma­ture sup­ply chain, fund­ing in­creases fol­low­ing proof of con­cept, and ex­ist­ing reg­u­la­tory ap­provals.

Ap­pendix 3 – How re­li­ably can we ac­tu­ally pre­dict cel­lu­lar agri­cul­ture?

‘There has been a flurry of in­ter­est – and not a lit­tle in­cre­dulity – about claims, of­ten made by com­pa­nies backed by billion­aires and run by bold sci­en­tists, that mar­ket-ready ______________ were just around the cor­ner.’

Cel­lu­lar agri­cul­ture prod­ucts or nu­clear fu­sion re­ac­tors?

The quote is ac­tu­ally about nu­clear fu­sion, which has con­tinu­ally been pre­dicted to be 20-30 years away for decades, be­cause large break­throughs are re­quired for one ma­jor out­come.[23] In 2030, will we still be say­ing cel­lu­lar agri­cul­ture is 10-20 /​ 30-50 years away?

In some ways, this could hap­pen. Every­thing might be fal­ling into place ex­cept for one big bit, like reg­u­la­tion or au­toma­tion.

But in other ways, cel­lu­lar agri­cul­ture is closer to some­thing like tran­sis­tors, where a steady pace of small in­no­va­tions dis­tributed across nu­mer­ous en­tities con­tributes to lin­ear growth.

Ba­si­cally, this is harder to pre­dict than pro­cess­ing speed in­creases, but not as hard to pre­dict as nu­clear fu­sion, be­cause cel­lu­lar agri­cul­ture lies be­tween them in terms of size of break­throughs needed, dis­tri­bu­tion across in­sti­tu­tions/​com­pa­nies, and the im­pact of one com­pany’s suc­cess on the rest of the in­dus­try.

Notes


  1. The goal is not for all re­search in this space to be open ac­cess: the main draw­back of open ac­cess re­search in this con­text is that it can pre­vent com­pa­nies from patent­ing spe­cific tech­nolo­gies, which may dis­cour­age fund­ing from pri­vate in­vestors. The ideal mix is prob­a­bly for foun­da­tional re­search to be open ac­cess and spe­cific in­no­va­tive tech­niques to be re­searched pri­vately. ↩︎

  2. Mosa Meat has also hired PhD can­di­dates to con­duct their re­search at Maas­tricht Univer­sity while be­ing em­ployed by Mosa Meat. ↩︎

  3. This is more likely to be true for the new wave of more main­stream plant-based foods, no­tably from Beyond and Im­pos­si­ble, whose cus­tomer base is 90-95% non-ve­gan, so they are likely re­plac­ing an­i­mal prod­ucts with non-an­i­mal prod­ucts. ↩︎

  4. An­i­mal and plant-based meat across fresh, frozen, and shelf-sta­ble. Plant-based meat in­cludes meat sub­sti­tutes such as veg­e­tar­ian sausages, bean burg­ers, and veg­e­tar­ian burg­ers made of tofu, soy, or my­co­pro­tein, as well as tra­di­tional prod­ucts such as tofu and tem­peh. Plant-based meat pen­e­tra­tion is likely to be over­es­ti­mated/​some­what un­cer­tain as some of these prod­ucts are not always true sub­sti­tutes for meat. ↩︎

  5. Based on mar­ket data, source on re­quest. ↩︎

  6. When I dis­cuss cel­lu­lar agri­cul­ture, the defin­ing char­ac­ter­is­tic of ‘pro­cessed meat’ here is its re­li­ance on ad­di­tional scaf­fold com­po­nents for struc­ture. In con­trast, to grow whole tis­sue meat in cel­lu­lar agri­cul­ture, the struc­ture comes from the in­ter­play of the cells them­selves. ↩︎

  7. GFI has iden­ti­fied nu­mer­ous un­ex­plored ar­eas in plant-based meat. Th­ese in­clude us­ing novel plant sources, im­prov­ing the farm­ing and flavour prop­er­ties of plants for plant-based meat, and fat en­cap­su­la­tion. ↩︎

