Flink, a quick-commerce startup out of Berlin that was an acquisition target of Gorillas, Getir, Amazon, and Gopuff, is spelling out how it plans to go forth on its own. TechCrunch has exclusively learned that the company has raised $150 million, which it will use to double down on business in Germany and the Netherlands in partnership with Just Eat Takeaway.com.
These days, when you hear about students and generative AI, chances are that you’re getting a taste of the debate over the adoption of tools like ChatGPT. Are they a help? (Yay! Great for research! Fast!) Or are they a harm? (Boo! Misinfo! Cheating!). But some startups are taking the arrival of generative AI in the school environment as a positive, and as a foregone conclusion. And they are building products to meet what they believe will be a certain market opportunity.
Consolidation is here in cybersecurity, as bigger players in the space pick up startups that will help them grapple with the ever-expanding attack surface for enterprises as they move more activity into the cloud. In the latest development, CyberArk — one of the army of larger security companies founded out of Israel — is acquiring Venafi, a specialist in machine identity, for $1.54 billion.
CyberArk will pay $1 billion in cash and approximately $540 million in shares. Shareholders in both companies have approved the deal, which is expected to close in the second half of 2024, the companies said.
Venafi is majority-owned by Thoma Bravo, and was valued at $1.15 billion when the private equity firm bought its controlling stake in 2020. In other words, Venafi’s selling price today represents a moderate increase since 2020.
The news confirms rumors of a deal between the two companies that had been doing the rounds over the last few days.
CyberArk’s interest in Venafi comes at a time when security teams are trying to get a better and more holistic understanding of the threat landscape and attack surface of their organizations. In today’s market, that is an extremely complex puzzle to solve thanks to the growth of mobile technology, cloud services and distributed working.
In essence, all of these have led to an explosion of computing endpoints, which include not just the many devices that people might use to connect to a network, but any other device on the network where data is being processed or stored. The rule of thumb is that there are 40 “machines” for every human on an enterprise network. All this has led to a major surge of business for companies that focus on identity security. Some startups in the space have raised large amounts of money too — Oasis Security and Silverfort are good examples.
Venafi’s tech focuses on securing and understanding the flow of data between those machines.
The startup is described as a specialist in PKI and certificate management, and CyberArk says that the deal will expand its own total addressable market by $10 billion (to a total of $60 billion).
“This acquisition marks a pivotal milestone for CyberArk, enabling us to further our vision to secure every identity – human and machine – with the right level of privilege controls,” said Matt Cohen, CEO of CyberArk, in a statement. “By combining forces with Venafi, we are expanding our abilities to secure machine identities in a cloud-first, GenAI, post-quantum world. Our integrated technologies, capabilities and expertise will address the needs of global enterprises and empower Chief Information Security Officers to defend against increasingly sophisticated attacks that leverage human and machine identities as part of the attack chain.”
The acquisition also underscores some themes playing out among cybersecurity companies around consolidation.
Some companies that raised money several years ago at higher valuations are finding those valuations under pressure as they, variously, fail to grow ARR or reach profitability, and approach the predictable end of their runways.
Those companies are now looking for an exit, and sometimes that comes at a price far below their last valuations. For example, in recent weeks: Akamai acquired Noname Security for $450 million, less than half its last valuation; and Wiz tried to acquire Lacework, last valued at $8.3 billion, for just over $150 million, returning around $800 million in cash Lacework had in the bank to investors — that deal has fallen through.
On the other hand, a select few cybersecurity businesses are seeing significant growth right now, and are being earmarked as the consolidators. Wiz raised $1 billion a couple of weeks ago to fuel an acquisition spree, and CyberArk, which has a market cap of over $10 billion, is clearly another in this category.
The consolidation trend is playing out amongst even those being acquired. In May 2020, Venafi acquired Jetstack to bring Kubernetes expertise into its fold. Just a day before that, CyberArk had acquired Idaptive.
Early-stage rounds continue to account for the majority of investments in the European startup market, and on Tuesday one of the biggest firms in the region announced a new fund to bolster that trend. Accel has raised $650 million to back startups from seed to Series A across the U.K., the Continent and Israel. The fund is the eighth of its kind for Accel since it first put down roots in London in 2000.
