“It was the best of times, it was the worst of times, it was the age of wisdom, it was the age of foolishness,……….. it was the spring of hope, it was the winter of despair.” the opening lines of “A tale of two cities“
This tweet and the opening lines pretty much sum up the venture world we live in.
On one hand, there is a significant slowdown in VC activity, lower investments and deal volumes, significant drop in early-stage funding, big squeeze on growth capital, layoffs galore, corporate governance issues, and write-downs. (A lot of it maybe justified, but more on this later.)
On the other hand, the market is abuzz with AI for the past six months. We have been looking from the sidelines as these developments unfold at a frantic pace. Using this A16Z framework to make sense of the ecosystem, here are the developments as we see them.
There has been a Cambrian explosion of GenAI startups, solving for generating text, content, video, code, mobile apps, games, etc. This compilation from Antler shows the number mushrooming.
We foresee a couple of winners maybe, but a lot of mortality in the long run in these type of applications space. There is the ever-present threat of tech majors offering these features for free in their existing applications. Besides that, the battle will be to make the solution relevant in the context of the user. This will mean tighter and seamless integration into other adjacent applications, more specialized focus on type of user.
There has been lot of action in other parts of the tech stack as well. There has been a race among various large cloud platforms to be active players in the model layer, either through partnerships (Google cloud & Snowflake or by direct ownership.) And as is the norm, when certain areas catch frenzy, $100M seed rounds to disrupt the majors begin to happen. Some other notable investments in the last 6 months are Anthropic’s $350M Series C to build next-gen AI Assistants, Adept’s $350M Series B to build AI that helps humans and machines work seamlessly (specific example of AI actually performing tasks using various software solutions), and SandboxAQ’s $500M raise to focus on cybersecurity.
In our view, the model space will be a commodity business over the mid-to-long term, will need significant capital as well as distribution chops, and will very likely be a big boys’ game. Vertical offerings and specialised solutions for niches are where the not-so-apparent opportunities lie, but access to proprietary data will be paramount in these spaces.
What has caught our attention recently is the M&A activity in this space, led by cloud data behemoths (Snowflake, Databricks).
Snowflake and Databricks announced their acquisitions of Neeva and MosaicML, respectively. Despite seemingly different objectives, Neeva has been helping Snowflake’s customers leverage their data better through better inferencing, and Databricks is looking to provide enterprises a secure way to build their own LLMs and train their own state-of-the-art models. The main driver is leveraging AI, without compromising or sharing own proprietary data.
A very interesting acquisition is that of Thomson Reuters acquiring CaseText. Casetext’s flagship product CoCounsel leverages AI to help attorneys and lawyers review documents, help with legal research memos, prepare depositions, and analyse contracts. Casetext was one of the few granted early access to OpenAI’s GPT-4 language model. While the stated objective being bringing AI to Thomson Reuters customers, given the vast amounts of data that CaseText has access to, we believe the objective is as much owning the data that can be used to build LLMs.
I suspect we will see a lot more M&A action driven by data ownership as well as investments in Vertical AI as time progresses.
Other stories from the AI world
On the other side of the world though, the story is not so great. The much-anticipated startup in China, Light Years Beyond or the so-called OpenAI for China is being bought out and capital being returned to its investors. Formal reasons notwithstanding, has this been driven by restrictions to semi-conductor access by China?
As for some unpleasant fallout of the AI rage, we just made it harder for brands to reach consumers and be credible. Read about how next-gen content farms powered by GenAI are siphoning off ad spends in this article. Surely, there is an opportunity for fraud prevention solutions in this space.
Meanwhile, in Europe, the draft AI Regulation was passed a couple weeks ago by EU Parliament. The regulation covers many aspects from labelling AI-generated content to publishing summaries of copyrighted material used to train models. It takes a risk-based approach, introducing restrictions on solutions based on how dangerous lawmakers perceive it to be. While there is a still some way before this gets passed as law, it could set the tone across the world for regulators.
Prompt engineering courses and guides have started mushrooming as well. Could this be a precursor to the next SEO?
Elsewhere in the startups and VC ecosystem…
In H1 2023, global funding reached $144 billion, marking a 51% decline from the $293 billion invested in H1 2022 and a 10% decline from the second half of 2022.
On the upside, startups continued to make significant strides towards becoming efficient, resulting in declining burn numbers.
As markets begin valuing capital efficiency over growth, Indian SaaS companies continue to be capital efficient (a hallmark of Indian B2B SaaS companies), with 8 out of 10 startups recording a burn multiple of less than 1.5x, (well below the global average of 2x-3x according to an EY-Upekkha report).
Latest Indian B2B Action
SaaS firm RevSure.AI scooped $10 million in a seed round. The company offers intelligence to enterprises for improving marketing and sales efficiencies.
Another SaaS firm Togai also raised a seed round. Togai is a B2B SaaS metering and pricing platform offering an innovative approach to crafting monetisation tools for dynamic billing and pricing.
Soptle, India's first SaaS-led B2B marketplace for FMCG manufacturers & retailers closed a USD 1 million pre-seed investment in its latest funding round in May.
Arali Perspectives
As we continue to evolve our thinking on the challenges faced by modern enterprises, we are increasingly diving deeper into the world of data governance & data privacy. Presenting a series of our perspectives on data privacy, governance, & data in the context of generative AI. Read Part 1 of the series.
From Arali’s Portfolio
First for the big news! Arali Ventures led a seed round along with Capital-A, Java Capital & others to invest in FRIGATE—a B2B manufacturing startup. Read more.
Oivi recently inaugurated its new Bangalore office with the Norwegian Ambassador to India Hans Jacob Frydenlund.
Unbox Robotics is part of the Oxagon x McLaren Accelerator. The company also won the 2023 IDEA India Digital Enabler Awards under the category “Best Use of AI/Bots.”
HBox was selected by a top-rated Michigan cardiac care centre for HBox Virtual Care. With this, HBox seeks to further expand its foothold in the Midwest through partnerships.
FinBox, along with IIFL and Finvu, was awarded for their work in MSME lending at the SamvAAd event in Mumbai. The company also recently participated in the prestigious Open Finance Summit in Indonesia.
WIZ was awarded "Fastest Growing Logistics Startup" at The 5th India Logistics & Warehousing Excellence Awards, organised by the Institute of Supply Chain Management (ISCM).
BimaKavach has partnered with EquityList to provide free business risk advisory and customised insurance recommendations to EquityList customers and help insure their business risks, founders, key assets and operations at optimal prices.
Cynlr launched Cybernetics H.I.V.E, a state-of-the-art research hub in Vision, Intelligence and Robotics, in Whitefield, Bangalore in February.
Prolance received the IKC Business Innovation Award 2023 and was recognised as a "Start-Up 50 Trailblazer" by Dun & Bradstreet in June.
Signing off, folks!
If you are building something in the enterprise tech space from India or associated with enterprise tech in general, we would love to hear from you.
Also, this post is public, so feel free to share it!
Glad that you share my guarded skepticism on AI opportunities.
It's the data that's going to ultimately be most critical. Tools will, eventually, be just that - tools.