  8. See ap­pen­dices for more de­tails on costs. ↩︎

  9. Ac­cord­ing to GFI, en­zymes can make up just <0.1% of a food product (e.g. bread) by vol­ume, while an egg pro­duced with acel­lu­lar agri­cul­ture would need c.10% egg white pro­tein. ↩︎

  10. Ac­cord­ing to GFI, these food pro­teins are more bulky and com­plex, which makes them more difficult to as­sem­ble and fold in a host cell. Fur­ther­more, they tend to ac­cu­mu­late within the host cell rather than be­ing se­creted, which makes cell har­vest­ing more difficult. How­ever, an ad­van­tage is that un­like en­zymes, these food pro­teins are in­ert, so pro­duc­ers don’t have to worry about host cells be­ing af­fected by the pro­tein. ↩︎

  11. A com­pany called Gel­tor already sells col­la­gen pro­duced with acel­lu­lar agri­cul­ture for skin­care (since 2017). They plan to move into col­la­gen and gelatin for food. ↩︎

  12. This ap­proval is given for a spe­cific product pro­duced un­der a spe­cific pro­cess. This does not guaran­tee that other acel­lu­lar prod­ucts or acel­lu­lar egg pro­teins will be ap­proved. ↩︎

  13. This is also a dis­tant pos­si­bil­ity, though it is more challeng­ing to grow a func­tion­ing or­gan than it is to grow a slab of flesh for food. ↩︎

  14. GFI has iden­ti­fied the key parts to op­ti­mise for cost and scale in cel­lu­lar agri­cul­ture: cell cul­ture me­dia, cell lines, biore­ac­tors, scaf­folds. Th­ese are also in­ter­de­pen­dent (e.g. the cell lines used im­pact the cell cul­ture me­dia needed), com­pound­ing the time and re­sources needed for ex­per­i­men­ta­tion. ↩︎

  15. GFI’s pa­per on cell cul­ture medium: https://​​www.gfi.org/​​what-in­no­va­tions-are-needed-to-make-cell ↩︎

  16. We might never be able to over­come some challenges. How­ever, for most things, re­searchers could prob­a­bly change their ap­proach to sub­vert them in­stead. For ex­am­ple, many cel­lu­lar agri­cul­ture com­pa­nies have been work­ing out how to use less cell cul­ture me­dia al­to­gether, so re­duc­ing its costs isn’t needed as much. ↩︎

  17. Source on re­quest. ↩︎

  18. Ad­di­tion­ally, avian cells tend to be more re­spon­sive to en­vi­ron­men­tal cues, so tri­als can be run more quickly. With cow cells, for ex­am­ple, stan­dard meth­ods for main­tain­ing fast-di­vid­ing stem cells and differ­en­ti­at­ing them into mus­cle cells re­sult in high cell wastage, which in­creases me­dia costs. Efforts to im­prove on this would re­quire a lot of time and en­ergy. ↩︎

  19. Some­thing that puz­zles me, how­ever, is the lack of shared culi­nary knowl­edge around al­mond milk prod­ucts in Europe to­day. Al­monds were a com­mon sub­sti­tute for dairy in the Mid­dle Ages, for two rea­sons: Lent and the short shelf life of dairy milk. Peo­ple used to make but­ter and cheese from al­mond milk, but this is now a lost art. ↩︎

  20. In­sects are eaten by many peo­ple around the world, but mainly in the global South, so the mar­ket is small in value terms. ↩︎

  21. This can also spawn in­no­va­tion, e.g. Fu­ture Meat is us­ing spon­ta­neously im­mor­tal­ised chicken fibrob­lasts. This means they are ‘se­ri­ally pas­sag­ing cells un­til a cell nat­u­rally de­vel­ops mu­ta­tions which make it con­tin­u­ous. The benefit of this method is that the mu­ta­tions are nat­u­rally oc­cur­ring, mean­ing it doesn’t yield GMO meat.’ ↩︎

  22. I don’t think this is a prob­lem, but an­i­mal ad­vo­cates should keep in mind that they are not the in­tended au­di­ence for these pub­lic es­ti­mates. The es­ti­mates have func­tions be­yond truth. ↩︎

  23. Based on con­ver­sa­tions with a plasma physi­cist. ↩︎