Accel has invested in more than 200 startups in the region to date, making it one of the more prolific VCs in this market.
One of the recurring laments you hear in Europe is that even if the region produces exceptional talent and ideas, companies on the continent are challenged when it comes to scaling. There have been a number of exceptions over the years, however, that test that claim, and part of Accel’s gravity as an investor comes from the fact that it’s been a backer in a number of them. They include some of the most successful startups to come out of Europe, such as Skype, Supercell and Spotify (also incidentally a trio of Nordic startups, respectively hatched in Estonia, Finland and Sweden).
In the years since those investments, Accel’s bet has been that the growth of startups in Europe has been strong enough to grow the pot of money that it’s raising to back them. Notably, the $650 million announced Tuesday is the same size as the firm’s early-stage fund in the U.S. (announced December 2023). Given that the U.S. is a considerably bigger market in terms of overall venture funding and number of startups, that speaks to Accel’s confidence in what’s playing out here.
“The European tech scene has really come of age,” said Harry Nelis, a longtime partner at Accel in London. Current investments include cybersecurity firms Cyera and Oasis, the care home marketplace Lottie, and the buzzy AI video startup Synthesia, among many others.
As you might expect from that list and recent headlines, the focus going forward will be on timely businesses tapping into the needs and interests of the day. That includes those that are building creative solutions to pressing problems (cybersecurity being a prime example of that), smart commerce solutions (including marketplaces that tap into societal and social needs), and — need I write it? — AI, AI, AI.
Venture investing in Q1 of this year, according to research from PitchBook, shows slight but still encouraging signs of recovery. In total some €16.3 billion was ploughed into startups across Europe in the first three months of this year. That’s up on Q1 of 2023, when €13.7 billion found its way into startups’ bank accounts, but both are down by many billions from the exuberant, internet-heady days of 2021 and 2022.
That drop might not be such a bad thing for the longer term: Right now the market is trying not to get knocked over by the wave of startups that were generously funded at precipitous valuations in years past, which are now crashing down as they find themselves struggling to reach their revenue projections, stand up their valuations, and unable to exit on the public markets or raise more funding.
Founders at the early stages of building their startups may have already created a strong solution, identified a gap in the market, or may simply have an inescapable and driving motivation to build their own business. Ideally, they have a good combination of all three. But do they have product-market fit? And what actually is product-market fit, anyway?
The investors at Sequoia, one of the world’s biggest venture capital firms, have come up with a very handy framework to answer those two questions. It distills the landscape into three archetypes.
“Hair on Fire” roughly means that your startup addresses an urgent problem. A security startup, for example, might fit here, especially if it can win initial business on the back of parachuting in to fix a breach or other problem already in progress. Or, think of the wave of companies that offered services to businesses and users when they were suddenly sheltering in place and working from home during the peak of Covid-19.
“Hard Fact” translates as a startup that solves an existing problem better than what’s already out there. Square, which emerged as a new point of sale product in a seemingly old and saturated market, is a good example of this.
Lastly, “Future Vision” relates to deep tech, moonshots, and products out of left field. These would include quantum startups, but also those building flying cars or even autonomous vehicles that would ply our roads (or any of the tech that will be needed to make such vehicles).
Each of these archetypes will have its own customer mindset, competitive market status, opportunity/general product goals, challenges, examples of those who got it right and those that did not, and so on. Sequoia partner Jess Lee, a specialist in early-stage investing, gave a big talk on the concept at TechCrunch’s Early Stage event in Boston in April. Sequoia has written about the framework here, too.
In sum, the theory goes like this: Startups all, more or less, fit into one of these three archetypes, so identifying which archetype a company fits in can help it focus and develop.
Sequoia is confident enough of the structure that it uses the framework in its Arc program to help early-stage founders focus on how they are building. It also helps the firm evaluate potential startup investments. Beyond that, and just as importantly, founders can lean on an archetype to better anticipate and articulate the challenges and opportunities in their space. That can be helpful for decision-making internally, of course, as well as for fundraising or pitching partnerships or customers.
During her presentation on the framework, Lee said that Sequoia does not have a favored category among the three.
“I think you can create great companies in all those categories,” Lee said. Still, she admitted that certain kinds of companies might find it especially challenging to raise money in the current climate.
For deep tech and moonshots — two common kinds of startups found in the “Future Vision” category — fundraising “was easier in a zero-interest-rate period when there was a ton of capital flowing in,” Lee said. “I don’t know if [those companies] would have been able to raise as much [starting out now] as they had to, to be able to get to where they are now.”
Lee was a co-founder at Polyvore, which combined social mechanics and e-commerce — its users contributed fashion and product clips from around the web and used those products to assemble mood boards, with affiliate marketing underpinning it all. Polyvore was eventually acquired by Yahoo, and she parted ways with it. Yet, that e-commerce and consumer focus has stayed with her, she said, adding that she’s still interested in trying to find new winners in that category despite the challenges of trying to break into the space these days.
“It can still be done,” she said. “I feel like many consumer companies fall in the ‘Hard Fact’ category, and I particularly love working with consumer companies. But you have to be good at both marketing your problem as well as marketing your solution and building this. So it takes a lot to get it right.
“It almost feels like alchemy. I can’t tell you how many founders I’ve met who said, ‘Oh, yeah I was working on Snapchat, too. Like, I had my own version.’ And it sounded like it was similar, but just the right number of details allowed Snapchat to be the one that broke away.”
None of this is to say that the third category, “Hair on Fire,” is exactly easy. “You have to ruthlessly execute,” Lee said. “[You need] so much velocity to stay ahead of everyone.”
Her conclusion drives home one of the most critical aspects of building an early-stage business. “I think there’s a little bit of founder-market fit that goes into each of these product-market fit categories.”
True to its business concept, Turkey’s “instant delivery” juggernaut Getir rose quickly. Now, with the quick commerce industry in free fall, it is nosediving just as fast. On Monday, the company — once valued close to $12 billion — announced it would shut down its operations the U.S., the U.K. and Europe to focus solely on its home market of Turkey.
The move puts a bitter end to the company’s very aggressive expansion strategy that saw it raise billions of dollars to grow organically and also snap up a number of equally aggressive, yet struggling, competitors to position itself as the market leader. The closures look like they will impact at least 6,000 jobs across the closing markets, but — according to the company — just 7% of its revenues. Alongside the closures, the company said it would get a new injection of investment as a lifeline to extend its runway.
“This decision will allow Getir to focus its financial resources on Turkey,” said a statement from the company.
Getir is not the only one in this space raising money to stay afloat while also retreating from global plans. Earlier this month, reports surfaced that Flink, an erstwhile rival of Getir’s in Germany, is raising some $106 million, with around one-third of that secured so far. It comes as Flink, too, is consolidating its position. Coinciding with the fundraise leak, the company also apparently “liquidated” its operations in France.
More details, including financials, below.
Layoffs: To be clear, Getir has only officially announced cuts of 1,500 in the U.K. in the short announcement that it sent out to journalists: no details on jobs impacted elsewhere. However, reports were surfacing over the last few days that it had started to send out notices to 1,800 employees in Germany — HQ of Gorillas (which it acquired at the end of 2022). We’ve been told by a source close to the company that the number is closer to 1,100 (one figure may include contractors).
Meanwhile, when Getir acquired FreshDirect in the U.S. — only six months ago, in November 2023 — it picked up 2,300 employees. Add those different numbers together and you get around 6,000, although since Getir was already active in the U.S. prior to that acquisition, there may well be more impacted. A year ago, the company had as many as 32,000 people working for it.
Its pandemic window of opportunity: The move is a grim chapter for the startup that was founded in 2015 and saw big traction in Turkey before the pandemic — Getir means ‘bring’ in Turkish. That led to aggressive investment and expansion that peaked during Covid-19, when consumers were shopping less in person — in part to minimise infection, in part because shopping in person became really challenging due to supply issues, long lines to stagger entries and more.
Just as ride-hailing companies like Uber raised aggressively to finance aggressive growth and competitive fights across the globe, so too did Getir: between its first outside investment in 2017 and September 2023, it raised more than $2.3 billion from some 36 investors, including Sequoia, Tiger Global, Silver Lake, Mubadala, Goodwater, G Squared and A*.
It also made some aggressive acquisitions of competitors to increase its position in the market — but notably, it was consolidation intended not just a power move, but a way for other struggling, cash-strapped players in the market to step out of the brutal race.
In addition to FreshDirect and Gorillas, Getir picked up operations in Spain, Italy and the U.K. at bargain prices. It was also reportedly interested at one point in Zapp in the U.K. and Flink in Germany, so it definitely saw itself as a consolidator in the troubled market. It was a strategy also taken by Getir’s biggest global competitor, GoPuff. Today’s news leaves easier waters for GoPuff in the U.S. and the U.K.
Its Turkish window of opportunity: This is a grim chapter, but not a final one. Getir also announced that it would be raising fresh money to double down on its home market, a round led by Mubadala and G Squared.
Getir did not disclose who else was participating, nor how much it raised, nor whether this is equity or debt, so it’s hard to say what this means beyond giving the company some runway and a chance at focusing on one market that has worked.
We have reached out to some of its previous investors, Sequoia and Tiger Global, to see if they would comment on whether they are remaining investors in the company now, or whether they have cashed out.
Right now, the strategy for the bigger players in the instant grocery delivery market seems to be: accept that our international strategies were not great ideas, and focus on just our core markets for now.
The writing on the wall: Getir, like its peers in the instant delivery market, has been struggling for a while. In May 2023, it cut 14% of staff and cancelled large parts of its geographic expansion plans as it scrambled to right-size the business ahead of more fundraising. Just weeks after that, it pulled out of Spain, Italy and Portugal in July 2023. At the time, it was well understood that it shut operations because those markets were just not thriving, but Getir was indeed trying to close another round of funding, so cutting loss-making operations makes sense in that context.
Documents have been shared with TechCrunch that indicate that the company, for the calendar year 2023, the company made $3.3 billion, with the U.S. and Europe (including the U.K.) accounting for around $1 billion of that across the year. (It’s not clear from Getir’s statement what the 7% figure relates to. We are asking.) From the documents that we have seen, as of the end of last year, the company was not Ebitda positive in any of its geographies.
Big, bad news in the chaotic market for instant delivery services, but given the states of the venture market, the current economy, and consumer behavior these days — yes, people buy online, but they are also very much back outside, shopping like before — it is likely not the last.
Perplexity, the AI search engine startup, is a hot property at the moment. TechCrunch has learned that the company is currently raising at least $250 million more at a valuation of between $2.5 billion and $3 billion.
The news comes on the heels of two other big fundraises that have seen company’s valuation leapfrog in the last four months: in January the company raised nearly $74 million at a valuation of $540 million (up from $121 million in April 2023). And at the beginning of March, the company closed $56 million on a valuation of $1 billion — a raise that has been quietly public since then (it was on PitchBook’s database for one), and which Bloomberg highlighted earlier today.
Those two reported rounds are not the full story. We understand from multiple sources close to the company that it is actually also raising a further round to capitalize on the attention it’s getting in the market. NEA and IVP, both previous backers of the company, are among those looking to invest in this larger round, according to sources.
Whether they or other previous backers participate, a source said, may depend on how willing Perplexity is to work with existing investors rather than diversity, expanding its cap table to bring in new investors.
“They are growing very rapidly,” a partner from an existing investor said. “Yes we will look to participate.”
The core of Perplexity’s product is a generative AI-based search engine that provides results using a chatbot-style interface. It’s definitely not the only company in generative AI pursuing the search opportunity: that is essentially how many people are using products like ChatGPT and Microsoft’s Bing (powered by OpenAI), and Google is making a big push to improve search results with its Gemini LLM.
But Perplexity is building its algorithms incorporating a variety of LLMs, the idea being that this produces a more accurate and richer response.
“Unlike other enterprise tools for knowledge work like Microsoft Copilot, Perplexity Enterprise Pro is also the only enterprise AI offering that offers all the cutting-edge foundation models in the market in one single product: OpenAI GPT-4, Anthropic Claude Opus, Mistral, and more to come,” CEO and co-founder Aravind Srinivas noted earlier today. “This gives customers and users choices to explore and customize their experience depending on their use cases.” That “more to come” may well be including more from Hugging Face and Meta, if Srinivas’s public endorsements and investor lists are anything to go by.
Considering that the company has only been around since 2022, Perplexity’s current investor list is already long, running to 46 names according to PitchBook data.
In addition to IVP and NEA, it includes other notable VCs such as Sequoia, Bessemer and Kindred; strategic backers like Nvidia, Databricks and Bezos Expeditions; and many recognizable individuals such as Jeff Bezos, Meta’s chief AI scientist Yann LeCun, Naval Ravikant, Susan Wojcicki, Elad Gil, Nat Friedman, and Clément Delangue from Hugging Face. A newer backer, Daniel Gross, led the $56 million round from March with other new backers Stanley Druckenmiller, Y Combinator head Garry Tan and Figma’s CEO Dylan Field also participating, among others.
One fundraise coming rapidly on the heels of another is reminiscent of rolling fundraising that we’ve seen from other big startups over the years. In the years leading up to is IPO during a time of rapid growth and major attention, Snap regularly appeared to be raising money on an ongoing basis. These days, it appears to be all about AI, with companies like OpenAI, Anthropic and Mistral all raising at a rapid pace and seeing their valuations skyrocket along with that.
In the case of Perplexity, the startup is standing out in the market for a couple of reasons. Most obviously, it’s one of the ambitious, albeit smaller, hopefuls in the race to build generative AI services. Its unique position in the market is that it’s not focused on the race to build multi-purpose large language models. Instead, taking a page from one of the biggest technology companies in the world today, it is tacking one specific product, at least for now: search.
Perplexity is not the only startup in AI that is building on very focused opportunities and by targeting enterprise. Synthesia in the UK is taking a similar approach with AI video tools, aiming them specifically at the business market, for the building of training and customer support video content.
In the case of Perplexity, the startup offers its tools on free and enterprise, paid tiers, and so far its processed 75 million queries this year and is currently on ARR of $20 million, according to Bloomberg.
Its reason for raising again so soon? Yes, perhaps to capitalize on customer and investor interest at what one investor described as a “zeitgeist moment” for the startup. But also because of the mechanics of building any kind of AI service right now.
“Compute is very expensive, so they may need to raise” for that reason alone, one said.
We have reached out to Srivinivas for comment and will update this post as we learn more.
If corporate boards of directors signal something about where a company wants to be focusing, Amazon’s just made an interesting move to change up its direction. The company has just announced that Andrew Ng — a prominent figure in the world of AI, and specifically building AI within large tech companies — is joining its board of directors. The company also announced, with a little less fanfare, that Judy McGrath — who is best known for her work as a longtime TV executive, running MTV and helping Viacom become a media powerhouse — will be stepping down as a director.
Taken together, the two moves sketch out an interesting picture about the tech giant.
After many costly years of wanting to go all-on on building out an entertainment empire — Amazon spent almost $19 billion on its video and music business in 2023 — it’s interesting that someone who would have been an important advocate and advisor on that strategy is not going to stand for reelection.
To be clear, this is certainly not a sign that Amazon is not going to continue to be a huge force in streaming entertainment, be it video or music, gaming or anything else. It’s now folding in advertising across Prime Video, one big reason to keep audiences strong. But it will be interesting to see how investment plays out in that segment in 2024 overall. Layoffs in the studio and video divisions, plus winding down activities in some regions, are two signals it could be smaller. And given the AI whiplash that every big tech company feels currently, it feels like a timely detail that now McGrath is stepping away from the board.
On that note, looking ahead, to stay at the front of the tech industry, Amazon will overall be looking for better thought leadership on how it takes its next steps in its artificial intelligence strategy.
It’s worth remembering that Amazon is and has been a leading player in AI for a long time. Its Alexa assistant and Echo devices arguably put voice recognition and connected assistants on the map, spurring many copycat products from rivals; it has been working on autonomous services in airborne and ground-level delivery and in-store purchasing; it uses machine learning to improve how products are targeted; AWS is a big player in AI compute; more recently it has poured billions into investments in big AI startups; and so on.
Yet for at least a year, in the wake of the GPT advancements made by OpenAI and Microsoft’s giant investment in that startup (and the massive scramble from a plethora of other tech companies following that), Amazon has grappled with the impression internally and externally that it is “falling behind” on the technology.
Is it true? Is it just optics? Regardless of what the real answer is, Ng’s appointment can only be helpful in advancing its AI profile, giving it more thought leadership on real innovation, not just making follow-on moves, in the space.
Ng is potentially a triple-threat style board appointment, with experience across academia, investing, and hands-on building, usually holding roles across all three simultaneously. His current roles include an adjunct professorship at Stanford; a role as general partner at the AI Fund, a venture studio building and backing AI startups; the head of edtech company DeepLearning.AI and the founder of Landing AI, a computer vision startup. Oh, and he’s also still chair of Coursera, another edtech startup he founded and used to lead.
He has also previously been the chief scientist and a VP at Baidu, the Chinese search giant; and he founded and led Google Brain, which was that search giant’s first big foray into building and applying AI tech across its products.
Amazon does not provide any statement from Ng himself in its announcement. We have reached out to him directly to see if he might tell us a bit more about what he sees as Amazon’s big opportunity and we’ll update when and if we hear back. Right now, it absolutely feels like a new wave of companies and thinkers are setting the pace in AI, but the Amazons of the world are certainly not standing by idly.
Travel and tourism are very much back on the map for consumers and the business world. Now, to underscore that surge, one of the startups building software in the space has closed a big round of funding. Guesty — which has built a platform for accommodation managers to manage all aspects of their business on platforms like Airbnb, Vrbo and directly to travellers — has raised $130 million.
The company, based out of New York with roots in Israel, says it expects to turn profitable this year and has grown revenues 5x in the last three years (without specifying actual revenue figures). Sources confirmed to TechCrunch that the Series F values Guesty at around $900 million post-money.
KKR is leading this round, with Apax Funds, BDT & MSD Partners and Sixth Street also participating.
To put the funding into some context: post-Covid, the tourism and travel sector has been on a strong rebound, with the World Tourism and Travel Council estimating that in 2024, globally it will generate a record-breaking $11.1 trillion in sales, despite the U.S. and China still not back to pre-pandemic levels of strength.
The upswing has been in progress for the last couple of years, and it’s played out with a number of nine-figure funding rounds not just for Guesty, but its competitors and startups in the space. Guesty’s last round, in August 2022, was a Series E of $170 million that valued it at $690 million. Guesty’s close competitor Hostaway raised $175 million in May last year (its first ever big funding round). Within a day of that news, GetYourGuide raised a monster $194 million at a $2 billion valuation.
And Mews, which like Guesty builds SaaS but for hoteliers, raised $110 million at a $1.2 billion valuation last month (March 2024). It is a reminder that investors are still willing to sign term sheets in the right circumstances.
“It’s definitely tough market. In every round I’ve raised, I would always get 40 no’s for every yes,” Amiad Soto, Guesty’s CEO, said in an interview. Now, with Guesty “closing in on becoming profitable this year,” he joked that “I still got 40 no’s, but also a lot more yes’s.”
Soto, who co-founded Guesty with his brother Koby (who is no longer with the company), plans to deploy the funding across a few different areas.
First of all, Guesty wants to continue expanding its existing platform for current customers. That business today already covers “hundreds of thousands” of properties, he said — he declined several times to give me a more specific figure — and it will double down on the one-stop-shop concept that a lot of other B2B tech companies are pursuing today.
The platform covers the basics of listing and booking management software, analytics, accounting tools, the ability to manage multiple properties, and CRM features. More recently, it’s added enhanced payment services and capital advances (built in-house not white labelled from third parties, Soto said), damage protection services (that is dipping into the area of insurance), website building tools, and price optimization services that all integrate with the dozens of interfaces where a property manager might list a room or home for travellers to book.
Second of all, while the main focus to date for Guesty has been short-term lets — booked typically for less than a month — now the company wants to expand into the medium-term space. This will open it up to more people who might be living temporarily in a location for a specific work assignment, for example.
Third of all, Soto said that Guesty wants to consider more acquisitions. This is because the market may not be looking favorably on all startups right now, but this is less a comment the strength of startups (talent and innovations) as it is on the state of venture right now. This basically means there are a lot of potentially very interesting companies out there that might be ready to entertain getting acquired at less bullish valuations.
Stephen Shanley, partner and head of Europe Tech Growth at KKR, along with Lauriane Requena, a principal at KKR Tech Growth, and Dennis Kavelman, Inovia Capital partner, are all joining the board with this round. “Guesty is a best-in-class operator and one of the clear leaders in the property management sector,” said Shanley in a statement. “There has been a significant shift towards the short- term rental market and this investment will support the company as it continues to meet that growing customer need.”
At an event in London on Tuesday, Meta confirmed that it plans an initial release of Llama 3 — the next generation of its large language model used to power generative AI assistants — within the next month.
This confirms a report published on Monday by The Information that Meta was getting close to launch.
“Within the next month, actually less, hopefully in a very short period of time, we hope to start rolling out our new suite of next-generation foundation models, Llama 3,” said Nick Clegg, Meta’s president of global affairs. He described what sounds like the release of several different iterations or versions of the product. “There will be a number of different models with different capabilities, different versatilities [released] during the course of this year, starting really very soon.”
The plan, Meta Chief Product Officer Chris Cox added, will be to power multiple products across Meta with Llama 3.
Meta has been scrambling to catch up to OpenAI, which took it and other big tech companies like Google by surprise when it launched ChatGPT over a year ago and the app went viral, turning generative AI question and answers into everyday, mainstream experiences.
Meta has largely taken a very cautious approach with AI, but that hasn’t gone over well with the public, with previous versions of Llama criticized as too limited. (Llama 2 was released publicly in July 2023. The first version of Llama was not released to the public, yet it still leaked online.)
Llama 3, which is bigger in scope than its predecessors, is expected to address this, with capabilities not just to answer questions more accurately but also to field a wider range of questions that might include more controversial topics. It hopes this will make the product catch on with users.
“Our goal over time is to make a Llama-powered Meta AI be the most useful assistant in the world,” said Joelle Pineau, Vice President AI Research. “There’s quite a bit of work remaining to get there.” The company did not talk about the size of the parameters it’s using in Llama 3, nor did it offer any demos of how it would work. It’s expected to have about 140 billion parameters, compared to 70 billion for the biggest Llama 2 model.
Most notably, Meta’s Llama families, built as open-source products, represent a different philosophical approach to how AI should develop as a wider technology. In doing so, Meta is hoping to play into wider favor with developers versus more proprietary models.
But Meta is also playing it more cautiously, it seems, especially when it comes to other generative AI beyond text generation. The company is not yet releasing Emu, its image generation tool, Pineau said.
“Latency matters a lot along with safety along with ease of use, to generate images that you’re proud of and that represent whatever your creative context is,” Cox said.
Ironically — or perhaps predictably (heh) — even as Meta works to launch Llama 3, it does have some significant generative AI skeptics in the house.
Yann LeCun, the celebrated AI academic who is also Meta’s chief AI scientist, took a swipe at the limitations of generative AI overall and said his bet is on what comes after it. He predicts that will be joint embedding predicting architecture (JEPA), a different approach both to training models and producing results, which Meta has been using to build more accurate predictive AI in the area of image generation.
“The future of AI is JEPA. It’s not generative AI,” he said. “We’re going to have to change the name of Chris’s product division.”
@2025 – All Right Reserved. Designed and Developed Startupnews